Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)
UvA-DARE (Digital Academic Repository)
MRI for Crohn's diseaseFrom quantification to automationPuylaert, C.A.J.
Link to publication
Creative Commons License (see https://creativecommons.org/use-remix/cc-licenses):Other
Citation for published version (APA):Puylaert, C. A. J. (2018). MRI for Crohn's disease: From quantification to automation.
General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).
Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.
Download date: 28 Sep 2020
MRI FOR CROHN’S DISEASE:FROM QUANTIFICATION TO AUTOMATION
Carl A.J. Puylaert
MR
I FO
R C
RO
HN
’S D
ISE
AS
E: F
RO
M Q
UA
NT
IFIC
AT
ION
TO
AU
TO
MA
TIO
N C
arl A
.J. P
uyla
ert
Cover.indd All Pages 12/9/18 21:32
UITNODIGINGVoor het bijwonen van de openbare
verdediging van het proefschrift
MRI FOR CROHN’S DISEASE:FROM QUANTIFICATION
TO AUTOMATION
door
Carl Puylaert
Op donderdag 15 november 2018 om 14:00 in de Agnietenkapel van
de Universiteit van AmsterdamOudezijds Voorburgwal 231
Amsterdam
Receptie ter plaatse na afloop van de promotie
Carl PuylaertRustenburgerstraat 229-3
1073 GA, Amsterdam
06-15215881
Paranimfen:
Catherine [email protected]
06-11070412
Olvert [email protected]
06-50640649
Uitnodiging promotie.indd 1 12/9/18 21:31
MRI FOR CROHN’S DISEASE:FROM QUANTIFICATION TO AUTOMATION
Carl A.J. Puylaert
Proefschrift2018-new2.indb 1 13/9/18 10:06
MRI FOR CROHN’S DISEASE:FROM QUANTIFICATION TO AUTOMATION
Carl A.J. Puylaert
Proefschrift2018-new2.indb 1 13/9/18 10:06
MRI for Crohn’s disease: from quantification to automation
© Copyright Carl Alejandro Julien Puylaert, Amsterdam 2018 No part of this thesis may be reproduced, stored or transmitted, in any for or by any means, without prior permission of the author.
ISBN: 978-94-6375-121-6Lay-out: Carl PuylaertCover design: Bram van LeeuwenPrinting: Ridderprint BV | www.ridderprint.nl
This thesis was prepared at the Department of Radiology and Nuclear Medicine, Amsterdam UMC, University of Amsterdam, the Netherlands.
The printing of this thesis was financially supported by Robarts Clinical Trials, Takeda Nederland BV and ABN AMRO.
The research described in this thesis was financially supported by the European Union’s Seventh Framework Programme (project number 270379).
Proefschrift2018-new2.indb 2 13/9/18 10:06
MRI for Crohn’s disease: from quantification to automation
© Copyright Carl Alejandro Julien Puylaert, Amsterdam 2018 No part of this thesis may be reproduced, stored or transmitted, in any for or by any means, without prior permission of the author.
ISBN: 978-94-6375-121-6Lay-out: Carl PuylaertCover design: Bram van LeeuwenPrinting: Ridderprint BV | www.ridderprint.nl
This thesis was prepared at the Department of Radiology and Nuclear Medicine, Amsterdam UMC, University of Amsterdam, the Netherlands.
The printing of this thesis was financially supported by Robarts Clinical Trials, Takeda Nederland BV and ABN AMRO.
The research described in this thesis was financially supported by the European Union’s Seventh Framework Programme (project number 270379).
Proefschrift2018-new2.indb 2 13/9/18 10:06
MRI FOR CROHN’S DISEASE:FROM QUANTIFICATION TO AUTOMATION
ACADEMISCH PROEFSCHRIFT
ter verkrijging van de graad van doctor
aan de Universiteit van Amsterdam
op gezag van de Rector Magnificus
prof. dr. ir. K.I.J. Maex
ten overstaan van een door het College voor Promoties ingestelde commissie,
in het openbaar te verdedigen in de Agnietenkapel
op donderdag 15 november 2018, te 14:00 uur
door
Carl Alejandro Julien Puylaert
geboren te ‘s-Gravenhage
Proefschrift2018-new2.indb 3 13/9/18 10:06
MRI FOR CROHN’S DISEASE:FROM QUANTIFICATION TO AUTOMATION
ACADEMISCH PROEFSCHRIFT
ter verkrijging van de graad van doctor
aan de Universiteit van Amsterdam
op gezag van de Rector Magnificus
prof. dr. ir. K.I.J. Maex
ten overstaan van een door het College voor Promoties ingestelde commissie,
in het openbaar te verdedigen in de Agnietenkapel
op donderdag 15 november 2018, te 14:00 uur
door
Carl Alejandro Julien Puylaert
geboren te ‘s-Gravenhage
Proefschrift2018-new2.indb 3 13/9/18 10:06
PROMOTIECOMMISSIE
Promotores: Prof. dr. J. Stoker AMC-UvA
Prof. dr. ir. L.J. van Vliet Technische Universiteit Delft
Co-promotores: Dr. C.Y. Ponsioen AMC-UvA
Dr. F.M. Vos Technische Universiteit Delft
Overige leden: Prof. dr. J.S. Laméris AMC-UvA
Prof. dr. G.R.A.M. D’Haens AMC-UvA
Prof. dr. W.A. Bemelman AMC-UvA
Prof. dr. W.M. Prokop Radboud Universiteit
Prof. dr. J.P.W. Pluim Technische Universiteit Eindhoven
Dr. D.J. de Jong Radboud Universiteit
Faculteit der Geneeskunde
Proefschrift2018-new2.indb 4 13/9/18 10:06
PROMOTIECOMMISSIE
Promotores: Prof. dr. J. Stoker AMC-UvA
Prof. dr. ir. L.J. van Vliet Technische Universiteit Delft
Co-promotores: Dr. C.Y. Ponsioen AMC-UvA
Dr. F.M. Vos Technische Universiteit Delft
Overige leden: Prof. dr. J.S. Laméris AMC-UvA
Prof. dr. G.R.A.M. D’Haens AMC-UvA
Prof. dr. W.A. Bemelman AMC-UvA
Prof. dr. W.M. Prokop Radboud Universiteit
Prof. dr. J.P.W. Pluim Technische Universiteit Eindhoven
Dr. D.J. de Jong Radboud Universiteit
Faculteit der Geneeskunde
Proefschrift2018-new2.indb 4 13/9/18 10:06
TABLE OF CONTENTS
Chapter 1 Introduction 7
Chapter 2 Grading of Crohn’s disease activity using CT, MRI, US and scintigraphy: a meta-analysis Eur Radiol. 2015 Nov;25(11):3295-313
17
Chapter 3 Comparison of MRI activity scoring systems and features for the terminal ileum in Crohn’s disease patientsAm J Roentgenol. Accepted for publication.
55
Chapter 4 Semiautomatic assessment of the terminal ileum and colon in patients with Crohn disease using MRI (the VIGOR++ project)Acad Radiol. 2018 Aug;25(8):1038-1045
79
Chapter 5 Comparison of contrast-enhanced and diffusion-weighted MRI in assessment of the terminal ileum in Crohn’s disease patientsAbdom Radiol. Accepted for publication.
109
Chapter 6 Quantified terminal ileal motility during MR enterography as a biomarker of Crohn’s disease activity: a prospective studyRadiology. Accepted for publication.
125
Chapter 7 Long-term performance of readers trained in grading Crohn’s disease activity using MRIAcad Radiol. 2016 Dec;23(12):1539-1544
149
Chapter 8 General discussion 165
Chapter 9 Summary 175
Chapter 10 Nederlandse samenvatting 181
Appendix List of contributing authors 188
Portfolio 190
List of publications 192
Curriculum vitae 195
Dankwoord 196
Proefschrift2018-new2.indb 5 13/9/18 10:06
TABLE OF CONTENTS
Chapter 1 Introduction 7
Chapter 2 Grading of Crohn’s disease activity using CT, MRI, US and scintigraphy: a meta-analysis Eur Radiol. 2015 Nov;25(11):3295-313
17
Chapter 3 Comparison of MRI activity scoring systems and features for the terminal ileum in Crohn’s disease patientsAm J Roentgenol. Accepted for publication.
55
Chapter 4 Semiautomatic assessment of the terminal ileum and colon in patients with Crohn disease using MRI (the VIGOR++ project)Acad Radiol. 2018 Aug;25(8):1038-1045
79
Chapter 5 Comparison of contrast-enhanced and diffusion-weighted MRI in assessment of the terminal ileum in Crohn’s disease patientsAbdom Radiol. Accepted for publication.
109
Chapter 6 Quantified terminal ileal motility during MR enterography as a biomarker of Crohn’s disease activity: a prospective studyRadiology. Accepted for publication.
125
Chapter 7 Long-term performance of readers trained in grading Crohn’s disease activity using MRIAcad Radiol. 2016 Dec;23(12):1539-1544
149
Chapter 8 General discussion 165
Chapter 9 Summary 175
Chapter 10 Nederlandse samenvatting 181
Appendix List of contributing authors 188
Portfolio 190
List of publications 192
Curriculum vitae 195
Dankwoord 196
Proefschrift2018-new2.indb 5 13/9/18 10:06
Proefschrift2018-new2.indb 6 13/9/18 10:06 Proefschrift2018-new2.indb 6 13/9/18 10:06
CHAPTER 1
Introduction
Proefschrift2018-new2.indb 7 13/9/18 10:06
CHAPTER 1
Introduction
Proefschrift2018-new2.indb 7 13/9/18 10:06
Chapter 1
8
INTRODUCTION
Crohn’s disease is a widely prevalent inflammatory bowel disease (IBD) that
manifests with variable degrees of progression and severity. The cause of Crohn’s
disease is unknown, although its incidence has been associated with environmental
and genetic factors [1]. Crohn’s disease has several notable characteristics, by
which it can be distinguished from other forms of IBD or gastrointestinal diseases.
In particular, it is commonly located in the small bowel and/or colon, although
inflammatory lesions may be present in any part of the gastrointestinal tract and
deep ulcerations can extend into the muscular layers of the intestinal wall. Affected
parts are often interspersed with stretches of healthy tissue (“skip lesions”). In their
lifetime, up to 50% of patients will develop a form of intestinal complication, such
as stricturing or penetrating disease [2].
Crohn’s disease treatment is focused on reduction of inflammation and relief of
symptoms, although the focus is nowadays shifting towards bringing and keeping
the mucosal inflammation into remission. Therapeutic drugs are the primary form of
treatment, while surgery is reserved for patients with refractory disease or in some
cases as an alternative to medical treatment [3]. Three out of four Crohn’s disease
patients will undergo surgery within 10 years of their diagnosis, of which 44%
will have had post-operative recurrence within 10 years of surgery [4]. Although
therapeutic methods are continuously improving, the management of Crohn’s
disease remains one of the major challenges in gastrointestinal medicine.
Crohn’s disease commonly manifests in a pattern of recurrence and remission,
therefore requiring attentive observation of intestinal inflammation. A combination
of clinical, radiologic, endoscopic and biochemical techniques can be used for
diagnosis and further evaluation of Crohn’s disease. Ileocolonoscopy is regarded
as the reference standard for evaluation of the terminal ileum and colon, the most
common locations for Crohn’s disease. Endoscopic scores, such as the Crohn’s
Disease Endoscopic Index of Severity (CDEIS) or the Simplified Endoscopic
Activity Score for Crohn’s Disease (SES-CD) can be used to grade endoscopic
disease activity [5,6]. Mucosal healing is frequently used as a primary aim for
therapy, as it is associated with lower rates of hospitalization and surgery [7].
Despite the capabilities of ileocolonoscopy, it has several limitations: firstly,
endoscopic techniques are invasive, present the risk of perforation and require
Proefschrift2018-new2.indb 8 13/9/18 10:06
Chapter 1
8
INTRODUCTION
Crohn’s disease is a widely prevalent inflammatory bowel disease (IBD) that
manifests with variable degrees of progression and severity. The cause of Crohn’s
disease is unknown, although its incidence has been associated with environmental
and genetic factors [1]. Crohn’s disease has several notable characteristics, by
which it can be distinguished from other forms of IBD or gastrointestinal diseases.
In particular, it is commonly located in the small bowel and/or colon, although
inflammatory lesions may be present in any part of the gastrointestinal tract and
deep ulcerations can extend into the muscular layers of the intestinal wall. Affected
parts are often interspersed with stretches of healthy tissue (“skip lesions”). In their
lifetime, up to 50% of patients will develop a form of intestinal complication, such
as stricturing or penetrating disease [2].
Crohn’s disease treatment is focused on reduction of inflammation and relief of
symptoms, although the focus is nowadays shifting towards bringing and keeping
the mucosal inflammation into remission. Therapeutic drugs are the primary form of
treatment, while surgery is reserved for patients with refractory disease or in some
cases as an alternative to medical treatment [3]. Three out of four Crohn’s disease
patients will undergo surgery within 10 years of their diagnosis, of which 44%
will have had post-operative recurrence within 10 years of surgery [4]. Although
therapeutic methods are continuously improving, the management of Crohn’s
disease remains one of the major challenges in gastrointestinal medicine.
Crohn’s disease commonly manifests in a pattern of recurrence and remission,
therefore requiring attentive observation of intestinal inflammation. A combination
of clinical, radiologic, endoscopic and biochemical techniques can be used for
diagnosis and further evaluation of Crohn’s disease. Ileocolonoscopy is regarded
as the reference standard for evaluation of the terminal ileum and colon, the most
common locations for Crohn’s disease. Endoscopic scores, such as the Crohn’s
Disease Endoscopic Index of Severity (CDEIS) or the Simplified Endoscopic
Activity Score for Crohn’s Disease (SES-CD) can be used to grade endoscopic
disease activity [5,6]. Mucosal healing is frequently used as a primary aim for
therapy, as it is associated with lower rates of hospitalization and surgery [7].
Despite the capabilities of ileocolonoscopy, it has several limitations: firstly,
endoscopic techniques are invasive, present the risk of perforation and require
Proefschrift2018-new2.indb 8 13/9/18 10:06
Introduction
9
bowel preparation. Furthermore, the proximal small bowel is frequently inaccessible
with routine endoscopic techniques. Lastly, no evaluation of extraluminal disease is
possible.
Cross-sectional imaging techniques, such as computerized tomography (CT),
ultrasound (US), scintigraphy and magnetic resonance imaging (MRI) overcome
these limitations and can provide imaging biomarkers for disease activity [8].
Visualization of abdominal complications, such as obstructive stenosis, fistulae, or
abscesses, is important to guide therapy and planning of surgical intervention. A
meta-analysis found no significant differences between different imaging techniques
(CT, MRI, US and scintigraphy) in terms of diagnostic accuracy [8]. Further use
of cross-sectional imaging techniques focuses on the evaluation and grading of
disease activity in patients with Crohn’s disease.
Magnetic resonance imaging
MRI is increasingly used for evaluation of small bowel Crohn’s disease, due to its
high spatial and contrast resolution, lack of ionizing radiation and large array of
MRI features that can be used to assess disease activity [9]. Luminal distension
of the small bowel is essential for evaluation, for which two methods can be used:
MR enterography, which uses oral administration of luminal contrast, and MR
enteroclysis, which requires placement of a nasojejunal tube under fluoroscopic
guidance. MR enterography is generally preferred, due to its similar accuracy to MR
enteroclysis and lower invasiveness [10]. A biphasic hyperosmotic contrast agent
(e.g. water + mannitol) is preferred, because of the good luminal distension and the
excellent image contrast between bowel wall and lumen on T1- and T2-weighted
images. For evaluation of the entire bowel, additional colonic distension is required
through the use of a rectal enema, as regular oral administration of luminal contrast
does not provide consistent distension of distal colonic segments [11]. On the other
hand, early intake of oral contrast might increase colonic distension, without the
need for a rectal enema.
Common findings on MRI of active inflammation include increased bowel wall
thickness, enhancement of the bowel wall on T1-weighted sequences after
intravenous contrast administration, increased signal intensity on fat saturated
T2-weighted sequences and extramural findings, such as the comb sign, creeping
Proefschrift2018-new2.indb 9 13/9/18 10:06
Introduction
9
bowel preparation. Furthermore, the proximal small bowel is frequently inaccessible
with routine endoscopic techniques. Lastly, no evaluation of extraluminal disease is
possible.
Cross-sectional imaging techniques, such as computerized tomography (CT),
ultrasound (US), scintigraphy and magnetic resonance imaging (MRI) overcome
these limitations and can provide imaging biomarkers for disease activity [8].
Visualization of abdominal complications, such as obstructive stenosis, fistulae, or
abscesses, is important to guide therapy and planning of surgical intervention. A
meta-analysis found no significant differences between different imaging techniques
(CT, MRI, US and scintigraphy) in terms of diagnostic accuracy [8]. Further use
of cross-sectional imaging techniques focuses on the evaluation and grading of
disease activity in patients with Crohn’s disease.
Magnetic resonance imaging
MRI is increasingly used for evaluation of small bowel Crohn’s disease, due to its
high spatial and contrast resolution, lack of ionizing radiation and large array of
MRI features that can be used to assess disease activity [9]. Luminal distension
of the small bowel is essential for evaluation, for which two methods can be used:
MR enterography, which uses oral administration of luminal contrast, and MR
enteroclysis, which requires placement of a nasojejunal tube under fluoroscopic
guidance. MR enterography is generally preferred, due to its similar accuracy to MR
enteroclysis and lower invasiveness [10]. A biphasic hyperosmotic contrast agent
(e.g. water + mannitol) is preferred, because of the good luminal distension and the
excellent image contrast between bowel wall and lumen on T1- and T2-weighted
images. For evaluation of the entire bowel, additional colonic distension is required
through the use of a rectal enema, as regular oral administration of luminal contrast
does not provide consistent distension of distal colonic segments [11]. On the other
hand, early intake of oral contrast might increase colonic distension, without the
need for a rectal enema.
Common findings on MRI of active inflammation include increased bowel wall
thickness, enhancement of the bowel wall on T1-weighted sequences after
intravenous contrast administration, increased signal intensity on fat saturated
T2-weighted sequences and extramural findings, such as the comb sign, creeping
Proefschrift2018-new2.indb 9 13/9/18 10:06
Chapter 1
10
fat and fistulizing disease [12]. These can be graded using predefined criteria,
although uniform definitions are lacking in current literature. Additional quantitative
measurements can be made, such as bowel wall thickness and length measurements,
or relative contrast enhancement (RCE) [11].
A number of limitations are associated with technical aspects of MRI and the
evaluation Crohn’s disease activity. The use of intravenous contrast has become
a standard part in MRI evaluation of Crohn’s disease, but several serious adverse
effects have been reported. Nephrogenic systemic fibrosis (NSF) is known to be
caused by some gadolinium chelated contrast media, occurring in about 1–7% of
patients with severely impaired renal function [13]. Newer contrast agents have
not yet shown clear association with NSF, although these should be thoroughly
investigated. Furthermore, long-term cerebral depositions after gadolinium use
have raised concerns about the safety of these contrast agents [14].
Several studies have reported varying reproducibility of MRI features, due to the
subjective interpretation of many features, lack of uniform definitions and the
complexity of the evaluation. A meta-analysis revealed the difficulty of MRI to
discern mild and remittent disease (both 62% grading accuracy), while moderate-
to-severe disease was graded with high accuracy (91%) [15]. Experience is important
in this respect, as MRI training with 100 cases of Crohn’s disease has been shown
to improve the grading accuracy of inexperienced readers (66% to 75%), with a
significant increase in grading accuracy for categories of mild, moderate and severe
disease activity [16].
MRI activity scores
In recent years, several activity scores for MRI have been proposed in an effort to
increase grading accuracy and reproducibility. The Magnetic Resonance Index of
Activity (MaRIA) was developed using the endoscopic CDEIS as reference standard,
while the Crohn’s Disease MRI Index (CDMI) and the so-called London score were
developed using the endoscopic biopsy Acute Inflammation Score (eAIS) as the
reference standard [11,17]. More recently, the Clermont score was developed as a
modified MaRIA, based on diffusion-weighted imaging (DWI), rather than contrast
enhancement [18]. The mentioned activity scores were developed using regression
techniques to provide a selection of important MRI features to which weightings
Proefschrift2018-new2.indb 10 13/9/18 10:06
Chapter 1
10
fat and fistulizing disease [12]. These can be graded using predefined criteria,
although uniform definitions are lacking in current literature. Additional quantitative
measurements can be made, such as bowel wall thickness and length measurements,
or relative contrast enhancement (RCE) [11].
A number of limitations are associated with technical aspects of MRI and the
evaluation Crohn’s disease activity. The use of intravenous contrast has become
a standard part in MRI evaluation of Crohn’s disease, but several serious adverse
effects have been reported. Nephrogenic systemic fibrosis (NSF) is known to be
caused by some gadolinium chelated contrast media, occurring in about 1–7% of
patients with severely impaired renal function [13]. Newer contrast agents have
not yet shown clear association with NSF, although these should be thoroughly
investigated. Furthermore, long-term cerebral depositions after gadolinium use
have raised concerns about the safety of these contrast agents [14].
Several studies have reported varying reproducibility of MRI features, due to the
subjective interpretation of many features, lack of uniform definitions and the
complexity of the evaluation. A meta-analysis revealed the difficulty of MRI to
discern mild and remittent disease (both 62% grading accuracy), while moderate-
to-severe disease was graded with high accuracy (91%) [15]. Experience is important
in this respect, as MRI training with 100 cases of Crohn’s disease has been shown
to improve the grading accuracy of inexperienced readers (66% to 75%), with a
significant increase in grading accuracy for categories of mild, moderate and severe
disease activity [16].
MRI activity scores
In recent years, several activity scores for MRI have been proposed in an effort to
increase grading accuracy and reproducibility. The Magnetic Resonance Index of
Activity (MaRIA) was developed using the endoscopic CDEIS as reference standard,
while the Crohn’s Disease MRI Index (CDMI) and the so-called London score were
developed using the endoscopic biopsy Acute Inflammation Score (eAIS) as the
reference standard [11,17]. More recently, the Clermont score was developed as a
modified MaRIA, based on diffusion-weighted imaging (DWI), rather than contrast
enhancement [18]. The mentioned activity scores were developed using regression
techniques to provide a selection of important MRI features to which weightings
Proefschrift2018-new2.indb 10 13/9/18 10:06
Introduction
11
are applied. The most often included MRI features in these scores were bowel wall
thickness, mural T2 signal and mural contrast enhancement. All the aforementioned
scores have been externally validated, although comparison in terms of diagnostic
and grading performance has not been done [19–21]. Furthermore, both the CDMI
and MaRIA have shown good reproducibility and moderate agreement to CDEIS
[21]. As these scores show, a small selection of important MRI features may be
sufficient to predict endoscopic disease activity. In the same reasoning, extramural
features – fistula, abscess, comb sign and perimural T2 signal – are less likely to be
included due to their less direct association with the endoscopic reference standard.
Computer-aided evaluation
Limitations in reproducibility of MRI features and activity scores have driven the
search for more reliable techniques to quantify disease activity. Specifically, the use
of specialized software has been proposed to reduce sources of variability, such
as differing levels of reader experience, lack of uniform definition and subjective
interpretation of MRI disease features [22]. Many aspects of conventional evaluation
require subjective and sometimes arbitrary choices to decide on a final grade or
measure. Bowel wall thickness and contrast enhancement are commonly graded
based on their maximum appearance. However, it is not known whether overall
estimates of bowel wall thickness and enhancement might provide more useful
information.
Recently, a method was developed for semiautomatic delineation of the bowel wall,
allowing for automatic wall thickness measurements [23]. These have shown to
increase agreement between multiple observers, when compared to conventional
measurements done by radiologists. Furthermore, these developments open up
possibilities for new types of measurements, such as the mean bowel wall thickness
or volume of a complete bowel segment. Simultaneously, advances have been made
in image registration (i.e. the alignment of temporally acquired image sequences)
for dynamic contrast enhancement (DCE) imaging [24]. The registration process
is essential for obtaining accurate measurements from DCE, and its improvement
has led to renewed interest into parameters, which were known to correlate with
Crohn’s disease activity, such as the initial slope of increase of the enhancement
curve. Integration of semiautomatic measurements in MRI activity scores could
potentially improve agreement by reducing subjective interpretation.
Proefschrift2018-new2.indb 11 13/9/18 10:06
Introduction
11
are applied. The most often included MRI features in these scores were bowel wall
thickness, mural T2 signal and mural contrast enhancement. All the aforementioned
scores have been externally validated, although comparison in terms of diagnostic
and grading performance has not been done [19–21]. Furthermore, both the CDMI
and MaRIA have shown good reproducibility and moderate agreement to CDEIS
[21]. As these scores show, a small selection of important MRI features may be
sufficient to predict endoscopic disease activity. In the same reasoning, extramural
features – fistula, abscess, comb sign and perimural T2 signal – are less likely to be
included due to their less direct association with the endoscopic reference standard.
Computer-aided evaluation
Limitations in reproducibility of MRI features and activity scores have driven the
search for more reliable techniques to quantify disease activity. Specifically, the use
of specialized software has been proposed to reduce sources of variability, such
as differing levels of reader experience, lack of uniform definition and subjective
interpretation of MRI disease features [22]. Many aspects of conventional evaluation
require subjective and sometimes arbitrary choices to decide on a final grade or
measure. Bowel wall thickness and contrast enhancement are commonly graded
based on their maximum appearance. However, it is not known whether overall
estimates of bowel wall thickness and enhancement might provide more useful
information.
Recently, a method was developed for semiautomatic delineation of the bowel wall,
allowing for automatic wall thickness measurements [23]. These have shown to
increase agreement between multiple observers, when compared to conventional
measurements done by radiologists. Furthermore, these developments open up
possibilities for new types of measurements, such as the mean bowel wall thickness
or volume of a complete bowel segment. Simultaneously, advances have been made
in image registration (i.e. the alignment of temporally acquired image sequences)
for dynamic contrast enhancement (DCE) imaging [24]. The registration process
is essential for obtaining accurate measurements from DCE, and its improvement
has led to renewed interest into parameters, which were known to correlate with
Crohn’s disease activity, such as the initial slope of increase of the enhancement
curve. Integration of semiautomatic measurements in MRI activity scores could
potentially improve agreement by reducing subjective interpretation.
Proefschrift2018-new2.indb 11 13/9/18 10:06
Chapter 1
12
New types of imaging sequences
Diffusion-weighted imaging (DWI) is a form of MRI, which exploits the random motion
of water molecules to obtain image contrast. Recent studies have investigated the
role of DWI as an addition or replacement to conventional MRI sequences for Crohn’s
disease evaluation [25]. The possibility to replace contrast enhanced sequences is
appealing, given the risk of adverse effects, such as neproghenic system fibrosis
and cerebral depositions, associated with gadolinum chelated contrast media [13].
In a comparative study with 44 Crohn’s disease patients, the addition of DWI to a
conventional MR enterography protocol showed no additional benefit in terms of
diagnostic accuracy. Although an increased sensitivity (mostly for mild disease)
was seen, the addition of DWI led to a decrease in specificity [25]. However, a
subsequent study found that DW-MRI without contrast-enhanced sequences was
non-inferior to contrast enhanced MRI and could be used as a safer alternative [26].
Conventional MRI evaluation of Crohn’s disease relied on visualization of bowel
structure, rather than its function. Using fast MRI acquisition, new motility
sequences can be made to show bowel peristalsis using high temporal resolution.
Previous studies have provided evidence for reduced segmental motility in affected
bowel segments and an inverse correlation between motility and histopathologic
inflammation [27,28]. However, this evidence requires validation in larger prospective
studies.
Proefschrift2018-new2.indb 12 13/9/18 10:06
Chapter 1
12
New types of imaging sequences
Diffusion-weighted imaging (DWI) is a form of MRI, which exploits the random motion
of water molecules to obtain image contrast. Recent studies have investigated the
role of DWI as an addition or replacement to conventional MRI sequences for Crohn’s
disease evaluation [25]. The possibility to replace contrast enhanced sequences is
appealing, given the risk of adverse effects, such as neproghenic system fibrosis
and cerebral depositions, associated with gadolinum chelated contrast media [13].
In a comparative study with 44 Crohn’s disease patients, the addition of DWI to a
conventional MR enterography protocol showed no additional benefit in terms of
diagnostic accuracy. Although an increased sensitivity (mostly for mild disease)
was seen, the addition of DWI led to a decrease in specificity [25]. However, a
subsequent study found that DW-MRI without contrast-enhanced sequences was
non-inferior to contrast enhanced MRI and could be used as a safer alternative [26].
Conventional MRI evaluation of Crohn’s disease relied on visualization of bowel
structure, rather than its function. Using fast MRI acquisition, new motility
sequences can be made to show bowel peristalsis using high temporal resolution.
Previous studies have provided evidence for reduced segmental motility in affected
bowel segments and an inverse correlation between motility and histopathologic
inflammation [27,28]. However, this evidence requires validation in larger prospective
studies.
Proefschrift2018-new2.indb 12 13/9/18 10:06
Introduction
13
OUTLINE OF THIS THESIS
This thesis studies the accuracy and reproducibility for grading of Crohn’s disease
activity using MRI.
Numerous studies have been performed using different cross-sectional imaging
techniques to grade disease activity in Crohn’s disease patients. However, only a
few studies have directly compared two different imaging techniques. In chapter 2,
we performed a meta-analysis of studies evaluating the grading of Crohn’s disease
activity using CT, MRI, US or scintigraphy.
Activity scores based on MRI disease features of Crohn’s disease are increasingly
used and investigated for the purpose of monitoring disease activity and therapy
evaluation. A comparison of the performance of several available scores in terms of
accuracy and interobserver agreement is presented in chapter 3.
MRI features and activity scores used for Crohn’s disease evaluation often have
limited interobserver agreement, due to the subjectivity in scoring of features.
In chapter 4, an MRI activity score is developed and validated, using new semi-
automatic measurements, such as bowel wall volume and dynamic contrast
enhancement, which are based on simple input from the radiologist.
New imaging techniques such as diffusion-weighted imaging (DWI) or motility
imaging have shown value in detection of disease activity in Crohn’s disease
patients. In chapter 5, we compare several imaging protocols using either DWI or
contrast-enhanced sequences, or a combination of both. In chapter 6, a study is
described evaluating parameters of motility against an endoscopic and histologic
reference standard.
Evaluation of Crohn’s disease on MRI requires considerable experience of readers,
due the complexity and diversity of the disease. Training in grading Crohn’s disease
on MRI can be effectively used to increase reader’s accuracy. Chapter 7 describes a
follow-up study wherein readers’ long-term performance is evaluated.
The thesis is concluded by a general discussion of the conclusions and implications
in chapter 8 and a summary in English and Dutch in chapters 9 and 10.
Proefschrift2018-new2.indb 13 13/9/18 10:06
Introduction
13
OUTLINE OF THIS THESIS
This thesis studies the accuracy and reproducibility for grading of Crohn’s disease
activity using MRI.
Numerous studies have been performed using different cross-sectional imaging
techniques to grade disease activity in Crohn’s disease patients. However, only a
few studies have directly compared two different imaging techniques. In chapter 2,
we performed a meta-analysis of studies evaluating the grading of Crohn’s disease
activity using CT, MRI, US or scintigraphy.
Activity scores based on MRI disease features of Crohn’s disease are increasingly
used and investigated for the purpose of monitoring disease activity and therapy
evaluation. A comparison of the performance of several available scores in terms of
accuracy and interobserver agreement is presented in chapter 3.
MRI features and activity scores used for Crohn’s disease evaluation often have
limited interobserver agreement, due to the subjectivity in scoring of features.
In chapter 4, an MRI activity score is developed and validated, using new semi-
automatic measurements, such as bowel wall volume and dynamic contrast
enhancement, which are based on simple input from the radiologist.
New imaging techniques such as diffusion-weighted imaging (DWI) or motility
imaging have shown value in detection of disease activity in Crohn’s disease
patients. In chapter 5, we compare several imaging protocols using either DWI or
contrast-enhanced sequences, or a combination of both. In chapter 6, a study is
described evaluating parameters of motility against an endoscopic and histologic
reference standard.
Evaluation of Crohn’s disease on MRI requires considerable experience of readers,
due the complexity and diversity of the disease. Training in grading Crohn’s disease
on MRI can be effectively used to increase reader’s accuracy. Chapter 7 describes a
follow-up study wherein readers’ long-term performance is evaluated.
The thesis is concluded by a general discussion of the conclusions and implications
in chapter 8 and a summary in English and Dutch in chapters 9 and 10.
Proefschrift2018-new2.indb 13 13/9/18 10:06
Chapter 1
14
REFERENCES 1. Loftus EV. Clinical epidemiology of inflammatory bowel disease: Incidence, prevalence,
and environmental influences. Gastroenterology. 2004;126:1504–17.
2. Thia KT, Sandborn WJ, Harmsen WS, et al. Risk factors associated with progression to
intestinal complications of Crohn’s disease in a population-based cohort. Gastroenterology
2010;139:1147–55.
3. Baumgart DC, Sandborn WJ. Crohn’s disease. 2012;6736:1–16.
4. Bernell O, Lapidus A, Hellers G. Risk factors for surgery and postoperative recurrence in
Crohn’s disease. Ann Surg 2000;231:38–45.
5. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Thérapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut 1989;30:983–9.
6. Daperno M, D’Haens G, Van Assche G, et al. Development and validation of a new,
simplified endoscopic activity score for Crohn’s disease: The SES-CD. Gastrointest Endosc
2004;60:505–12.
7. Mazzuoli S, Guglielmi FW, Antonelli E, et al. Definition and evaluation of mucosal healing
in clinical practice. Dig Liver Dis 2013;45:969–77.
8. Horsthuis K, Bipat S, Bennink RJ, et al. Inflammatory bowel disease diagnosed with US,
MR, scintigraphy, and CT: meta-analysis of prospective studies. Radiology 2008;247:64–
79.
9. Panes J, Bouhnik Y, Reinisch W, et al. Imaging techniques for assessment of inflammatory
bowel disease: Joint ECCO and ESGAR evidence-based consensus guidelines. J Crohn’s
Colitis 2013;7:556–85.
10. Negaard A, Paulsen V, Sandvik L, et al. A prospective randomized comparison between
two MRI studies of the small bowel in Crohn’s disease, the oral contrast method and MR
enteroclysis. Eur Radiol 2007;17:2294–301.
11. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment of
disease activity and severity in ileocolonic Crohn’s disease. Gut 2009;58:1113–20.
12. Maccioni F, Bruni A, Viscido A, et al. MR imaging in patients with Crohn disease: value
of T2- versus T1-weighted gadolinium-enhanced MR sequences with use of an oral
superparamagnetic contrast agent. Radiology 2006;238:517–30.
13. ACR Committee on drugs and contrast Media. ACR Manual on Contrast Media. 2017.
14. McDonald RJ, McDonald JS, Kallmes DF, et al. Intracranial Gadolinium Deposition after
Contrast-enhanced MR Imaging. Radiology 2015;0:150025.
15. Horsthuis K, Bipat S, Stokkers PCF, et al. Magnetic resonance imaging for evaluation of
disease activity in Crohn’s disease: a systematic review. Eur Radiol 2009;19:1450–60.
16. Tielbeek JAW, Bipat S, Boellaard TN, et al. Training readers to improve their accuracy in
grading Crohn’s disease activity on MRI. Eur Radiol 2014;:1059–67.
17. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: Derivation and histopathological validation of an MR-
based activity index. Eur J Radiol 2012;81:2080–8.
18. Buisson A, Joubert A, Montoriol P-F, et al. Diffusion-weighted magnetic resonance imaging
for detecting and assessing ileal inflammation in Crohn’s disease. Aliment Pharmacol Ther
2013;37:537–45.
19. Rimola J, Ordás I, Rodriguez S, et al. Magnetic resonance imaging for evaluation of Crohn’s
disease: Validation of parameters of severity and quantitative index of activity. Inflamm
Proefschrift2018-new2.indb 14 13/9/18 10:06
Chapter 1
14
REFERENCES 1. Loftus EV. Clinical epidemiology of inflammatory bowel disease: Incidence, prevalence,
and environmental influences. Gastroenterology. 2004;126:1504–17.
2. Thia KT, Sandborn WJ, Harmsen WS, et al. Risk factors associated with progression to
intestinal complications of Crohn’s disease in a population-based cohort. Gastroenterology
2010;139:1147–55.
3. Baumgart DC, Sandborn WJ. Crohn’s disease. 2012;6736:1–16.
4. Bernell O, Lapidus A, Hellers G. Risk factors for surgery and postoperative recurrence in
Crohn’s disease. Ann Surg 2000;231:38–45.
5. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Thérapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut 1989;30:983–9.
6. Daperno M, D’Haens G, Van Assche G, et al. Development and validation of a new,
simplified endoscopic activity score for Crohn’s disease: The SES-CD. Gastrointest Endosc
2004;60:505–12.
7. Mazzuoli S, Guglielmi FW, Antonelli E, et al. Definition and evaluation of mucosal healing
in clinical practice. Dig Liver Dis 2013;45:969–77.
8. Horsthuis K, Bipat S, Bennink RJ, et al. Inflammatory bowel disease diagnosed with US,
MR, scintigraphy, and CT: meta-analysis of prospective studies. Radiology 2008;247:64–
79.
9. Panes J, Bouhnik Y, Reinisch W, et al. Imaging techniques for assessment of inflammatory
bowel disease: Joint ECCO and ESGAR evidence-based consensus guidelines. J Crohn’s
Colitis 2013;7:556–85.
10. Negaard A, Paulsen V, Sandvik L, et al. A prospective randomized comparison between
two MRI studies of the small bowel in Crohn’s disease, the oral contrast method and MR
enteroclysis. Eur Radiol 2007;17:2294–301.
11. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment of
disease activity and severity in ileocolonic Crohn’s disease. Gut 2009;58:1113–20.
12. Maccioni F, Bruni A, Viscido A, et al. MR imaging in patients with Crohn disease: value
of T2- versus T1-weighted gadolinium-enhanced MR sequences with use of an oral
superparamagnetic contrast agent. Radiology 2006;238:517–30.
13. ACR Committee on drugs and contrast Media. ACR Manual on Contrast Media. 2017.
14. McDonald RJ, McDonald JS, Kallmes DF, et al. Intracranial Gadolinium Deposition after
Contrast-enhanced MR Imaging. Radiology 2015;0:150025.
15. Horsthuis K, Bipat S, Stokkers PCF, et al. Magnetic resonance imaging for evaluation of
disease activity in Crohn’s disease: a systematic review. Eur Radiol 2009;19:1450–60.
16. Tielbeek JAW, Bipat S, Boellaard TN, et al. Training readers to improve their accuracy in
grading Crohn’s disease activity on MRI. Eur Radiol 2014;:1059–67.
17. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: Derivation and histopathological validation of an MR-
based activity index. Eur J Radiol 2012;81:2080–8.
18. Buisson A, Joubert A, Montoriol P-F, et al. Diffusion-weighted magnetic resonance imaging
for detecting and assessing ileal inflammation in Crohn’s disease. Aliment Pharmacol Ther
2013;37:537–45.
19. Rimola J, Ordás I, Rodriguez S, et al. Magnetic resonance imaging for evaluation of Crohn’s
disease: Validation of parameters of severity and quantitative index of activity. Inflamm
Proefschrift2018-new2.indb 14 13/9/18 10:06
Introduction
15
Bowel Dis 2011;17:1759–68.
20. Hordonneau C, Buisson A, Scanzi J, et al. Diffusion-weighted magnetic resonance
imaging in ileocolonic Crohn’s disease: validation of quantitative index of activity. Am J
Gastroenterol 2014;109:89–98.
21. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading crohn disease activity with MRI:
Interobserver variability of MRI features, MRI scoring of severity, and correlation with
Crohn disease endoscopic index of severity. Am J Roentgenol 2013;201:1220–8.
22. Tielbeek JAW, Vos FM, Stoker J. A computer-assisted model for detection of MRI signs of
Crohn’s disease activity: Future or fiction? Abdom Imaging 2012;37:967–73.
23. Naziroglu RE, Puylaert CAJ, Tielbeek JAW, et al. Semi-automatic bowel wall thickness
measurements on MR enterography in patients with Crohn’s disease. Br J Radiol 2017;90.
24. Li Z, Tielbeek JAW, Caan MWA, et al. Expiration-Phase Template-Based Motion Correction
of Free-Breathing Abdominal Dynamic Contrast Enhanced MRI. IEEE Trans Biomed Eng
2015;62:1215–25.
25. Kim K-J, Lee Y, Park SH, et al. Diffusion-weighted MR enterography for evaluating Crohn’s
disease: how does it add diagnostically to conventional MR enterography? Inflamm Bowel
Dis 2015;21:101–9.
26. Seo N, Park SH, Kim K-J, et al. MR Enterography for the Evaluation of Small-Bowel
Inflammation in Crohn Disease by Using Diffusion-weighted Imaging without Intravenous
Contrast Material: A Prospective Noninferiority Study. Radiology 2016;278:762–72.
27. Menys A, Atkinson D, Odille F, et al. Quantified terminal ileal motility during MR
enterography as a potential biomarker of Crohn’s disease activity: A preliminary study.
Eur Radiol 2012;22:2494–501.
28. Cullmann JL, Bickelhaupt S, Froehlich JM, et al. MR imaging in Crohn’s disease: Correlation
of MR motility measurement with histopathology in the terminal ileum. Neurogastroenterol
Motil 2013;25.
Proefschrift2018-new2.indb 15 13/9/18 10:06
Introduction
15
Bowel Dis 2011;17:1759–68.
20. Hordonneau C, Buisson A, Scanzi J, et al. Diffusion-weighted magnetic resonance
imaging in ileocolonic Crohn’s disease: validation of quantitative index of activity. Am J
Gastroenterol 2014;109:89–98.
21. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading crohn disease activity with MRI:
Interobserver variability of MRI features, MRI scoring of severity, and correlation with
Crohn disease endoscopic index of severity. Am J Roentgenol 2013;201:1220–8.
22. Tielbeek JAW, Vos FM, Stoker J. A computer-assisted model for detection of MRI signs of
Crohn’s disease activity: Future or fiction? Abdom Imaging 2012;37:967–73.
23. Naziroglu RE, Puylaert CAJ, Tielbeek JAW, et al. Semi-automatic bowel wall thickness
measurements on MR enterography in patients with Crohn’s disease. Br J Radiol 2017;90.
24. Li Z, Tielbeek JAW, Caan MWA, et al. Expiration-Phase Template-Based Motion Correction
of Free-Breathing Abdominal Dynamic Contrast Enhanced MRI. IEEE Trans Biomed Eng
2015;62:1215–25.
25. Kim K-J, Lee Y, Park SH, et al. Diffusion-weighted MR enterography for evaluating Crohn’s
disease: how does it add diagnostically to conventional MR enterography? Inflamm Bowel
Dis 2015;21:101–9.
26. Seo N, Park SH, Kim K-J, et al. MR Enterography for the Evaluation of Small-Bowel
Inflammation in Crohn Disease by Using Diffusion-weighted Imaging without Intravenous
Contrast Material: A Prospective Noninferiority Study. Radiology 2016;278:762–72.
27. Menys A, Atkinson D, Odille F, et al. Quantified terminal ileal motility during MR
enterography as a potential biomarker of Crohn’s disease activity: A preliminary study.
Eur Radiol 2012;22:2494–501.
28. Cullmann JL, Bickelhaupt S, Froehlich JM, et al. MR imaging in Crohn’s disease: Correlation
of MR motility measurement with histopathology in the terminal ileum. Neurogastroenterol
Motil 2013;25.
Proefschrift2018-new2.indb 15 13/9/18 10:06
Proefschrift2018-new2.indb 16 13/9/18 10:06 Proefschrift2018-new2.indb 16 13/9/18 10:06
CHAPTER 2
Grading of Crohn’s disease activity using CT, MRI, US and scintigraphy: a meta-analysis
Carl A.J. Puylaert, Jeroen A.W. Tielbeek, Shandra Bipat, Jaap Stoker
Proefschrift2018-new2.indb 17 13/9/18 10:06
CHAPTER 2
Grading of Crohn’s disease activity using CT, MRI, US and scintigraphy: a meta-analysis
Carl A.J. Puylaert, Jeroen A.W. Tielbeek, Shandra Bipat, Jaap Stoker
Proefschrift2018-new2.indb 17 13/9/18 10:06
18
Chapter 2
ABSTRACT
Objectives
To assess the grading of Crohn’s disease activity using CT, MRI, US and scintigraphy.
Materials and Methods
MEDLINE, EMBASE and Cochrane databases were searched (January 1983–March
2014) for studies evaluating CT, MRI, US and scintigraphy in grading Crohn’s
disease activity compared to endoscopy, biopsies or intraoperative findings. Two
independent reviewers assessed the data. Three-by-three tables (none, mild, frank
disease) were constructed for all studies, and estimates of accurate, over- and
under-grading were calculated/summarized by fixed or random effects models.
Results
Our search yielded 9356 articles, 19 of which were included. Per-patient data
showed accurate grading values for CT, MRI, US and scintigraphy of 86% (95% CI:
75–93%), 84% (95% CI: 67–93%), 44% (95% CI: 28–61%) and 40% (95% CI: 16–70%),
respectively. In the per-patient analysis, CT and MRI showed similar accurate grading
estimates (P=0.8). Per-segment data showed accurate grading values for CT and
scintigraphy of 87% (95% CI: 77–93%) and 86% (95% CI: 80–91%), respectively. MRI
and US showed grading accuracies of 67–82% and 56–75%, respectively.
Conclusions
CT and MRI showed comparable high accurate grading estimates in the per-patient
analysis. Results for US and scintigraphy were inconsistent, and limited data were
available.
Proefschrift2018-new2.indb 18 13/9/18 10:06
18
Chapter 2
ABSTRACT
Objectives
To assess the grading of Crohn’s disease activity using CT, MRI, US and scintigraphy.
Materials and Methods
MEDLINE, EMBASE and Cochrane databases were searched (January 1983–March
2014) for studies evaluating CT, MRI, US and scintigraphy in grading Crohn’s
disease activity compared to endoscopy, biopsies or intraoperative findings. Two
independent reviewers assessed the data. Three-by-three tables (none, mild, frank
disease) were constructed for all studies, and estimates of accurate, over- and
under-grading were calculated/summarized by fixed or random effects models.
Results
Our search yielded 9356 articles, 19 of which were included. Per-patient data
showed accurate grading values for CT, MRI, US and scintigraphy of 86% (95% CI:
75–93%), 84% (95% CI: 67–93%), 44% (95% CI: 28–61%) and 40% (95% CI: 16–70%),
respectively. In the per-patient analysis, CT and MRI showed similar accurate grading
estimates (P=0.8). Per-segment data showed accurate grading values for CT and
scintigraphy of 87% (95% CI: 77–93%) and 86% (95% CI: 80–91%), respectively. MRI
and US showed grading accuracies of 67–82% and 56–75%, respectively.
Conclusions
CT and MRI showed comparable high accurate grading estimates in the per-patient
analysis. Results for US and scintigraphy were inconsistent, and limited data were
available.
Proefschrift2018-new2.indb 18 13/9/18 10:06
19
Grading CD using CT, MRI, US and scintigraphy
INTRODUCTION
Cross-sectional imaging techniques are widely used for diagnosis and evaluation of
Crohn’s disease. Numerous studies have evaluated the diagnostic accuracy of cross-
sectional imaging techniques in patients with Crohn’s disease, and a meta-analysis
was published that investigated the diagnostic accuracy of computed tomography
(CT), magnetic resonance imaging (MRI), ultrasound (US) and scintigraphy [1].
However, clinical monitoring and choice of therapy largely rely on grading of disease
activity.
Clinical symptoms and inflammatory lesions can exist independently, so assessment
of the bowel is essential in guiding therapy decisions [2]. If inflammation is
present, it is important to distinguish between mild, moderate and severe disease,
as medical management differs among these stages [3]. Ileocolonoscopy, the
current reference standard for luminal Crohn’s disease, is accurate for assessing
mucosal abnormalities, but it has several drawbacks, as it is an invasive technique,
is associated with the risk of bowel perforation, is incapable of assessing trans- and
extraluminal disease, and is limited to the colon and terminal ileum [4]. Video capsule
endoscopy (VCE) is a well-tolerated and accurate alternative to ileocolonoscopy
that allows assessment of the whole gastrointestinal tract, although it has shown
lower specificity and bears the risk of capsule retention, which occurs in up to 13%
of patients with Crohn’s disease [5].
Cross-sectional imaging techniques that could accurately grade disease severity
would be preferable to ileocolonoscopy, as they are non-invasive and not limited to
the colon and terminal ileum. Several studies have looked at the use of cross-sectional
imaging for assessing the severity of Crohn’s disease, but offered no comparison
between imaging techniques, as no meta-analysis was performed [2, 6]. To our
knowledge, only one such meta-analysis has been performed, but it evaluated only
MRI and used a search period that ended in April 2007 [7]. This study showed
that MRI correctly graded disease activity in 91% of patients with frank (moderate-
to-severe) disease. However, correct grading was limited in patients with disease
in remission and with mild disease (62% for both). Furthermore, no comparison
with other imaging techniques was made and numerous articles on the grading of
Crohn’s disease using MRI have been published since 2007.
Proefschrift2018-new2.indb 19 13/9/18 10:06
19
Grading CD using CT, MRI, US and scintigraphy
INTRODUCTION
Cross-sectional imaging techniques are widely used for diagnosis and evaluation of
Crohn’s disease. Numerous studies have evaluated the diagnostic accuracy of cross-
sectional imaging techniques in patients with Crohn’s disease, and a meta-analysis
was published that investigated the diagnostic accuracy of computed tomography
(CT), magnetic resonance imaging (MRI), ultrasound (US) and scintigraphy [1].
However, clinical monitoring and choice of therapy largely rely on grading of disease
activity.
Clinical symptoms and inflammatory lesions can exist independently, so assessment
of the bowel is essential in guiding therapy decisions [2]. If inflammation is
present, it is important to distinguish between mild, moderate and severe disease,
as medical management differs among these stages [3]. Ileocolonoscopy, the
current reference standard for luminal Crohn’s disease, is accurate for assessing
mucosal abnormalities, but it has several drawbacks, as it is an invasive technique,
is associated with the risk of bowel perforation, is incapable of assessing trans- and
extraluminal disease, and is limited to the colon and terminal ileum [4]. Video capsule
endoscopy (VCE) is a well-tolerated and accurate alternative to ileocolonoscopy
that allows assessment of the whole gastrointestinal tract, although it has shown
lower specificity and bears the risk of capsule retention, which occurs in up to 13%
of patients with Crohn’s disease [5].
Cross-sectional imaging techniques that could accurately grade disease severity
would be preferable to ileocolonoscopy, as they are non-invasive and not limited to
the colon and terminal ileum. Several studies have looked at the use of cross-sectional
imaging for assessing the severity of Crohn’s disease, but offered no comparison
between imaging techniques, as no meta-analysis was performed [2, 6]. To our
knowledge, only one such meta-analysis has been performed, but it evaluated only
MRI and used a search period that ended in April 2007 [7]. This study showed
that MRI correctly graded disease activity in 91% of patients with frank (moderate-
to-severe) disease. However, correct grading was limited in patients with disease
in remission and with mild disease (62% for both). Furthermore, no comparison
with other imaging techniques was made and numerous articles on the grading of
Crohn’s disease using MRI have been published since 2007.
Proefschrift2018-new2.indb 19 13/9/18 10:06
20
Chapter 2
Our purpose was to systematically review and compare the accuracy of CT, MRI,
US, scintigraphy and positron emission tomography–computed tomography (PET-
CT) in grading Crohn’s disease activity on a per-patient or per-segment basis as
compared to endoscopy, biopsies or intraoperative findings by performing a meta-
analysis. Furthermore, we aimed to investigate the degree of over- and under-
grading for these imaging techniques.
MATERIALS AND METHODS
This review was conducted according to the Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA) guidelines [8]. The review protocol was not
published or registered in advance.
Literature search and strategy
We performed an electronic search in MEDLINE, EMBASE and Cochrane databases
for studies examining the accuracy of CT, MRI, US, scintigraphy and PET (-CT) for
grading Crohn’s disease activity in human subjects. Search terms ‘Crohn’s disease’
and ‘inflammatory bowel disease’ were combined using ‘OR’ and search terms
for imaging modalities were combined using ‘OR’ as well. These two groups were
combined using ‘AND’. The search period was limited from January 1983 to March
2014. Details of the search strategy are provided in the electronic supplementary
material (Appendix E1).
Study selection on title and abstract
All articles retrieved from the electronic search were assessed by one observer (CP).
Non-relevant articles and articles in the form of a review, case report, comment
or letter were excluded. Subsequently, the remaining titles and abstracts were
independently assessed by two observers (CP, JT) to identify potentially eligible
articles. In cases of uncertainty, articles were deemed potentially eligible and
retrieved as full text.
Study selection on full text
The full texts of the remaining articles were retrieved. Two observers (CP, JT)
Proefschrift2018-new2.indb 20 13/9/18 10:06
20
Chapter 2
Our purpose was to systematically review and compare the accuracy of CT, MRI,
US, scintigraphy and positron emission tomography–computed tomography (PET-
CT) in grading Crohn’s disease activity on a per-patient or per-segment basis as
compared to endoscopy, biopsies or intraoperative findings by performing a meta-
analysis. Furthermore, we aimed to investigate the degree of over- and under-
grading for these imaging techniques.
MATERIALS AND METHODS
This review was conducted according to the Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA) guidelines [8]. The review protocol was not
published or registered in advance.
Literature search and strategy
We performed an electronic search in MEDLINE, EMBASE and Cochrane databases
for studies examining the accuracy of CT, MRI, US, scintigraphy and PET (-CT) for
grading Crohn’s disease activity in human subjects. Search terms ‘Crohn’s disease’
and ‘inflammatory bowel disease’ were combined using ‘OR’ and search terms
for imaging modalities were combined using ‘OR’ as well. These two groups were
combined using ‘AND’. The search period was limited from January 1983 to March
2014. Details of the search strategy are provided in the electronic supplementary
material (Appendix E1).
Study selection on title and abstract
All articles retrieved from the electronic search were assessed by one observer (CP).
Non-relevant articles and articles in the form of a review, case report, comment
or letter were excluded. Subsequently, the remaining titles and abstracts were
independently assessed by two observers (CP, JT) to identify potentially eligible
articles. In cases of uncertainty, articles were deemed potentially eligible and
retrieved as full text.
Study selection on full text
The full texts of the remaining articles were retrieved. Two observers (CP, JT)
Proefschrift2018-new2.indb 20 13/9/18 10:06
21
Grading CD using CT, MRI, US and scintigraphy
independently reviewed all eligible articles for the following inclusion criteria: (a)
ten or more patients were included (fewer were considered case-series); (b) CT,
MRI, US, scintigraphy or PET (-CT) was used to grade Crohn’s disease activity; (c)
patients with clinically suspected inflammatory bowel disease (IBD) or known IBD/
Crohn’s disease were included; (d) endoscopy, biopsies or intraoperative findings
were used as a reference test; (e) imaging features used for grading disease activity
were defined; (f) raw data were available to construct 3 x 3 tables; (g) articles were
written in English, Italian, Spanish, French, German or Dutch; and (h) patients with
Crohn’s disease could be analysed separately from other IBD patients. No patient
age limits were applied. Articles in the form of a review, case report, conference
abstract, comment or letter were excluded. In the case of duplicate publications, we
excluded the studies with the lower number of patients. Disagreement regarding
potential eligibility and inclusion was resolved by consensus. The observers were
not blinded to author and journal names.
Table 1. Methodological characteristics from the QUADAS tool and their corresponding
signalling questions [9, 10]. The risk of bias is determined for every domain using the signalling
questions.
Modified QUADAS Methodological Characteristics
Domain Signalling questions (yes, no, unclear)
Patient selection
Was a consecutive or random sample of patients enrolled? Was a case-control design avoided? Did the study avoid inappropriate exclusions?
Index test Were the index test results interpreted without knowledge of the results of the reference test? If a threshold was used, was it pre-specified?Was the execution and interpretation (expertise, image analysis) of the index test described in sufficient detail to permit its replication?a
Reference test Is the reference test likely to correctly classify the target condition? Was the reference test results interpreted without knowledge of the results of the index test?Was the execution and interpretation of the reference test described in sufficient detail to permit its replication? a
Patient flow Was there an appropriate interval between index test(s) and reference test (>1 month)? Did all patients receive a reference test? Did all patients receive the same reference test? Were all patients included in the analysis?
Prospective / Retrospectiveb
Was the data collected after the research question was defined?
a. This signalling question was added from QUADAS 1 [10]. We considered this question essential for quality assessment, while it is not part of QUADAS 2 [9].b. This item was not part of the QUADAS tool
Proefschrift2018-new2.indb 21 13/9/18 10:06
21
Grading CD using CT, MRI, US and scintigraphy
independently reviewed all eligible articles for the following inclusion criteria: (a)
ten or more patients were included (fewer were considered case-series); (b) CT,
MRI, US, scintigraphy or PET (-CT) was used to grade Crohn’s disease activity; (c)
patients with clinically suspected inflammatory bowel disease (IBD) or known IBD/
Crohn’s disease were included; (d) endoscopy, biopsies or intraoperative findings
were used as a reference test; (e) imaging features used for grading disease activity
were defined; (f) raw data were available to construct 3 x 3 tables; (g) articles were
written in English, Italian, Spanish, French, German or Dutch; and (h) patients with
Crohn’s disease could be analysed separately from other IBD patients. No patient
age limits were applied. Articles in the form of a review, case report, conference
abstract, comment or letter were excluded. In the case of duplicate publications, we
excluded the studies with the lower number of patients. Disagreement regarding
potential eligibility and inclusion was resolved by consensus. The observers were
not blinded to author and journal names.
Table 1. Methodological characteristics from the QUADAS tool and their corresponding
signalling questions [9, 10]. The risk of bias is determined for every domain using the signalling
questions.
Modified QUADAS Methodological Characteristics
Domain Signalling questions (yes, no, unclear)
Patient selection
Was a consecutive or random sample of patients enrolled? Was a case-control design avoided? Did the study avoid inappropriate exclusions?
Index test Were the index test results interpreted without knowledge of the results of the reference test? If a threshold was used, was it pre-specified?Was the execution and interpretation (expertise, image analysis) of the index test described in sufficient detail to permit its replication?a
Reference test Is the reference test likely to correctly classify the target condition? Was the reference test results interpreted without knowledge of the results of the index test?Was the execution and interpretation of the reference test described in sufficient detail to permit its replication? a
Patient flow Was there an appropriate interval between index test(s) and reference test (>1 month)? Did all patients receive a reference test? Did all patients receive the same reference test? Were all patients included in the analysis?
Prospective / Retrospectiveb
Was the data collected after the research question was defined?
a. This signalling question was added from QUADAS 1 [10]. We considered this question essential for quality assessment, while it is not part of QUADAS 2 [9].b. This item was not part of the QUADAS tool
Proefschrift2018-new2.indb 21 13/9/18 10:06
22
Chapter 2
Study characteristics
Methodological characteristics
Both reviewers extracted study characteristics independently for all included articles
using a standardized form. To assess the quality of the study design, we used a
modified Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS 2) tool [9,
10], as it separately assesses risk of bias in several methodological domains (patient
selection, index test, reference test and patient flow) using a number of signalling
questions (Table 1). Risk of bias for each domain was described as high, low or
unclear. In addition, concerns regarding the applicability of the patient population,
index and reference test to the review question were rated by the observers as high,
low or unclear. Disagreements were resolved by discussion.
Patient characteristics
The following patient characteristics were recorded: number of patients included,
number of patients in the analysis, whether patients were recruited consecutively,
age characteristics, gender ratio, patient spectrum (i.e. known or suspected IBD or
Crohn’s disease) and other selection criteria for patient inclusion.
Imaging characteristics
Imaging characteristics concerning type of equipment and basic specifications
(type of scanner for CT, field strength and coil type for MRI, and transducer type
for US), techniques used for evaluation (sequences for MRI, use of Doppler for US,
labelling target and tracer type for scintigraphy), bowel preparation (fasting and/
or laxatives), use of luminal and/or intravenous contrast medium, timing of post-
contrast scans and use of spasmolytic drugs were extracted.
Reference test
All reference tests (i.e. endoscopy, biopsies or intraoperative findings) used for
analysis were recorded.
Imaging and reference test interpretation
We recorded the following information regarding interpretation of imaging and
reference tests: the interval in days between index and reference tests, bowel
Proefschrift2018-new2.indb 22 13/9/18 10:06
22
Chapter 2
Study characteristics
Methodological characteristics
Both reviewers extracted study characteristics independently for all included articles
using a standardized form. To assess the quality of the study design, we used a
modified Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS 2) tool [9,
10], as it separately assesses risk of bias in several methodological domains (patient
selection, index test, reference test and patient flow) using a number of signalling
questions (Table 1). Risk of bias for each domain was described as high, low or
unclear. In addition, concerns regarding the applicability of the patient population,
index and reference test to the review question were rated by the observers as high,
low or unclear. Disagreements were resolved by discussion.
Patient characteristics
The following patient characteristics were recorded: number of patients included,
number of patients in the analysis, whether patients were recruited consecutively,
age characteristics, gender ratio, patient spectrum (i.e. known or suspected IBD or
Crohn’s disease) and other selection criteria for patient inclusion.
Imaging characteristics
Imaging characteristics concerning type of equipment and basic specifications
(type of scanner for CT, field strength and coil type for MRI, and transducer type
for US), techniques used for evaluation (sequences for MRI, use of Doppler for US,
labelling target and tracer type for scintigraphy), bowel preparation (fasting and/
or laxatives), use of luminal and/or intravenous contrast medium, timing of post-
contrast scans and use of spasmolytic drugs were extracted.
Reference test
All reference tests (i.e. endoscopy, biopsies or intraoperative findings) used for
analysis were recorded.
Imaging and reference test interpretation
We recorded the following information regarding interpretation of imaging and
reference tests: the interval in days between index and reference tests, bowel
Proefschrift2018-new2.indb 22 13/9/18 10:06
23
Grading CD using CT, MRI, US and scintigraphy
segments that were examined, grading criteria used for imaging and reference
tests, imaging features used for evaluation of disease activity, and whether grading
was performed on a per-patient and/or per-bowel-segment basis.
Data extraction
Grading results for imaging and reference tests were extracted with the grading
scales used in individual studies (i.e. three-, four-, or five-grade scales). From this
data, three-by-three contingency tables comparing results from index and reference
tests categorized as none, mild or frank disease were constructed for each study.
These categories did not use predefined criteria, but were formed either by using the
original grading from each study (in the case of a three-grade scale) or by merging
certain grades to form a three-grade scale. If a four-grade scale was used (none,
mild, moderate or severe disease), groups with moderate and severe disease were
merged into frank disease. For five-grade scales, the second and third scales were
grouped into mild disease and the fourth and fifth were grouped into frank disease.
When studies used multiple reference tests, we used intraoperative findings as the
reference standard. In other cases, histological findings from biopsies were preferred
over endoscopic findings. Because the imaging results in these studies were based
on the most severe lesion, we considered histological data from biopsies as more
lesion-specific and better resembling imaging results than endoscopic results.
Publication bias
To study publication bias, we followed the method by Deeks et al., as recommended
in the Cochrane handbook for DTA reviews [11]. We first calculated effective sample
sizes (ESS) for each study. We then performed linear regression analyses if enough
datasets were available in a group (n > 5), with the proportion of accurate grading
per study as the independent variable and 1/√ESS as the dependent variable.
A significant regression coefficient (P < 0.05) was deemed sufficient to indicate
publication bias.
Data analysis
For each study, we constructed three proportions: ‘accurate grading’, defined as
the number of correctly graded patients or segments; ‘under-grading’, defined as
the number of patients or segments on which the index test graded lower than the
reference test; and ‘over-grading’, defined as the number of patients or segments
Proefschrift2018-new2.indb 23 13/9/18 10:06
23
Grading CD using CT, MRI, US and scintigraphy
segments that were examined, grading criteria used for imaging and reference
tests, imaging features used for evaluation of disease activity, and whether grading
was performed on a per-patient and/or per-bowel-segment basis.
Data extraction
Grading results for imaging and reference tests were extracted with the grading
scales used in individual studies (i.e. three-, four-, or five-grade scales). From this
data, three-by-three contingency tables comparing results from index and reference
tests categorized as none, mild or frank disease were constructed for each study.
These categories did not use predefined criteria, but were formed either by using the
original grading from each study (in the case of a three-grade scale) or by merging
certain grades to form a three-grade scale. If a four-grade scale was used (none,
mild, moderate or severe disease), groups with moderate and severe disease were
merged into frank disease. For five-grade scales, the second and third scales were
grouped into mild disease and the fourth and fifth were grouped into frank disease.
When studies used multiple reference tests, we used intraoperative findings as the
reference standard. In other cases, histological findings from biopsies were preferred
over endoscopic findings. Because the imaging results in these studies were based
on the most severe lesion, we considered histological data from biopsies as more
lesion-specific and better resembling imaging results than endoscopic results.
Publication bias
To study publication bias, we followed the method by Deeks et al., as recommended
in the Cochrane handbook for DTA reviews [11]. We first calculated effective sample
sizes (ESS) for each study. We then performed linear regression analyses if enough
datasets were available in a group (n > 5), with the proportion of accurate grading
per study as the independent variable and 1/√ESS as the dependent variable.
A significant regression coefficient (P < 0.05) was deemed sufficient to indicate
publication bias.
Data analysis
For each study, we constructed three proportions: ‘accurate grading’, defined as
the number of correctly graded patients or segments; ‘under-grading’, defined as
the number of patients or segments on which the index test graded lower than the
reference test; and ‘over-grading’, defined as the number of patients or segments
Proefschrift2018-new2.indb 23 13/9/18 10:06
24
Chapter 2
on which the index test graded disease activity higher than the reference test.
Datasets were sorted into groups by type of imaging, which were then subdivided
by target of evaluation (per-patient or per-segment). To quantify heterogeneity we
calculated the I2-statistic for each group. Data were pooled if more than one dataset
was available in a group and the data were not too heterogeneous (I2<75%) [12].
For the pooled data, we calculated mean logit accurate grading and under- and
over-grading values with corresponding standard errors using non-linear fixed or
random effects models based on the Akaike information criterion (AIC) statistic (a
lower AIC value indicates a better fit) [13, 14]. Using anti-logit transformation, we
obtained summary estimates with 95% confidence intervals (95% CI) for accurate
grading and over- and under-grading. In several studies, multiple datasets were
available (i.e. multiple readers). Because we used all datasets for analysis, we
adjusted the correlation between datasets from the same study by adding the same
number for each study in the subject statement of the random effects approach.
Comparison of CT, MRI, US and scintigraphy was performed with Z-tests using the
logit values of the pooled data. For data that was not pooled, we performed logit
transformation using proportion and sample size (n) to enable comparison. To
calculate logit values for proportions of 0 or 100, we added 0.5 to the number of
events [15]. P values less than 0.05 indicated a statistically significant difference. All
data analyses were performed using Excel 2010 (Microsoft Corporation, Redmond,
WA, USA), SPSS 22.0 (IBM SPSS Statistics for Macintosh, Version 22.0; IBM Corp.,
Armonk, NY, USA), and SAS 9.3 (SAS Institute, Cary, NC, USA) software programs.
RESULTS
Search and study selection
The search yielded 9356 articles. After selection on title and/or abstract, 149 articles
remained and were retrieved as full-text articles (Fig. 1). Of these remaining articles,
130 did not fulfill the eligibility criteria (Appendix E2). Nineteen articles met all
inclusion criteria and were included for further data extraction. CT was evaluated
in 3 [16–18], MRI in 11 [19–29], US in 3 [30–32], and scintigraphy in 3 [18, 33, 34]. No
articles evaluating PET-CT were found that met our criteria.
Proefschrift2018-new2.indb 24 13/9/18 10:06
24
Chapter 2
on which the index test graded disease activity higher than the reference test.
Datasets were sorted into groups by type of imaging, which were then subdivided
by target of evaluation (per-patient or per-segment). To quantify heterogeneity we
calculated the I2-statistic for each group. Data were pooled if more than one dataset
was available in a group and the data were not too heterogeneous (I2<75%) [12].
For the pooled data, we calculated mean logit accurate grading and under- and
over-grading values with corresponding standard errors using non-linear fixed or
random effects models based on the Akaike information criterion (AIC) statistic (a
lower AIC value indicates a better fit) [13, 14]. Using anti-logit transformation, we
obtained summary estimates with 95% confidence intervals (95% CI) for accurate
grading and over- and under-grading. In several studies, multiple datasets were
available (i.e. multiple readers). Because we used all datasets for analysis, we
adjusted the correlation between datasets from the same study by adding the same
number for each study in the subject statement of the random effects approach.
Comparison of CT, MRI, US and scintigraphy was performed with Z-tests using the
logit values of the pooled data. For data that was not pooled, we performed logit
transformation using proportion and sample size (n) to enable comparison. To
calculate logit values for proportions of 0 or 100, we added 0.5 to the number of
events [15]. P values less than 0.05 indicated a statistically significant difference. All
data analyses were performed using Excel 2010 (Microsoft Corporation, Redmond,
WA, USA), SPSS 22.0 (IBM SPSS Statistics for Macintosh, Version 22.0; IBM Corp.,
Armonk, NY, USA), and SAS 9.3 (SAS Institute, Cary, NC, USA) software programs.
RESULTS
Search and study selection
The search yielded 9356 articles. After selection on title and/or abstract, 149 articles
remained and were retrieved as full-text articles (Fig. 1). Of these remaining articles,
130 did not fulfill the eligibility criteria (Appendix E2). Nineteen articles met all
inclusion criteria and were included for further data extraction. CT was evaluated
in 3 [16–18], MRI in 11 [19–29], US in 3 [30–32], and scintigraphy in 3 [18, 33, 34]. No
articles evaluating PET-CT were found that met our criteria.
Proefschrift2018-new2.indb 24 13/9/18 10:06
25
Grading CD using CT, MRI, US and scintigraphy
Figure 1. Flow diagram showing study selection.
Study characteristics
Methodological characteristics
Evaluation of the imaging tests was performed blinded from the reference test in
13 studies [17, 18, 21, 22, 24–30, 33, 34]. The reference test was performed blinded
to the imaging results in 12 studies [16, 17, 19, 21, 24, 26–30, 33, 34]. The remaining
studies did not specify whether observers were blinded to other results [20, 23, 31,
32]. Fifteen of the studies included patients prospectively [16–26, 28, 30, 31, 34].
Signalling questions for the QUADAS tool were answered with ‘yes’ in 78.9% of
cases (Fig. 2). Patient selection and index test domains showed less risk of bias than
reference test and patient flow domains. Concern about applicability of patient
selection and index and reference tests was generally low (Fig. 3).
Proefschrift2018-new2.indb 25 13/9/18 10:06
25
Grading CD using CT, MRI, US and scintigraphy
Figure 1. Flow diagram showing study selection.
Study characteristics
Methodological characteristics
Evaluation of the imaging tests was performed blinded from the reference test in
13 studies [17, 18, 21, 22, 24–30, 33, 34]. The reference test was performed blinded
to the imaging results in 12 studies [16, 17, 19, 21, 24, 26–30, 33, 34]. The remaining
studies did not specify whether observers were blinded to other results [20, 23, 31,
32]. Fifteen of the studies included patients prospectively [16–26, 28, 30, 31, 34].
Signalling questions for the QUADAS tool were answered with ‘yes’ in 78.9% of
cases (Fig. 2). Patient selection and index test domains showed less risk of bias than
reference test and patient flow domains. Concern about applicability of patient
selection and index and reference tests was generally low (Fig. 3).
Proefschrift2018-new2.indb 25 13/9/18 10:06
26
Chapter 2
Figure 2. QUADAS signalling questions (Table 1) per domain (from up to down: patient selec-
tion, index test, reference test and patient flow). The last column shows whether studies
included patients prospectively.
Figure 3. QUADAS risk of bias per domain and concerns regarding applicability for domains
of patient selection, index test and reference test.
Patient selectionPatient selection applicability
Index testIndex test applicability
Reference testReference test applicability
Patient �ow
Low risk / concernsHigh risk / concernsUnclear
0 20 40 60 80 100 %
Proefschrift2018-new2.indb 26 13/9/18 10:06
26
Chapter 2
Figure 2. QUADAS signalling questions (Table 1) per domain (from up to down: patient selec-
tion, index test, reference test and patient flow). The last column shows whether studies
included patients prospectively.
Figure 3. QUADAS risk of bias per domain and concerns regarding applicability for domains
of patient selection, index test and reference test.
Patient selectionPatient selection applicability
Index testIndex test applicability
Reference testReference test applicability
Patient �ow
Low risk / concernsHigh risk / concernsUnclear
0 20 40 60 80 100 %
Proefschrift2018-new2.indb 26 13/9/18 10:06
27
Grading CD using CT, MRI, US and scintigraphy
Table 2. Study characteristics
Study
Im-aging mo-dality
No. of Patients includ-ed
No. of patients in analysis
Con-secu-tive
Age, mean (range) or mean±SD
Male/Female Ratio
Patient Spectrum Inclusion Criteria
Mao 2013 [17]
CT 32 32 Y 30 (18–51) b 22:10 Known CD
Suspected recurrence after ileocolic resection
Mohamed 2012 [16]
CT 26 26 ? 43.4 (19–69) 18:8 Known CD
Referred to further assess-ment with CTE
Kolkman 1996 [18]
CT, SG
32 17 ? 36 (17–65) 11:6 Known/suspect-ed IBD
Suspected IBD or IBD exacerbations or suspected abdominal complications
Schill 2013 [29]
MRI 76 76 N 31.5 (16–76) b 40:36 Known CD
Patients scheduled for CD surgery
Gallego 2011 [28]
MRI 61 61 Y 36.1 (14–65) 29:32 Known CD
NA
Koilakou 2010 [27]
MRI 26 26 ? 36.5 (22–69) b
16:13 Known CD
Patients with previous ileocolic resection
Horsthuis 2010 [26]
MRI 33 15 Y 14 (8–17) a, b 15:18 a Suspect-ed IBD
Age 8–18 years
Girometti 2008 [25]
MRI 52 45 Y 42.5 (18–67) 23:29 a Known/suspect-ed CD
Referred for CC with biopsy and MRI for relapse or suspected onset of CD
Horsthuis 2006 [24]
MRI 20 20 Y 36±13 7:13 Known CD
Scheduled for CC
Florie 2005 [21]
MRI 31 31 ? 36±12 22:9 Known CD
Scheduled for ICC because of clinical suspicion of relapsing CD
Shoenut 1994 [20]
MRI 20 12 Y 42.6(20–70) a
12:8a Suspect-ed IBD
NA
Shoenut 1993 [19]
MRI 28 19 Y 34.1(20–58) 17:11a Known IBD
Referred to MRI for evalua-tion and on medical therapy
Schreyer 2005 [22]
MRI 30 30 Y 29 (18–65) 8:22 Known CD
Routine small bowel MRI
Schreyer 2005 [23]
MRI 22 12 Y 33.4 (19–55) 5:7 Known/suspect-ed IBD
NA
Drews 2009 [32]
US 32 32 N 38.8 (17–71) 14:18 Known CD
NA
Neye 2004 [31]
US 22 22 Y 33.7 (16–56) 9:13 Known CD
Referred to gastroenter-ologist
Bozkurt 1996 [30]
US 88 32 ? 39 (16–87) a 48:40 a Suspect-ed IBD
NA
Biancone 1997 [34]
SG 17 10 ? 43±11 a 9:8 a Known CD
Patients 6–12 months after ileocecal resection
Sciarretta 1998 [33]
SG 103 31 ? 38.3 (15–78) a 54:49 a Suspect-ed IBD
NA
CC = Colonoscopy, CD = Crohn’s Disease, CTE = Computed Tomography Enterography, IBD = Inflammatory Bowel Dis-
ease, ICC = Ileocolonoscopy, MRI = Magnetic Resonance Imaging, NA = Not Applicable, Y = Yes, N = No, ? = Unclear
a. These values reflect the total number of included patients in their respective studies, not only the patients used in this
analysis. b. Median (range)
Proefschrift2018-new2.indb 27 13/9/18 10:06
27
Grading CD using CT, MRI, US and scintigraphy
Table 2. Study characteristics
Study
Im-aging mo-dality
No. of Patients includ-ed
No. of patients in analysis
Con-secu-tive
Age, mean (range) or mean±SD
Male/Female Ratio
Patient Spectrum Inclusion Criteria
Mao 2013 [17]
CT 32 32 Y 30 (18–51) b 22:10 Known CD
Suspected recurrence after ileocolic resection
Mohamed 2012 [16]
CT 26 26 ? 43.4 (19–69) 18:8 Known CD
Referred to further assess-ment with CTE
Kolkman 1996 [18]
CT, SG
32 17 ? 36 (17–65) 11:6 Known/suspect-ed IBD
Suspected IBD or IBD exacerbations or suspected abdominal complications
Schill 2013 [29]
MRI 76 76 N 31.5 (16–76) b 40:36 Known CD
Patients scheduled for CD surgery
Gallego 2011 [28]
MRI 61 61 Y 36.1 (14–65) 29:32 Known CD
NA
Koilakou 2010 [27]
MRI 26 26 ? 36.5 (22–69) b
16:13 Known CD
Patients with previous ileocolic resection
Horsthuis 2010 [26]
MRI 33 15 Y 14 (8–17) a, b 15:18 a Suspect-ed IBD
Age 8–18 years
Girometti 2008 [25]
MRI 52 45 Y 42.5 (18–67) 23:29 a Known/suspect-ed CD
Referred for CC with biopsy and MRI for relapse or suspected onset of CD
Horsthuis 2006 [24]
MRI 20 20 Y 36±13 7:13 Known CD
Scheduled for CC
Florie 2005 [21]
MRI 31 31 ? 36±12 22:9 Known CD
Scheduled for ICC because of clinical suspicion of relapsing CD
Shoenut 1994 [20]
MRI 20 12 Y 42.6(20–70) a
12:8a Suspect-ed IBD
NA
Shoenut 1993 [19]
MRI 28 19 Y 34.1(20–58) 17:11a Known IBD
Referred to MRI for evalua-tion and on medical therapy
Schreyer 2005 [22]
MRI 30 30 Y 29 (18–65) 8:22 Known CD
Routine small bowel MRI
Schreyer 2005 [23]
MRI 22 12 Y 33.4 (19–55) 5:7 Known/suspect-ed IBD
NA
Drews 2009 [32]
US 32 32 N 38.8 (17–71) 14:18 Known CD
NA
Neye 2004 [31]
US 22 22 Y 33.7 (16–56) 9:13 Known CD
Referred to gastroenter-ologist
Bozkurt 1996 [30]
US 88 32 ? 39 (16–87) a 48:40 a Suspect-ed IBD
NA
Biancone 1997 [34]
SG 17 10 ? 43±11 a 9:8 a Known CD
Patients 6–12 months after ileocecal resection
Sciarretta 1998 [33]
SG 103 31 ? 38.3 (15–78) a 54:49 a Suspect-ed IBD
NA
CC = Colonoscopy, CD = Crohn’s Disease, CTE = Computed Tomography Enterography, IBD = Inflammatory Bowel Dis-
ease, ICC = Ileocolonoscopy, MRI = Magnetic Resonance Imaging, NA = Not Applicable, Y = Yes, N = No, ? = Unclear
a. These values reflect the total number of included patients in their respective studies, not only the patients used in this
analysis. b. Median (range)
Proefschrift2018-new2.indb 27 13/9/18 10:06
28
Chapter 2
Patient characteristics
A total of 549 patients were included (75 for CT, 347 for MRI, 86 for US, and 58
for scintigraphy). The mean study size was 29 patients (range, 10–76). Study
characteristics are presented in Table 2. In ten of the studies, patients were recruited
consecutively [17, 19, 20, 22–26, 28, 31]. Studies included patients with clinically
suspected IBD, known IBD/Crohn’s disease, or a combination of both (12, 4, and 3
studies, respectively).
Imaging characteristics
Imaging equipment and specifications are presented in Tables 3, 4, 5 and 6. Bowel
preparation (fasting and/or laxatives) was used in eight studies (1 CT, 7 MRI) [17,
21–26, 28]. Luminal contrast medium was used in ten studies (3 CT, 7 MRI) [16–18,
21–23, 25, 27–29], of which one used enteroclysis [27]. Intravenous contrast medium
was used in 13 studies (2 CT, 11 MRI) [16, 17, 19–29].
Reference test
Endoscopy, biopsies and intraoperative findings were used in 11, 8 and 4 studies,
respectively (Table 7). Three studies recorded results for both endoscopy and
histology from biopsies, for which we used the histological data in our analysis [30,
33, 34].
Table 3. CT characteristics
Study Type of scanner
Bowel prepara-tion
Luminal contrast Enterogra-phy (EG) / enterocly-sis (EC)
I.V. con-trast
Post-con-trast scan timing
Mao 2013 [17]
Multiple-slice helical CT with 64 de-tector rings
1 night fasting
2000 mL 2.5% Mannitol solution 1 hr prior
EG 100 mL Iopra-mide
28 s and 60 s
Mo-hamed 2012 [16]
Multiple-slice helical CT
NS 1500–2000 mL water EG 100–150 mL Iopa-miro 300
60 s
Kolkman 1996 [18]
Siemens Somatom Plus 4
NS 500 ml water with 15 ml Rayvist 60% both on the evening prior to CT and immediately preceding scan. 500 mL with 30 mL Rayvist 1 hr prior to CT
EG NS NS
CT = Computed Tomography, NS = Not Specified
Proefschrift2018-new2.indb 28 13/9/18 10:06
28
Chapter 2
Patient characteristics
A total of 549 patients were included (75 for CT, 347 for MRI, 86 for US, and 58
for scintigraphy). The mean study size was 29 patients (range, 10–76). Study
characteristics are presented in Table 2. In ten of the studies, patients were recruited
consecutively [17, 19, 20, 22–26, 28, 31]. Studies included patients with clinically
suspected IBD, known IBD/Crohn’s disease, or a combination of both (12, 4, and 3
studies, respectively).
Imaging characteristics
Imaging equipment and specifications are presented in Tables 3, 4, 5 and 6. Bowel
preparation (fasting and/or laxatives) was used in eight studies (1 CT, 7 MRI) [17,
21–26, 28]. Luminal contrast medium was used in ten studies (3 CT, 7 MRI) [16–18,
21–23, 25, 27–29], of which one used enteroclysis [27]. Intravenous contrast medium
was used in 13 studies (2 CT, 11 MRI) [16, 17, 19–29].
Reference test
Endoscopy, biopsies and intraoperative findings were used in 11, 8 and 4 studies,
respectively (Table 7). Three studies recorded results for both endoscopy and
histology from biopsies, for which we used the histological data in our analysis [30,
33, 34].
Table 3. CT characteristics
Study Type of scanner
Bowel prepara-tion
Luminal contrast Enterogra-phy (EG) / enterocly-sis (EC)
I.V. con-trast
Post-con-trast scan timing
Mao 2013 [17]
Multiple-slice helical CT with 64 de-tector rings
1 night fasting
2000 mL 2.5% Mannitol solution 1 hr prior
EG 100 mL Iopra-mide
28 s and 60 s
Mo-hamed 2012 [16]
Multiple-slice helical CT
NS 1500–2000 mL water EG 100–150 mL Iopa-miro 300
60 s
Kolkman 1996 [18]
Siemens Somatom Plus 4
NS 500 ml water with 15 ml Rayvist 60% both on the evening prior to CT and immediately preceding scan. 500 mL with 30 mL Rayvist 1 hr prior to CT
EG NS NS
CT = Computed Tomography, NS = Not Specified
Proefschrift2018-new2.indb 28 13/9/18 10:06
29
Grading CD using CT, MRI, US and scintigraphy
Table 4. MRI characteristics
Study Field strength Coil
Bowel prepara-tion
Luminal contrast
Ad-min-istra-tion
I.V. con-trast
Post-con-trast sequence timing
Spas-molytic agent Sequences
Schill 2013 [29]
1.5 T Body and spine array coils
NS 1.5–2 L Man-nitol solution orally 45 min prior and 0.5–1 L 0.9% NaCl rectally
EG 0.2 mL/kg Gd-DT-PA
70 s 40 mg Busco-pan iv
3D T2-SPACE, bSSFP, RARE, T1-FLASH, T1-FLASH (post-contrast), T1-FLASH with fat suppression (post-contrast)
Gallego 2011 [28]
1.0 T Body coil
8 hrs fasting
1.5 L PEG and mineral salts
EG 0.1 mmol/kg Gd-DT-PA
40 s, 70 s (used for RCE) 120 s, 180 s
Busco-pan iv
bSSFP, interpo-lated 3D T1 with fat suppression (pre-/post-con-trast), RARE
Koilakou 2010 [27]
1.5 T NS NS 100–150 mL/min 0.5% methylcellu-lose solution
EC 0.1 mmol/kg Gd-DT-PA
NS 20 mg Busco-pan iv
Interpolated 3D T1 with fat suppression (post-contrast), SSFP, T2 with fat suppression
Horsthu-is 2010 [26]
3.0 T Phased array coil
Meta-mucil in 250 mL water 4 hrs prior
NS EG 0.1 mL/kg Ga-dodi-amide
NS Busco-pan iv
Interpolated 3D T1 with fat suppression (post-contrast), RARE
Girom-etti 2008 [25]
1.5 T Phased array coil
8 hrs fasting
2 L PEG EG 0.2 mL/kg Gd-DT-PA
30 s, 45 s, 60 s, 75 s, 90 s, 150 s
10 mg Busco-pan iv
bSSFP, cine, interpolated 3D T1 with fat sup-pression (pre-/post-contrast), RARE
Horsthu-is 2006 [24]
3.0 T Phased array body coil
Meta-mucil in 250 mL water 4 hrs prior
NS EG 0.05 mmol/kg Ga-dodi-amide
70 s 20 mg Busco-pan iv or 1 mg glucagon hydro-chloride
bFFE, T2-TSE, T1-FFE with fat suppression (post-contrast)
Florie 2005 [21]
1.5 T NS 4 hrs fasting
1 L water 2 hrs prior
EG 0.1 mmol/kg Gd-DT-PA
NS 20 mg Busco-pan iv or 1 mg glucagon hydro-chloride
bSSFP, interpolated 3D T1 (pre-/post-contrast), out-of-phase fast low angle shot, RARE
Shoenut 1994 [20]
1.5 T NS NS None NA 0.1 mmol/kg Gd-DT-PA
5 s, 30 s (used for RCE), 5 min, 10 min
NS T1-FLASH (pre-/post-contrast), T1 spin echo with fat suppression (post-contrast)
Shoenut 1993 [19]
1.5 T NS NS NS NA 0.1 mmol/kg Gd-DT-PA
5 s, 30 s (used for RCE), 10 min
NS T1-FLASH, T1-FLASH (post-contrast), T1 with fat sup-pression (pre-/post-contrast)
Schreyer 2005 [22]
1.5 T Phased array body coil
12 hrs fasting
2 L Mannitol solution with carob seed 1 hr prior orally and 0.4–1.0 L 0.9% NaCl rectally
NA 0.2 mmol/kg Gd-DT-PA
70 s 40 mg Busco-pan iv
2D T1-FLASH, 2D & 3D T1-FLASH with fat suppres-sion (post-con-trast), bSSFP, RARE
Schreyer 2005 [23]
1.5 T Phased array body coil
Macro-gol 3350
1.5 L Gd (5 mmol/L) mixture with water rectally
NA 0.1 mmol/kg Gd-DT-PA
NS 40 mg Busco-pan iv
2D T1-FLASH, 2D T1-FLASH fat sat (post-contrast), 3D T1-FLASH, bSSFP, RARE,
bFFE = balanced Fast Field Echo, (b)SSFP = (Balanced) Steady-State Free Precession, Gd(-DTPA) = Gadolinium(-di-ethylenetriaminepentaacetic acid), iv = intravenous, MRI = Magnetic Resonance Imaging, NaCl = Sodium chloride, NS = Not Specified, PEG = Polyethylene glycol, RARE = Rapid Acquisition with Refocusing Echoes, RCE = Relative Contrast Enhancement, T = Tesla, T1-FLASH = T1-weighted Fast Low Angle Shot, TSE = Turbo Spin Echo
Proefschrift2018-new2.indb 29 13/9/18 10:06
29
Grading CD using CT, MRI, US and scintigraphy
Table 4. MRI characteristics
Study Field strength Coil
Bowel prepara-tion
Luminal contrast
Ad-min-istra-tion
I.V. con-trast
Post-con-trast sequence timing
Spas-molytic agent Sequences
Schill 2013 [29]
1.5 T Body and spine array coils
NS 1.5–2 L Man-nitol solution orally 45 min prior and 0.5–1 L 0.9% NaCl rectally
EG 0.2 mL/kg Gd-DT-PA
70 s 40 mg Busco-pan iv
3D T2-SPACE, bSSFP, RARE, T1-FLASH, T1-FLASH (post-contrast), T1-FLASH with fat suppression (post-contrast)
Gallego 2011 [28]
1.0 T Body coil
8 hrs fasting
1.5 L PEG and mineral salts
EG 0.1 mmol/kg Gd-DT-PA
40 s, 70 s (used for RCE) 120 s, 180 s
Busco-pan iv
bSSFP, interpo-lated 3D T1 with fat suppression (pre-/post-con-trast), RARE
Koilakou 2010 [27]
1.5 T NS NS 100–150 mL/min 0.5% methylcellu-lose solution
EC 0.1 mmol/kg Gd-DT-PA
NS 20 mg Busco-pan iv
Interpolated 3D T1 with fat suppression (post-contrast), SSFP, T2 with fat suppression
Horsthu-is 2010 [26]
3.0 T Phased array coil
Meta-mucil in 250 mL water 4 hrs prior
NS EG 0.1 mL/kg Ga-dodi-amide
NS Busco-pan iv
Interpolated 3D T1 with fat suppression (post-contrast), RARE
Girom-etti 2008 [25]
1.5 T Phased array coil
8 hrs fasting
2 L PEG EG 0.2 mL/kg Gd-DT-PA
30 s, 45 s, 60 s, 75 s, 90 s, 150 s
10 mg Busco-pan iv
bSSFP, cine, interpolated 3D T1 with fat sup-pression (pre-/post-contrast), RARE
Horsthu-is 2006 [24]
3.0 T Phased array body coil
Meta-mucil in 250 mL water 4 hrs prior
NS EG 0.05 mmol/kg Ga-dodi-amide
70 s 20 mg Busco-pan iv or 1 mg glucagon hydro-chloride
bFFE, T2-TSE, T1-FFE with fat suppression (post-contrast)
Florie 2005 [21]
1.5 T NS 4 hrs fasting
1 L water 2 hrs prior
EG 0.1 mmol/kg Gd-DT-PA
NS 20 mg Busco-pan iv or 1 mg glucagon hydro-chloride
bSSFP, interpolated 3D T1 (pre-/post-contrast), out-of-phase fast low angle shot, RARE
Shoenut 1994 [20]
1.5 T NS NS None NA 0.1 mmol/kg Gd-DT-PA
5 s, 30 s (used for RCE), 5 min, 10 min
NS T1-FLASH (pre-/post-contrast), T1 spin echo with fat suppression (post-contrast)
Shoenut 1993 [19]
1.5 T NS NS NS NA 0.1 mmol/kg Gd-DT-PA
5 s, 30 s (used for RCE), 10 min
NS T1-FLASH, T1-FLASH (post-contrast), T1 with fat sup-pression (pre-/post-contrast)
Schreyer 2005 [22]
1.5 T Phased array body coil
12 hrs fasting
2 L Mannitol solution with carob seed 1 hr prior orally and 0.4–1.0 L 0.9% NaCl rectally
NA 0.2 mmol/kg Gd-DT-PA
70 s 40 mg Busco-pan iv
2D T1-FLASH, 2D & 3D T1-FLASH with fat suppres-sion (post-con-trast), bSSFP, RARE
Schreyer 2005 [23]
1.5 T Phased array body coil
Macro-gol 3350
1.5 L Gd (5 mmol/L) mixture with water rectally
NA 0.1 mmol/kg Gd-DT-PA
NS 40 mg Busco-pan iv
2D T1-FLASH, 2D T1-FLASH fat sat (post-contrast), 3D T1-FLASH, bSSFP, RARE,
bFFE = balanced Fast Field Echo, (b)SSFP = (Balanced) Steady-State Free Precession, Gd(-DTPA) = Gadolinium(-di-ethylenetriaminepentaacetic acid), iv = intravenous, MRI = Magnetic Resonance Imaging, NaCl = Sodium chloride, NS = Not Specified, PEG = Polyethylene glycol, RARE = Rapid Acquisition with Refocusing Echoes, RCE = Relative Contrast Enhancement, T = Tesla, T1-FLASH = T1-weighted Fast Low Angle Shot, TSE = Turbo Spin Echo
Proefschrift2018-new2.indb 29 13/9/18 10:06
30
Chapter 2
Imaging and reference test interpretation
Thirteen of the studies used an interval of less than one month between imaging and
reference test [17, 19–23, 26, 28, 29, 31–34]. The imaging features most commonly
used for evaluation were bowel wall thickness and post-contrast enhancement (or
tracer uptake for scintigraphy), which were both used in 17 studies (Table 7). The
reference test and imaging criteria for each study are presented in Tables 8 and 9.
Publication bias
Linear regression analysis on MRI per-patient data showed a regression coefficient
of 0.4 (95% CI: 0.9–0.9), with no significant relationship between accurate grading
and 1/√ESS (P = 0.09). Data in other groups were deemed insufficient for performing
linear regression analyses.
Table 5. US characteristics
Study (US) Transducer type + frequency Bowel preparation
Luminal contrast
I.V. con-trast
Doppler + type
Drews 2009 [32]
Linear 5–12MHz (neoterminal ile-um) and convex 2–5 MHz (entire abdomen)
NS NS NS Power Doppler
Neye 2004 [31]
Linear 5–12Mhz and dynamic sector scanner 4–7 MHz
NS NS NS Pulsed Doppler and Power Doppler
Bozkurt 1996 [30]
Linear 7.5MHz NS NS NS NA
MHz = Megahertz, NA = Not Applicable, NS = Not Specified, US = Ultrasound
Table 6. Scintigraphy characteristics
Study (Scintig-raphy)
Labeling of: Labeling by:
Amount of tracer
Scans Criteria used for image analysis
Kolkman 1996 [18]
Antigranulocyte antibodies
Tc-99m HMPAO
NS 2 scans (at 1 hrs and 4 hrs)
Uptake of tracer compared to bone marrow and liver
Biancone 1997 [34]
Leukocytes Tc-99m HMPAO
185 MBq 2 scans (at 30 min and 3 hrs)
Uptake of tracer compared to bone marrow and liver
Sciarretta 1998 [33]
Leukocytes Tc-99m HMPAO
370–555 MBq 3 scans (at 30 min, 2–2.5 hrs and 24 hrs)
Uptake of tracer compared to bone marrow and liver
MBq = Megabecquerel, NS = Not Specified, Tc-99m HMPAO = Technetium hexamethylpropyleneamine oxime
Proefschrift2018-new2.indb 30 13/9/18 10:06
30
Chapter 2
Imaging and reference test interpretation
Thirteen of the studies used an interval of less than one month between imaging and
reference test [17, 19–23, 26, 28, 29, 31–34]. The imaging features most commonly
used for evaluation were bowel wall thickness and post-contrast enhancement (or
tracer uptake for scintigraphy), which were both used in 17 studies (Table 7). The
reference test and imaging criteria for each study are presented in Tables 8 and 9.
Publication bias
Linear regression analysis on MRI per-patient data showed a regression coefficient
of 0.4 (95% CI: 0.9–0.9), with no significant relationship between accurate grading
and 1/√ESS (P = 0.09). Data in other groups were deemed insufficient for performing
linear regression analyses.
Table 5. US characteristics
Study (US) Transducer type + frequency Bowel preparation
Luminal contrast
I.V. con-trast
Doppler + type
Drews 2009 [32]
Linear 5–12MHz (neoterminal ile-um) and convex 2–5 MHz (entire abdomen)
NS NS NS Power Doppler
Neye 2004 [31]
Linear 5–12Mhz and dynamic sector scanner 4–7 MHz
NS NS NS Pulsed Doppler and Power Doppler
Bozkurt 1996 [30]
Linear 7.5MHz NS NS NS NA
MHz = Megahertz, NA = Not Applicable, NS = Not Specified, US = Ultrasound
Table 6. Scintigraphy characteristics
Study (Scintig-raphy)
Labeling of: Labeling by:
Amount of tracer
Scans Criteria used for image analysis
Kolkman 1996 [18]
Antigranulocyte antibodies
Tc-99m HMPAO
NS 2 scans (at 1 hrs and 4 hrs)
Uptake of tracer compared to bone marrow and liver
Biancone 1997 [34]
Leukocytes Tc-99m HMPAO
185 MBq 2 scans (at 30 min and 3 hrs)
Uptake of tracer compared to bone marrow and liver
Sciarretta 1998 [33]
Leukocytes Tc-99m HMPAO
370–555 MBq 3 scans (at 30 min, 2–2.5 hrs and 24 hrs)
Uptake of tracer compared to bone marrow and liver
MBq = Megabecquerel, NS = Not Specified, Tc-99m HMPAO = Technetium hexamethylpropyleneamine oxime
Proefschrift2018-new2.indb 30 13/9/18 10:06
31
Grading CD using CT, MRI, US and scintigraphy
Data analysis
Results from our data analysis are presented in Table 10. Three-by-three contingency
tables for each study can be found in the supplementary materials (Appendix E3).
Per-patient
Data was provided on a per-patient basis in 13 studies (evaluating CT in 2, MRI in
9, US in 1 and scintigraphy in 1) (Fig. 4). I2 values for overall grading accuracy for
groups with more than one dataset were as follows: 67.7% (95% CI: 42.6–81.8%) for
CT, and 73.9% (95% CI: 56.2–84.4%) for MRI.
CT and MRI data were pooled for each modality (I2< 75%). US and scintigraphy
were not pooled, as only one dataset was available for each modality. CT, MRI, US
and scintigraphy showed accurate grading estimates of 86% (95% CI: 75–93%), 84%
(95% CI: 67–93%), 44% (95% CI: 28–61%) and 40% (95% CI: 16–70%), respectively.
CT and MRI showed similar overall grading accuracy (P = 0.8), both higher than US
(P = 0.0001 and P = 0.001, respectively) and scintigraphy (P = 0.003 and P = 0.01,
respectively). CT and MRI showed similar over-grading (P = 0.8) and under-grading
(P = 0.5). Both showed less under-grading than US (P = 0.002 and P = 0.003,
respectively) and scintigraphy (P=0.0005 and P=0.001, respectively).
Per-segment
Data were provided on a per-segment basis in seven articles, of which one evaluated
both CT and scintigraphy, two evaluated MRI, two evaluated US, and two evaluated
scintigraphy, respectively (Fig. 4). I2 values were 86.3% (95% CI: 66.4–94.4%) for
MRI, 91.5% (95% CI: 79.1–96.6%) for US, and 0% for scintigraphy. MRI and US data
were not pooled, as data were too heterogeneous (I2 ≥ 75%). Data on CT were also
not pooled, as only one dataset was available. The overall grading accuracy was
87% (95% CI: 77–93%) for CT and 86% (95% CI: 80–91%) for scintigraphy. CT and
scintigraphy showed similar overall grading accuracy (P=0.8), over-grading (P=0.2)
and under-grading (P=0.5). Accuracy for MRI and US ranged from 67 to 82% and
56 to 75%, respectively.
Proefschrift2018-new2.indb 31 13/9/18 10:06
31
Grading CD using CT, MRI, US and scintigraphy
Data analysis
Results from our data analysis are presented in Table 10. Three-by-three contingency
tables for each study can be found in the supplementary materials (Appendix E3).
Per-patient
Data was provided on a per-patient basis in 13 studies (evaluating CT in 2, MRI in
9, US in 1 and scintigraphy in 1) (Fig. 4). I2 values for overall grading accuracy for
groups with more than one dataset were as follows: 67.7% (95% CI: 42.6–81.8%) for
CT, and 73.9% (95% CI: 56.2–84.4%) for MRI.
CT and MRI data were pooled for each modality (I2< 75%). US and scintigraphy
were not pooled, as only one dataset was available for each modality. CT, MRI, US
and scintigraphy showed accurate grading estimates of 86% (95% CI: 75–93%), 84%
(95% CI: 67–93%), 44% (95% CI: 28–61%) and 40% (95% CI: 16–70%), respectively.
CT and MRI showed similar overall grading accuracy (P = 0.8), both higher than US
(P = 0.0001 and P = 0.001, respectively) and scintigraphy (P = 0.003 and P = 0.01,
respectively). CT and MRI showed similar over-grading (P = 0.8) and under-grading
(P = 0.5). Both showed less under-grading than US (P = 0.002 and P = 0.003,
respectively) and scintigraphy (P=0.0005 and P=0.001, respectively).
Per-segment
Data were provided on a per-segment basis in seven articles, of which one evaluated
both CT and scintigraphy, two evaluated MRI, two evaluated US, and two evaluated
scintigraphy, respectively (Fig. 4). I2 values were 86.3% (95% CI: 66.4–94.4%) for
MRI, 91.5% (95% CI: 79.1–96.6%) for US, and 0% for scintigraphy. MRI and US data
were not pooled, as data were too heterogeneous (I2 ≥ 75%). Data on CT were also
not pooled, as only one dataset was available. The overall grading accuracy was
87% (95% CI: 77–93%) for CT and 86% (95% CI: 80–91%) for scintigraphy. CT and
scintigraphy showed similar overall grading accuracy (P=0.8), over-grading (P=0.2)
and under-grading (P=0.5). Accuracy for MRI and US ranged from 67 to 82% and
56 to 75%, respectively.
Proefschrift2018-new2.indb 31 13/9/18 10:06
32
Chapter 2
Tab
le 7
. Im
ag
ing
an
d r
efe
ren
ce t
est
in
terp
reta
tio
n
Stud
yIm
-ag
ing
m
o-
dal
ity
Ref
er-
ence
tes
t us
ed in
an
alys
is
Ana
lysi
s p
er p
a-ti
ent/
per
se
gm
ent
Tim
e in
terv
al
(day
s) in
dex
&
ref
eren
ce
test
Par
t o
f G
I tr
act
exam
ined
Gra
din
g s
cale
in
dex
tes
tG
rad
ing
sca
le
refe
renc
e te
stIm
agin
g f
eatu
res
used
fo
r g
rad
ing
dis
ease
act
ivit
y
Mao
20
13
[17]
CT
ICC
Pati
en
t<
=7
Neo
term
i-n
al ile
um
0–3
i0–i
4 (
Ru
tgeert
s sc
ore
)B
ow
el w
all
thic
kn
ess
, p
ost
-co
ntr
ast
en
han
cem
en
t, m
uco
sal ir
reg
ula
r-it
ies/
hyp
erd
en
siti
es,
mu
ral st
rati
ficati
on
, st
en
osi
s an
d p
rest
en
oti
c d
il-ata
tio
n a
nd
extr
alu
min
al fi
nd
ing
s (l
ym
ph
no
des,
ab
scess
es,
fis
tula
s,
co
mb
sig
n, cre
ep
ing
fat)
Mo
ham
ed
20
12 [
16]
CT
B, S
SP
ati
en
t<
=7a
Co
lon
an
d T
IM
ild, m
od
era
te,
severe
Mild
, m
od
era
te,
severe
B
ow
el w
all
thic
kn
ess
, p
ost
-co
ntr
ast
en
han
cem
en
t, e
xtr
alu
min
al
fin
din
gs
(lym
ph
no
des,
ab
scess
es,
fis
tula
s, c
om
b s
ign
, cre
ep
ing
fat,
ed
em
a)
Ko
lkm
an
19
96
[18
]C
T,
SG
B, IC
C,
SS
Seg
men
t1–
50
C
olo
n
an
d T
I0
–3 (
CT
), 0
–4 (
SG
)0
–3B
ow
el w
all
thic
kn
ess
, T
1 en
han
cem
en
t an
d p
att
ern
, u
lcera
tio
n,
do
ub
le-h
alo
sig
n a
nd
extr
alu
min
al fi
nd
ing
s (c
reep
ing
fat,
mese
nte
ric
fib
rovasc
ula
r st
ran
ds)
(C
T).
Up
take o
f tr
acer
co
mp
are
d t
o b
on
e
marr
ow
an
d liv
er
(SG
)
Sch
ill
20
13 [
29
]M
RI
SS
Pati
en
t<
=28
N
SB
1, B
2, B
3 (
Mo
ntr
e-
al cla
ss.)
B1,
B2, B
3 (
Mo
n-
treal cla
ss.)
Targ
et
sig
n, T
2 m
ura
l si
gn
al in
ten
sity
, in
flam
mato
ry m
ass
, st
en
osi
s w
ith
pre
sten
oti
c d
ilata
tio
n a
nd
extr
alu
min
al fi
nd
ing
s (l
ym
ph
no
des,
ab
scess
es,
fis
tula
s, c
om
b s
ign
)
Galle
go
20
11 [
28
]M
RI
ICC
Pati
en
t<
=15
Ileu
mN
on
e, m
ild, m
od
er-
ate
/severe
0
–3 (
SE
S-C
D)
Bo
wel w
all
thic
kn
ess
an
d e
dem
a, T
1 en
han
cem
en
t, m
uco
sal ab
-n
orm
alit
ies,
in
flam
mato
ry m
ass
, m
oti
lity,
ste
no
sis
an
d e
xtr
alu
min
al
fin
din
gs
(lym
ph
no
des,
fis
tula
s)
Ko
ilako
u
20
10 [
27]
MR
IIC
CP
ati
en
tN
SN
eo
term
i-n
al ile
um
0–3
i0–i
4 (
Ru
tgeert
s sc
ore
)B
ow
el w
all
thic
kn
ess
, T
1 en
han
cem
en
t, T
2 m
ura
l si
gn
al,
mu
co
sal
irre
gu
lari
ties,
in
filt
rate
, ed
em
a, st
en
osi
s an
d p
rest
en
oti
c d
ilata
tio
n,
extr
alu
min
al fi
nd
ing
s (a
bsc
ess
es,
fis
tula
s)
Ho
rsth
uis
20
10 [
26
]M
RI
EG
D, IC
CP
ati
en
t<
= 1
4
Co
lon
, T
I an
d d
uo
de-
nu
m
No
ne, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
No
ne, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, s
ten
osi
s an
d p
rest
en
oti
c
dila
tati
on
.
Gir
om
ett
i 20
08
[25
]M
RI
BP
ati
en
tN
ST
IN
on
e, m
ild, m
od
er-
ate
/severe
No
ne, m
ild, m
od
-era
te/s
evere
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, m
uco
sal ab
no
rmalit
ies,
in
flam
mato
ry m
ass
, m
ese
nte
ric invo
lvem
en
t, m
oti
lity,
ste
no
sis
an
d
extr
alu
min
al fi
nd
ing
s (l
ym
ph
no
des,
fis
tula
s)
Ho
rsth
uis
20
06
[2
4]
MR
IIC
CP
ati
en
t1–
48
C
olo
n
an
d T
IN
on
e, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
No
ne, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, u
lcera
tio
n, le
ng
th o
f d
isease
d
seg
men
t, c
ob
ble
sto
nin
g, extr
alu
min
al fi
nd
ing
s (l
ym
ph
no
des,
ab
-sc
ess
es,
fis
tula
s, c
om
b s
ign
an
d c
reep
ing
fat)
Flo
rie
20
05
[21]
MR
IIC
CP
ati
en
t<
=14
C
olo
n
an
d T
IN
on
e, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
No
ne, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, s
ten
osi
s, t
arg
et
sig
n, co
b-
ble
sto
nin
g
B =
Bio
psi
es,
CC
= C
olo
no
sco
py,
CT
= C
om
pu
ted
to
mo
gra
ph
y, E
GD
= E
sop
hag
og
ast
rod
uo
den
osc
op
y, IC
C =
Ile
oco
lon
osc
op
y, M
RI =
Mag
neti
c R
eso
nan
ce Im
ag
ing
, N
S =
No
t S
pecif
ied
, S
ES
-C
D =
Sim
ple
En
do
sco
pic
Sco
re f
or
Cro
hn
’s D
isease
, S
G =
Scin
tig
rap
hy,
SS
= S
urg
ical S
pecim
en
s, T
I =
Term
inal Ileu
m, U
S =
Ult
raso
un
da. T
ime in
terv
al w
as
no
t sp
ecif
ied
fo
r p
ati
en
ts u
nd
erg
oin
g s
urg
ery
.
Proefschrift2018-new2.indb 32 13/9/18 10:06
32
Chapter 2
Tab
le 7
. Im
ag
ing
an
d r
efe
ren
ce t
est
in
terp
reta
tio
n
Stud
yIm
-ag
ing
m
o-
dal
ity
Ref
er-
ence
tes
t us
ed in
an
alys
is
Ana
lysi
s p
er p
a-ti
ent/
per
se
gm
ent
Tim
e in
terv
al
(day
s) in
dex
&
ref
eren
ce
test
Par
t o
f G
I tr
act
exam
ined
Gra
din
g s
cale
in
dex
tes
tG
rad
ing
sca
le
refe
renc
e te
stIm
agin
g f
eatu
res
used
fo
r g
rad
ing
dis
ease
act
ivit
y
Mao
20
13
[17]
CT
ICC
Pati
en
t<
=7
Neo
term
i-n
al ile
um
0–3
i0–i
4 (
Ru
tgeert
s sc
ore
)B
ow
el w
all
thic
kn
ess
, p
ost
-co
ntr
ast
en
han
cem
en
t, m
uco
sal ir
reg
ula
r-it
ies/
hyp
erd
en
siti
es,
mu
ral st
rati
ficati
on
, st
en
osi
s an
d p
rest
en
oti
c d
il-ata
tio
n a
nd
extr
alu
min
al fi
nd
ing
s (l
ym
ph
no
des,
ab
scess
es,
fis
tula
s,
co
mb
sig
n, cre
ep
ing
fat)
Mo
ham
ed
20
12 [
16]
CT
B, S
SP
ati
en
t<
=7a
Co
lon
an
d T
IM
ild, m
od
era
te,
severe
Mild
, m
od
era
te,
severe
B
ow
el w
all
thic
kn
ess
, p
ost
-co
ntr
ast
en
han
cem
en
t, e
xtr
alu
min
al
fin
din
gs
(lym
ph
no
des,
ab
scess
es,
fis
tula
s, c
om
b s
ign
, cre
ep
ing
fat,
ed
em
a)
Ko
lkm
an
19
96
[18
]C
T,
SG
B, IC
C,
SS
Seg
men
t1–
50
C
olo
n
an
d T
I0
–3 (
CT
), 0
–4 (
SG
)0
–3B
ow
el w
all
thic
kn
ess
, T
1 en
han
cem
en
t an
d p
att
ern
, u
lcera
tio
n,
do
ub
le-h
alo
sig
n a
nd
extr
alu
min
al fi
nd
ing
s (c
reep
ing
fat,
mese
nte
ric
fib
rovasc
ula
r st
ran
ds)
(C
T).
Up
take o
f tr
acer
co
mp
are
d t
o b
on
e
marr
ow
an
d liv
er
(SG
)
Sch
ill
20
13 [
29
]M
RI
SS
Pati
en
t<
=28
N
SB
1, B
2, B
3 (
Mo
ntr
e-
al cla
ss.)
B1,
B2, B
3 (
Mo
n-
treal cla
ss.)
Targ
et
sig
n, T
2 m
ura
l si
gn
al in
ten
sity
, in
flam
mato
ry m
ass
, st
en
osi
s w
ith
pre
sten
oti
c d
ilata
tio
n a
nd
extr
alu
min
al fi
nd
ing
s (l
ym
ph
no
des,
ab
scess
es,
fis
tula
s, c
om
b s
ign
)
Galle
go
20
11 [
28
]M
RI
ICC
Pati
en
t<
=15
Ileu
mN
on
e, m
ild, m
od
er-
ate
/severe
0
–3 (
SE
S-C
D)
Bo
wel w
all
thic
kn
ess
an
d e
dem
a, T
1 en
han
cem
en
t, m
uco
sal ab
-n
orm
alit
ies,
in
flam
mato
ry m
ass
, m
oti
lity,
ste
no
sis
an
d e
xtr
alu
min
al
fin
din
gs
(lym
ph
no
des,
fis
tula
s)
Ko
ilako
u
20
10 [
27]
MR
IIC
CP
ati
en
tN
SN
eo
term
i-n
al ile
um
0–3
i0–i
4 (
Ru
tgeert
s sc
ore
)B
ow
el w
all
thic
kn
ess
, T
1 en
han
cem
en
t, T
2 m
ura
l si
gn
al,
mu
co
sal
irre
gu
lari
ties,
in
filt
rate
, ed
em
a, st
en
osi
s an
d p
rest
en
oti
c d
ilata
tio
n,
extr
alu
min
al fi
nd
ing
s (a
bsc
ess
es,
fis
tula
s)
Ho
rsth
uis
20
10 [
26
]M
RI
EG
D, IC
CP
ati
en
t<
= 1
4
Co
lon
, T
I an
d d
uo
de-
nu
m
No
ne, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
No
ne, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, s
ten
osi
s an
d p
rest
en
oti
c
dila
tati
on
.
Gir
om
ett
i 20
08
[25
]M
RI
BP
ati
en
tN
ST
IN
on
e, m
ild, m
od
er-
ate
/severe
No
ne, m
ild, m
od
-era
te/s
evere
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, m
uco
sal ab
no
rmalit
ies,
in
flam
mato
ry m
ass
, m
ese
nte
ric invo
lvem
en
t, m
oti
lity,
ste
no
sis
an
d
extr
alu
min
al fi
nd
ing
s (l
ym
ph
no
des,
fis
tula
s)
Ho
rsth
uis
20
06
[2
4]
MR
IIC
CP
ati
en
t1–
48
C
olo
n
an
d T
IN
on
e, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
No
ne, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, u
lcera
tio
n, le
ng
th o
f d
isease
d
seg
men
t, c
ob
ble
sto
nin
g, extr
alu
min
al fi
nd
ing
s (l
ym
ph
no
des,
ab
-sc
ess
es,
fis
tula
s, c
om
b s
ign
an
d c
reep
ing
fat)
Flo
rie
20
05
[21]
MR
IIC
CP
ati
en
t<
=14
C
olo
n
an
d T
IN
on
e, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
No
ne, m
ild,
mo
dera
te, se
vere
(s
ub
jecti
ve)
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, s
ten
osi
s, t
arg
et
sig
n, co
b-
ble
sto
nin
g
B =
Bio
psi
es,
CC
= C
olo
no
sco
py,
CT
= C
om
pu
ted
to
mo
gra
ph
y, E
GD
= E
sop
hag
og
ast
rod
uo
den
osc
op
y, IC
C =
Ile
oco
lon
osc
op
y, M
RI =
Mag
neti
c R
eso
nan
ce Im
ag
ing
, N
S =
No
t S
pecif
ied
, S
ES
-C
D =
Sim
ple
En
do
sco
pic
Sco
re f
or
Cro
hn
’s D
isease
, S
G =
Scin
tig
rap
hy,
SS
= S
urg
ical S
pecim
en
s, T
I =
Term
inal Ileu
m, U
S =
Ult
raso
un
da. T
ime in
terv
al w
as
no
t sp
ecif
ied
fo
r p
ati
en
ts u
nd
erg
oin
g s
urg
ery
.
Proefschrift2018-new2.indb 32 13/9/18 10:06
33
Grading CD using CT, MRI, US and scintigraphy
Stud
yIm
-ag
ing
m
o-
dal
ity
Ref
er-
ence
tes
t us
ed in
an
alys
is
Ana
lysi
s p
er p
a-ti
ent/
per
se
gm
ent
Tim
e in
terv
al
(day
s) in
dex
&
ref
eren
ce
test
Par
t o
f G
I tr
act
exam
ined
Gra
din
g s
cale
in
dex
tes
tG
rad
ing
sca
le
refe
renc
e te
stIm
agin
g f
eatu
res
used
fo
r g
rad
ing
dis
ease
act
ivit
y
Sh
oen
ut
199
4 [
20
]M
RI
BP
ati
en
t<
=3
C
olo
n
an
d T
IM
ild, m
od
era
te,
severe
Mild
, m
od
era
te,
severe
(su
bje
c-
tive)
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, len
gth
of
dis
ease
d s
eg
men
t
Sh
oen
ut
199
3 [
19]
MR
IC
C, S
SP
ati
en
t<
=7
Co
lon
an
d T
IM
ild, m
od
era
te,
severe
Mild
, m
od
era
te,
severe
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, len
gth
of
dis
ease
d s
eg
men
t
Sch
reyer
20
05
[2
2]
MR
IIC
CS
eg
men
t<
=7
Co
lon
an
d T
I0
–20
–2B
ow
el w
all
thic
kn
ess
, T
1 en
han
cem
en
t, s
ten
osi
s, lym
ph
no
des,
lo
cal
inje
cti
on
fo
r in
flam
mati
on
ass
ess
men
t
Sch
reyer
20
05
[2
3]
MR
IIC
CS
eg
men
t1
Co
lon
an
d T
I0
–20
–2B
ow
el w
all
thic
kn
ess
, T
1 en
han
cem
en
t, lym
ph
no
des,
mese
nte
ric
inje
cti
on
Dre
ws
20
09
[3
2]
US
BP
ati
en
t<
=5
C
olo
n a
nd
(n
eo
-)
term
inal
ileu
m
0–4
0–4
Vasc
ula
rizati
on
an
d t
hic
kn
ess
of
the b
ow
el w
all,
pre
serv
ati
on
of
five-l
ayer
stru
ctu
re, le
ng
th o
f d
isease
d s
eg
men
t
Neye
20
04
[3
1]
US
ICC
Seg
men
t<
=3
C
olo
n
an
d T
I0
–30
–3V
asc
ula
rizati
on
an
d t
hic
kn
ess
of
the b
ow
el w
all
Bo
zku
rt
199
6 [
30
]U
SB
Seg
men
tN
SC
olo
n0
–20
–2 (
sub
jecti
ve)
Bo
wel w
all
thic
kn
ess
, ech
og
en
icit
y o
f th
e b
ow
el w
all,
sm
oo
thn
ess
of
bo
un
dari
es,
vis
ibili
ty o
f in
div
idu
al b
ow
el w
all
layers
Bia
nco
ne
199
7 [
34
]S
GB
Pati
en
t<
=14
N
eo
term
i-n
al ile
um
0–3
0–3
(su
bje
cti
ve)
Up
take o
f tr
acer
co
mp
are
d t
o b
on
e m
arr
ow
an
d liv
er
Scia
rrett
a
199
8 [
33
]S
GB
Seg
men
t<
= 7
C
olo
n
an
d T
I0
–30
–3 (
sub
jecti
ve)
Up
take o
f tr
acer
co
mp
are
d t
o b
on
e m
arr
ow
an
d liv
er
B =
Bio
psi
es,
CC
= C
olo
no
sco
py,
CT
= C
om
pu
ted
to
mo
gra
ph
y, E
GD
= E
sop
hag
og
ast
rod
uo
den
osc
op
y, IC
C =
Ile
oco
lon
osc
op
y, M
RI =
Mag
neti
c R
eso
nan
ce Im
ag
ing
, N
S =
No
t S
pecif
ied
, S
ES
-C
D =
Sim
ple
En
do
sco
pic
Sco
re f
or
Cro
hn
’s D
isease
, S
G =
Scin
tig
rap
hy,
SS
= S
urg
ical S
pecim
en
s, T
I =
Term
inal Ileu
m, U
S =
Ult
raso
un
da. T
ime in
terv
al w
as
no
t sp
ecif
ied
fo
r p
ati
en
ts u
nd
erg
oin
g s
urg
ery
.
Proefschrift2018-new2.indb 33 13/9/18 10:06
33
Grading CD using CT, MRI, US and scintigraphy
Stud
yIm
-ag
ing
m
o-
dal
ity
Ref
er-
ence
tes
t us
ed in
an
alys
is
Ana
lysi
s p
er p
a-ti
ent/
per
se
gm
ent
Tim
e in
terv
al
(day
s) in
dex
&
ref
eren
ce
test
Par
t o
f G
I tr
act
exam
ined
Gra
din
g s
cale
in
dex
tes
tG
rad
ing
sca
le
refe
renc
e te
stIm
agin
g f
eatu
res
used
fo
r g
rad
ing
dis
ease
act
ivit
y
Sh
oen
ut
199
4 [
20
]M
RI
BP
ati
en
t<
=3
C
olo
n
an
d T
IM
ild, m
od
era
te,
severe
Mild
, m
od
era
te,
severe
(su
bje
c-
tive)
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, len
gth
of
dis
ease
d s
eg
men
t
Sh
oen
ut
199
3 [
19]
MR
IC
C, S
SP
ati
en
t<
=7
Co
lon
an
d T
IM
ild, m
od
era
te,
severe
Mild
, m
od
era
te,
severe
Bo
wel w
all
thic
kn
ess
, T
1 en
han
cem
en
t, len
gth
of
dis
ease
d s
eg
men
t
Sch
reyer
20
05
[2
2]
MR
IIC
CS
eg
men
t<
=7
Co
lon
an
d T
I0
–20
–2B
ow
el w
all
thic
kn
ess
, T
1 en
han
cem
en
t, s
ten
osi
s, lym
ph
no
des,
lo
cal
inje
cti
on
fo
r in
flam
mati
on
ass
ess
men
t
Sch
reyer
20
05
[2
3]
MR
IIC
CS
eg
men
t1
Co
lon
an
d T
I0
–20
–2B
ow
el w
all
thic
kn
ess
, T
1 en
han
cem
en
t, lym
ph
no
des,
mese
nte
ric
inje
cti
on
Dre
ws
20
09
[3
2]
US
BP
ati
en
t<
=5
C
olo
n a
nd
(n
eo
-)
term
inal
ileu
m
0–4
0–4
Vasc
ula
rizati
on
an
d t
hic
kn
ess
of
the b
ow
el w
all,
pre
serv
ati
on
of
five-l
ayer
stru
ctu
re, le
ng
th o
f d
isease
d s
eg
men
t
Neye
20
04
[3
1]
US
ICC
Seg
men
t<
=3
C
olo
n
an
d T
I0
–30
–3V
asc
ula
rizati
on
an
d t
hic
kn
ess
of
the b
ow
el w
all
Bo
zku
rt
199
6 [
30
]U
SB
Seg
men
tN
SC
olo
n0
–20
–2 (
sub
jecti
ve)
Bo
wel w
all
thic
kn
ess
, ech
og
en
icit
y o
f th
e b
ow
el w
all,
sm
oo
thn
ess
of
bo
un
dari
es,
vis
ibili
ty o
f in
div
idu
al b
ow
el w
all
layers
Bia
nco
ne
199
7 [
34
]S
GB
Pati
en
t<
=14
N
eo
term
i-n
al ile
um
0–3
0–3
(su
bje
cti
ve)
Up
take o
f tr
acer
co
mp
are
d t
o b
on
e m
arr
ow
an
d liv
er
Scia
rrett
a
199
8 [
33
]S
GB
Seg
men
t<
= 7
C
olo
n
an
d T
I0
–30
–3 (
sub
jecti
ve)
Up
take o
f tr
acer
co
mp
are
d t
o b
on
e m
arr
ow
an
d liv
er
B =
Bio
psi
es,
CC
= C
olo
no
sco
py,
CT
= C
om
pu
ted
to
mo
gra
ph
y, E
GD
= E
sop
hag
og
ast
rod
uo
den
osc
op
y, IC
C =
Ile
oco
lon
osc
op
y, M
RI =
Mag
neti
c R
eso
nan
ce Im
ag
ing
, N
S =
No
t S
pecif
ied
, S
ES
-C
D =
Sim
ple
En
do
sco
pic
Sco
re f
or
Cro
hn
’s D
isease
, S
G =
Scin
tig
rap
hy,
SS
= S
urg
ical S
pecim
en
s, T
I =
Term
inal Ileu
m, U
S =
Ult
raso
un
da. T
ime in
terv
al w
as
no
t sp
ecif
ied
fo
r p
ati
en
ts u
nd
erg
oin
g s
urg
ery
.
Proefschrift2018-new2.indb 33 13/9/18 10:06
34
Chapter 2
Tab
le 8
. Ori
gin
al re
fere
nce t
est
cri
teri
a a
nd
cate
go
rizati
on
fo
r th
is s
tud
y
Stud
yN
one
Mild
Seve
re
Mao
et
al.
his
tolo
gic
al
sco
re (
Ru
tgeert
s sc
ore
) [1
7]
i0: N
o lesi
on
si1
: L
ess
th
an
5 a
ph
tho
us
lesi
on
si2
: M
ore
th
an
5 a
ph
tho
us
lesi
on
s w
ith
no
rmal
mu
co
sa b
etw
een
th
e
lesi
on
s o
r sk
ip a
reas
of
larg
er
lesi
on
s o
r le
sio
ns
co
nfi
ned
to
ile
oco
lon
ic
an
ast
om
osi
s
i3: D
iffu
se a
ph
tho
us
ileit
is w
ith
dif
fuse
ly
infl
am
ed
mu
co
sa
i4: D
iffu
se in
flam
ma-
tio
n w
ith
alr
ead
y larg
e
ulc
ers
, n
od
ule
s, a
nd
/or
narr
ow
ing
Mo
ham
ed
et
al.
his
tolo
g-
ical sc
ore
(su
bje
cti
ve)
[16
]
-M
ild
Mo
dera
teS
evere
Ko
lkm
an
et
al h
isto
log
i-cal sc
ore
[18
]0
: N
o a
bn
orm
alit
ies,
or
pla
in f
ibro
sis
1: S
om
e in
filt
rati
on
of
po
lym
orp
ho
nu
cle
ar
leu
ko
-cyte
s, n
o u
lcera
tio
n2: M
od
era
te in
filt
rati
on
o
f p
oly
mo
rph
on
ucle
ar
leu
ko
cyte
s, s
om
e u
lcer-
ati
on
pre
sen
t
3: S
evere
ly u
lcera
ted
w
ith
mass
ive in
filt
rati
on
o
f p
oly
mo
rph
on
ucle
ar
leu
ko
cyte
s
Sch
ill e
t al.
surg
ical sc
ore
(b
ase
d o
n M
on
treal cla
s-si
ficati
on
) [2
9]
-B
1: N
on
-str
ictu
rin
g a
nd
no
n-p
en
etr
ati
ng
B2: S
tric
turi
ng
B3
: P
en
etr
ati
ng
Galle
go
et
al.
en
do
sco
pic
sc
ore
(S
ES
-CD
)a
[28
]0
-2 p
oin
ts: In
acti
ve
3-6
po
ints
: m
ild d
isease
≥7 p
oin
ts: m
od
era
te/s
evere
dis
ease
Ko
ilako
u e
t al.
his
tolo
g-
ical sc
ore
(R
utg
eert
s sc
ore
) [1
8]
i0: N
o lesi
on
si1
: L
ess
th
an
5 a
ph
tho
us
lesi
on
si2
: M
ore
th
an
5 a
ph
tho
us
lesi
on
s w
ith
no
rmal
mu
co
sa b
etw
een
th
e
lesi
on
s o
r sk
ip a
reas
of
larg
er
lesi
on
s o
r le
sio
ns
co
nfi
ned
to
ile
oco
lon
ic
an
ast
om
osi
s
i3: D
iffu
se a
ph
tho
us
ileit
is w
ith
dif
fuse
ly
infl
am
ed
mu
co
sa
i4: D
iffu
se in
flam
ma-
tio
n w
ith
alr
ead
y larg
e
ulc
ers
, n
od
ule
s, a
nd
/or
narr
ow
ing
Ho
rsth
uis
et
al.
en
do
-sc
op
ic s
co
re (
sub
jecti
ve)
[26
]
No
dis
ease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
Gir
om
ett
i et
al.
his
tolo
g-
ical sc
ore
(su
bje
cti
ve)
[25
]
No
dis
ease
(o
r ch
ron
ic,
qu
iesc
en
t d
isease
)M
ild d
isease
Mo
dera
te-t
o-s
evere
dis
ease
a. 0
–3 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
siz
e o
f u
lcers
(0
: N
on
e, 1:
Ap
hto
us
ulc
ers
(0
.1–0
.5 c
m),
2: L
arg
e u
lcers
(0
.5–2
cm
), 3
: V
ery
la
rge u
lcers
(>
2 c
m))
, u
lcera
ted
su
rface (
0: N
on
e, 1:
<10
%, 2: 10
–30
%, 3
: >
30
%),
aff
ecte
d s
urf
ace (
0: N
on
e, 1:
<5
0%
, 2: 5
0–7
5%
, 3
: >
75
%)
an
d p
rese
nce o
f n
arr
ow
-in
g (
0: N
on
e, 1:
Sin
gle
, can
be p
ass
ed
, 2: M
ult
iple
, can
be p
ass
ed
, 3
: C
an
no
t b
e p
ass
ed
).
Proefschrift2018-new2.indb 34 13/9/18 10:06
34
Chapter 2
Tab
le 8
. Ori
gin
al re
fere
nce t
est
cri
teri
a a
nd
cate
go
rizati
on
fo
r th
is s
tud
y
Stud
yN
one
Mild
Seve
re
Mao
et
al.
his
tolo
gic
al
sco
re (
Ru
tgeert
s sc
ore
) [1
7]
i0: N
o lesi
on
si1
: L
ess
th
an
5 a
ph
tho
us
lesi
on
si2
: M
ore
th
an
5 a
ph
tho
us
lesi
on
s w
ith
no
rmal
mu
co
sa b
etw
een
th
e
lesi
on
s o
r sk
ip a
reas
of
larg
er
lesi
on
s o
r le
sio
ns
co
nfi
ned
to
ile
oco
lon
ic
an
ast
om
osi
s
i3: D
iffu
se a
ph
tho
us
ileit
is w
ith
dif
fuse
ly
infl
am
ed
mu
co
sa
i4: D
iffu
se in
flam
ma-
tio
n w
ith
alr
ead
y larg
e
ulc
ers
, n
od
ule
s, a
nd
/or
narr
ow
ing
Mo
ham
ed
et
al.
his
tolo
g-
ical sc
ore
(su
bje
cti
ve)
[16
]
-M
ild
Mo
dera
teS
evere
Ko
lkm
an
et
al h
isto
log
i-cal sc
ore
[18
]0
: N
o a
bn
orm
alit
ies,
or
pla
in f
ibro
sis
1: S
om
e in
filt
rati
on
of
po
lym
orp
ho
nu
cle
ar
leu
ko
-cyte
s, n
o u
lcera
tio
n2: M
od
era
te in
filt
rati
on
o
f p
oly
mo
rph
on
ucle
ar
leu
ko
cyte
s, s
om
e u
lcer-
ati
on
pre
sen
t
3: S
evere
ly u
lcera
ted
w
ith
mass
ive in
filt
rati
on
o
f p
oly
mo
rph
on
ucle
ar
leu
ko
cyte
s
Sch
ill e
t al.
surg
ical sc
ore
(b
ase
d o
n M
on
treal cla
s-si
ficati
on
) [2
9]
-B
1: N
on
-str
ictu
rin
g a
nd
no
n-p
en
etr
ati
ng
B2: S
tric
turi
ng
B3
: P
en
etr
ati
ng
Galle
go
et
al.
en
do
sco
pic
sc
ore
(S
ES
-CD
)a
[28
]0
-2 p
oin
ts: In
acti
ve
3-6
po
ints
: m
ild d
isease
≥7 p
oin
ts: m
od
era
te/s
evere
dis
ease
Ko
ilako
u e
t al.
his
tolo
g-
ical sc
ore
(R
utg
eert
s sc
ore
) [1
8]
i0: N
o lesi
on
si1
: L
ess
th
an
5 a
ph
tho
us
lesi
on
si2
: M
ore
th
an
5 a
ph
tho
us
lesi
on
s w
ith
no
rmal
mu
co
sa b
etw
een
th
e
lesi
on
s o
r sk
ip a
reas
of
larg
er
lesi
on
s o
r le
sio
ns
co
nfi
ned
to
ile
oco
lon
ic
an
ast
om
osi
s
i3: D
iffu
se a
ph
tho
us
ileit
is w
ith
dif
fuse
ly
infl
am
ed
mu
co
sa
i4: D
iffu
se in
flam
ma-
tio
n w
ith
alr
ead
y larg
e
ulc
ers
, n
od
ule
s, a
nd
/or
narr
ow
ing
Ho
rsth
uis
et
al.
en
do
-sc
op
ic s
co
re (
sub
jecti
ve)
[26
]
No
dis
ease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
Gir
om
ett
i et
al.
his
tolo
g-
ical sc
ore
(su
bje
cti
ve)
[25
]
No
dis
ease
(o
r ch
ron
ic,
qu
iesc
en
t d
isease
)M
ild d
isease
Mo
dera
te-t
o-s
evere
dis
ease
a. 0
–3 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
siz
e o
f u
lcers
(0
: N
on
e, 1:
Ap
hto
us
ulc
ers
(0
.1–0
.5 c
m),
2: L
arg
e u
lcers
(0
.5–2
cm
), 3
: V
ery
la
rge u
lcers
(>
2 c
m))
, u
lcera
ted
su
rface (
0: N
on
e, 1:
<10
%, 2: 10
–30
%, 3
: >
30
%),
aff
ecte
d s
urf
ace (
0: N
on
e, 1:
<5
0%
, 2: 5
0–7
5%
, 3
: >
75
%)
an
d p
rese
nce o
f n
arr
ow
-in
g (
0: N
on
e, 1:
Sin
gle
, can
be p
ass
ed
, 2: M
ult
iple
, can
be p
ass
ed
, 3
: C
an
no
t b
e p
ass
ed
).
Proefschrift2018-new2.indb 34 13/9/18 10:06
35
Grading CD using CT, MRI, US and scintigraphy
Stud
yN
one
Mild
Seve
re
Ho
rsth
uis
et
al.
en
do
-sc
op
ic s
co
re (
sub
jecti
ve)
[24
]
No
dis
ease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
Flo
rie e
t al.
en
do
sco
pic
sc
ore
(su
bje
cti
ve)
[21]
No
dis
ease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
Sh
oen
ut
et
al.
en
do
sco
p-
ic s
co
re [
20
]-
Mild
: m
uco
sal ery
them
a, fr
iab
ility
an
d g
ran
ula
rity
Mo
dera
te: m
ark
ed
ed
em
a, lin
ear
or
patc
hy
ulc
era
tio
n
Severe
: C
oale
scin
g
ulc
era
tio
n, exu
dati
ve
co
litis
Sh
oen
ut
et
al.
his
tolo
g-
ical sc
ore
(su
bje
cti
ve)
[19
]
-M
ild d
isease
Mo
dera
te d
isease
Severe
dis
ease
Sch
reyer
et
al.
en
do
-sc
op
ic s
co
re [
22]
0: N
o f
ind
ing
s1:
Ery
them
a, d
ecre
ase
d o
r ab
sen
t vasc
ula
r p
att
ern
, fr
iab
ility
of
the m
uco
sa, si
ng
le o
r m
ult
iple
ap
hth
ou
s le
sio
ns,
an
d s
mall
ulc
ers
2: P
rese
nce o
f la
rge u
lcero
us
lesi
on
s, n
od
ule
s, a
nd
/o
r n
arr
ow
ing
Sch
reyer
et
al.
en
do
-sc
op
ic s
co
re [
23
]0
: N
o f
ind
ing
s1:
Ery
them
a, d
ecre
ase
d o
r ab
sen
t vasc
ula
r p
att
ern
, fr
iab
ility
of
the m
uco
sa, si
ng
le o
r m
ult
iple
ap
hth
ou
s le
sio
ns,
an
d s
mall
ulc
ers
2: P
rese
nce o
f sp
on
tan
eo
us
ble
ed
ing
, an
d larg
e
ulc
ero
us
lesi
on
s, n
od
ule
s, a
nd
/or
narr
ow
ing
Dre
ws
et
al.
his
tolo
gic
al
sco
re [
32]
0: N
o in
flam
mati
on
1: C
hro
nic
no
n-a
cti
ve
infl
am
mati
on
2: M
ild a
cti
ve in
flam
-m
ati
on
3: M
od
era
te a
cti
ve
infl
am
mati
on
4: S
evere
acti
ve in
flam
-m
ati
on
Neye e
t al.
en
do
sco
pic
sc
ore
[3
1]0
: N
o lesi
on
s1:
Ap
hte
s2: A
ph
tes
an
d u
lcers
<
50
%3
: A
ph
tes
an
d u
lcers
>
50
%
Bo
zku
rt e
t al.
his
tolo
gic
al
sco
re (
sub
jecti
ve)
[30
]0
12
Bia
nco
ne e
t al.
his
tolo
g-
ical sc
ore
(su
bje
cti
ve)
[34
]
01
23
Scia
rrett
a e
t al.
his
tolo
g-
ical sc
ore
(su
bje
cti
ve)
[33
]
01
2
a. 0
–3 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
siz
e o
f u
lcers
(0
: N
on
e, 1:
Ap
hto
us
ulc
ers
(0
.1–0
.5 c
m),
2: L
arg
e u
lcers
(0
.5–2
cm
), 3
: V
ery
la
rge u
lcers
(>
2 c
m))
, u
lcera
ted
su
rface (
0: N
on
e, 1:
<10
%, 2: 10
–30
%, 3
: >
30
%),
aff
ecte
d s
urf
ace (
0: N
on
e, 1:
<5
0%
, 2: 5
0–7
5%
, 3
: >
75
%)
an
d p
rese
nce o
f n
arr
ow
-in
g (
0: N
on
e, 1:
Sin
gle
, can
be p
ass
ed
, 2: M
ult
iple
, can
be p
ass
ed
, 3
: C
an
no
t b
e p
ass
ed
).
Proefschrift2018-new2.indb 35 13/9/18 10:06
35
Grading CD using CT, MRI, US and scintigraphy
Stud
yN
one
Mild
Seve
re
Ho
rsth
uis
et
al.
en
do
-sc
op
ic s
co
re (
sub
jecti
ve)
[24
]
No
dis
ease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
Flo
rie e
t al.
en
do
sco
pic
sc
ore
(su
bje
cti
ve)
[21]
No
dis
ease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
Sh
oen
ut
et
al.
en
do
sco
p-
ic s
co
re [
20
]-
Mild
: m
uco
sal ery
them
a, fr
iab
ility
an
d g
ran
ula
rity
Mo
dera
te: m
ark
ed
ed
em
a, lin
ear
or
patc
hy
ulc
era
tio
n
Severe
: C
oale
scin
g
ulc
era
tio
n, exu
dati
ve
co
litis
Sh
oen
ut
et
al.
his
tolo
g-
ical sc
ore
(su
bje
cti
ve)
[19
]
-M
ild d
isease
Mo
dera
te d
isease
Severe
dis
ease
Sch
reyer
et
al.
en
do
-sc
op
ic s
co
re [
22]
0: N
o f
ind
ing
s1:
Ery
them
a, d
ecre
ase
d o
r ab
sen
t vasc
ula
r p
att
ern
, fr
iab
ility
of
the m
uco
sa, si
ng
le o
r m
ult
iple
ap
hth
ou
s le
sio
ns,
an
d s
mall
ulc
ers
2: P
rese
nce o
f la
rge u
lcero
us
lesi
on
s, n
od
ule
s, a
nd
/o
r n
arr
ow
ing
Sch
reyer
et
al.
en
do
-sc
op
ic s
co
re [
23
]0
: N
o f
ind
ing
s1:
Ery
them
a, d
ecre
ase
d o
r ab
sen
t vasc
ula
r p
att
ern
, fr
iab
ility
of
the m
uco
sa, si
ng
le o
r m
ult
iple
ap
hth
ou
s le
sio
ns,
an
d s
mall
ulc
ers
2: P
rese
nce o
f sp
on
tan
eo
us
ble
ed
ing
, an
d larg
e
ulc
ero
us
lesi
on
s, n
od
ule
s, a
nd
/or
narr
ow
ing
Dre
ws
et
al.
his
tolo
gic
al
sco
re [
32]
0: N
o in
flam
mati
on
1: C
hro
nic
no
n-a
cti
ve
infl
am
mati
on
2: M
ild a
cti
ve in
flam
-m
ati
on
3: M
od
era
te a
cti
ve
infl
am
mati
on
4: S
evere
acti
ve in
flam
-m
ati
on
Neye e
t al.
en
do
sco
pic
sc
ore
[3
1]0
: N
o lesi
on
s1:
Ap
hte
s2: A
ph
tes
an
d u
lcers
<
50
%3
: A
ph
tes
an
d u
lcers
>
50
%
Bo
zku
rt e
t al.
his
tolo
gic
al
sco
re (
sub
jecti
ve)
[30
]0
12
Bia
nco
ne e
t al.
his
tolo
g-
ical sc
ore
(su
bje
cti
ve)
[34
]
01
23
Scia
rrett
a e
t al.
his
tolo
g-
ical sc
ore
(su
bje
cti
ve)
[33
]
01
2
a. 0
–3 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
siz
e o
f u
lcers
(0
: N
on
e, 1:
Ap
hto
us
ulc
ers
(0
.1–0
.5 c
m),
2: L
arg
e u
lcers
(0
.5–2
cm
), 3
: V
ery
la
rge u
lcers
(>
2 c
m))
, u
lcera
ted
su
rface (
0: N
on
e, 1:
<10
%, 2: 10
–30
%, 3
: >
30
%),
aff
ecte
d s
urf
ace (
0: N
on
e, 1:
<5
0%
, 2: 5
0–7
5%
, 3
: >
75
%)
an
d p
rese
nce o
f n
arr
ow
-in
g (
0: N
on
e, 1:
Sin
gle
, can
be p
ass
ed
, 2: M
ult
iple
, can
be p
ass
ed
, 3
: C
an
no
t b
e p
ass
ed
).
Proefschrift2018-new2.indb 35 13/9/18 10:06
36
Chapter 2
Tab
le 9
. Ori
gin
al im
ag
ing
cri
teri
a a
nd
cate
go
rizati
on
fo
r th
is s
tud
y
Stud
yN
one
Mild
Seve
re
Mao
et
al.
CT
sco
re
[17]
0: N
o f
ind
ing
s1:
Min
or
mu
co
sal ir
reg
ula
riti
es
wit
h s
ligh
t w
all
thic
ken
ing
an
d m
ura
l co
ntr
ast
en
han
cem
en
t2: M
uco
sal h
yp
erd
en
sity
wit
h
dis
tin
ct
bo
wel w
all
thic
ken
ing
, n
o s
ten
osi
s, o
r st
en
osi
s w
ith
ou
t p
rest
en
oti
c d
ilata
tio
n
3: M
ajo
r m
uco
sal ab
no
rmalit
ies,
d
isti
nct
bo
wel w
all
thic
ken
ing
w
ith
targ
et
sig
n a
nd
extr
avis
cera
l si
gn
s su
ch
as
peri
en
teri
c s
tran
d-
ing
, co
mb
sig
n, fi
bro
fatt
y p
rolif
-era
tio
n, st
en
osi
s w
ith
pre
sten
oti
c
dila
tati
on
an
d/o
r th
e p
rese
nce o
f co
mp
licati
on
s
Mo
ham
ed
et
al.
CT
sc
ore
[16
]-
Mild
: M
uco
sal h
yp
ere
nh
an
cem
en
tM
od
era
te: A
bn
orm
al m
uco
sal
en
han
cem
en
t an
d w
all
thic
ken
ing
(>
3 m
m)
Severe
: A
bn
orm
al m
uco
sal en
-h
an
cem
en
t, w
all
thic
ken
ing
(>
3
mm
) an
d o
ne o
r m
ore
extr
a-e
n-
teri
c m
an
ifest
ati
on
s (e
dem
a o
f th
e m
ese
nte
ric f
at,
en
go
rged
vasa
recta
, ly
mp
had
en
op
ath
y,
fist
ula
, ab
scess
)
Ko
lkm
an
et
al C
T
sco
re [
18]
0: N
o t
hic
ken
ing
of
the b
ow
-el w
all,
no
rmal m
ese
nte
ry1:
Th
icken
ed
bo
wel w
all,
ho
mo
gen
ou
s asp
ect,
no
en
-h
an
cem
en
t w
ith
in
traven
ou
s co
ntr
ast
, n
o d
ou
ble
-halo
si
gn
2: T
hic
ken
ed
bo
wel w
all,
en
han
cem
en
t w
ith
in
traven
ou
s co
ntr
ast
or
do
ub
le-h
alo
sig
n, u
l-cera
tio
n, o
r m
ese
nte
ric f
ibro
fatt
y
pro
lifera
tio
n
3: T
hic
ken
ed
bo
wel w
all,
en
-h
an
cem
en
t w
ith
in
traven
ou
s co
n-
trast
, u
lcera
tio
n, an
d m
ese
nte
ric
fib
rovasc
ula
r st
ran
ds
Ko
lkm
an
et
al sc
inti
-g
rap
hic
sco
re [
18]
0: N
o a
cti
vit
y1:
Up
take less
th
an
bo
ne
marr
ow
2: U
pta
ke e
qu
al to
bo
ne
marr
ow
3: U
pta
ke h
igh
er
than
bo
ne m
ar-
row
, b
ut
less
th
an
liv
er
4: U
pta
ke e
qu
al o
r h
igh
er
than
liv
er
Sch
ill e
t al.
MR
I sc
ore
(b
ase
d o
n M
on
treal
cla
ssif
icati
on
) [2
9]
-B
1: N
on
-str
ictu
rin
g a
nd
no
n-p
en
etr
ati
ng
B2: S
tric
turi
ng
B3
: P
en
etr
ati
ng
Galle
go
et
al.
MR
I sc
ore
a [
28
]0
–1 p
oin
ts: N
o d
isease
2–6
po
ints
: M
ild d
isease
≥7 p
oin
ts: M
od
era
te/s
evere
dis
ease
Ko
ilako
u e
t al.
MR
I sc
ore
[18
]0
: 1:
2:
3:
Ho
rsth
uis
et
al.
MR
I sc
ore
(su
bje
cti
ve)
[26
]N
o d
isease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
Gir
om
ett
i et
al.
MR
I sc
ore
b [
25
]0
–1 p
oin
ts: N
o d
isease
2–6
po
ints
: M
ild d
isease
≥7 p
oin
ts: M
od
era
te/s
evere
dis
ease
Ho
rsth
uis
et
al.
MR
I sc
ore
(su
bje
cti
ve)
[24
]N
o d
isease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
a. 0
–2 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
bo
wel w
all
thic
kn
ess
(0
: <
3 m
m, 1:
3–4
mm
, 2: >
4 m
m),
rela
tive e
nh
an
cem
en
t (0
: <
70
%, 1:
70
–10
0%
, 2: >
100
%),
m
oti
lity (
0: n
orm
al,
1: r
ed
uced
, 2: ab
sen
t), p
erc
en
tag
e s
ten
osi
s (0
: ≤6
0%
, 1:
>6
0%
), b
ow
el w
all
ed
em
a (
0: ab
sen
t, 1
: p
rese
nt)
, m
uco
sal ab
no
rmalit
ies
(0: ab
sen
t, 1
: p
rese
nt)
, ly
mp
h n
od
es
(0:
ab
sen
t, 1
: p
rese
nt)
, fi
stu
lae o
r si
nu
s tr
acts
(0
: ab
sen
t, 1
: p
rese
nt)
, in
flam
mato
ry m
ass
es
(0: ab
sen
t, 1
: p
rese
nt)
.b
. 0
–2 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
bo
wel w
all
thic
kn
ess
(0
: <
3 m
m, 1:
3–4
mm
, 2: >
4 m
m),
wall-
co
ntr
ast
en
han
cem
en
t (0
: <
70
%, 1:
70
–10
0%
, 2:
>10
0%
), p
erc
en
tag
e s
ten
osi
s (0
: <
50
%, 1:
50
–80
%, 2: >
80
%),
mu
co
sal ab
no
rmalit
ies
(0: ab
sen
t, 1
: p
rese
nt)
, la
yere
d w
all
en
han
cem
en
t (0
: ab
sen
t, 1
: p
rese
nt)
, p
eri
stals
is (
0: p
rese
nt,
1: ab
-se
nt)
, d
iste
nsi
bili
ty (
0: p
rese
nt,
1: ab
sen
t), m
ese
nte
ric invo
lvem
en
t (0
: ab
sen
t, 1
: p
rese
nt)
, p
ath
olo
gic
lym
ph
no
des
(n>
3)
(0: ab
sen
t, 1
: p
rese
nt)
, fi
stu
lae o
r si
nu
s tr
acts
(0
: ab
sen
t, 1
: p
rese
nt)
, in
flam
mato
ry m
ass
es
(0: ab
sen
t, 1
: p
rese
nt)
.
Proefschrift2018-new2.indb 36 13/9/18 10:06
36
Chapter 2
Tab
le 9
. Ori
gin
al im
ag
ing
cri
teri
a a
nd
cate
go
rizati
on
fo
r th
is s
tud
y
Stud
yN
one
Mild
Seve
re
Mao
et
al.
CT
sco
re
[17]
0: N
o f
ind
ing
s1:
Min
or
mu
co
sal ir
reg
ula
riti
es
wit
h s
ligh
t w
all
thic
ken
ing
an
d m
ura
l co
ntr
ast
en
han
cem
en
t2: M
uco
sal h
yp
erd
en
sity
wit
h
dis
tin
ct
bo
wel w
all
thic
ken
ing
, n
o s
ten
osi
s, o
r st
en
osi
s w
ith
ou
t p
rest
en
oti
c d
ilata
tio
n
3: M
ajo
r m
uco
sal ab
no
rmalit
ies,
d
isti
nct
bo
wel w
all
thic
ken
ing
w
ith
targ
et
sig
n a
nd
extr
avis
cera
l si
gn
s su
ch
as
peri
en
teri
c s
tran
d-
ing
, co
mb
sig
n, fi
bro
fatt
y p
rolif
-era
tio
n, st
en
osi
s w
ith
pre
sten
oti
c
dila
tati
on
an
d/o
r th
e p
rese
nce o
f co
mp
licati
on
s
Mo
ham
ed
et
al.
CT
sc
ore
[16
]-
Mild
: M
uco
sal h
yp
ere
nh
an
cem
en
tM
od
era
te: A
bn
orm
al m
uco
sal
en
han
cem
en
t an
d w
all
thic
ken
ing
(>
3 m
m)
Severe
: A
bn
orm
al m
uco
sal en
-h
an
cem
en
t, w
all
thic
ken
ing
(>
3
mm
) an
d o
ne o
r m
ore
extr
a-e
n-
teri
c m
an
ifest
ati
on
s (e
dem
a o
f th
e m
ese
nte
ric f
at,
en
go
rged
vasa
recta
, ly
mp
had
en
op
ath
y,
fist
ula
, ab
scess
)
Ko
lkm
an
et
al C
T
sco
re [
18]
0: N
o t
hic
ken
ing
of
the b
ow
-el w
all,
no
rmal m
ese
nte
ry1:
Th
icken
ed
bo
wel w
all,
ho
mo
gen
ou
s asp
ect,
no
en
-h
an
cem
en
t w
ith
in
traven
ou
s co
ntr
ast
, n
o d
ou
ble
-halo
si
gn
2: T
hic
ken
ed
bo
wel w
all,
en
han
cem
en
t w
ith
in
traven
ou
s co
ntr
ast
or
do
ub
le-h
alo
sig
n, u
l-cera
tio
n, o
r m
ese
nte
ric f
ibro
fatt
y
pro
lifera
tio
n
3: T
hic
ken
ed
bo
wel w
all,
en
-h
an
cem
en
t w
ith
in
traven
ou
s co
n-
trast
, u
lcera
tio
n, an
d m
ese
nte
ric
fib
rovasc
ula
r st
ran
ds
Ko
lkm
an
et
al sc
inti
-g
rap
hic
sco
re [
18]
0: N
o a
cti
vit
y1:
Up
take less
th
an
bo
ne
marr
ow
2: U
pta
ke e
qu
al to
bo
ne
marr
ow
3: U
pta
ke h
igh
er
than
bo
ne m
ar-
row
, b
ut
less
th
an
liv
er
4: U
pta
ke e
qu
al o
r h
igh
er
than
liv
er
Sch
ill e
t al.
MR
I sc
ore
(b
ase
d o
n M
on
treal
cla
ssif
icati
on
) [2
9]
-B
1: N
on
-str
ictu
rin
g a
nd
no
n-p
en
etr
ati
ng
B2: S
tric
turi
ng
B3
: P
en
etr
ati
ng
Galle
go
et
al.
MR
I sc
ore
a [
28
]0
–1 p
oin
ts: N
o d
isease
2–6
po
ints
: M
ild d
isease
≥7 p
oin
ts: M
od
era
te/s
evere
dis
ease
Ko
ilako
u e
t al.
MR
I sc
ore
[18
]0
: 1:
2:
3:
Ho
rsth
uis
et
al.
MR
I sc
ore
(su
bje
cti
ve)
[26
]N
o d
isease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
Gir
om
ett
i et
al.
MR
I sc
ore
b [
25
]0
–1 p
oin
ts: N
o d
isease
2–6
po
ints
: M
ild d
isease
≥7 p
oin
ts: M
od
era
te/s
evere
dis
ease
Ho
rsth
uis
et
al.
MR
I sc
ore
(su
bje
cti
ve)
[24
]N
o d
isease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
a. 0
–2 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
bo
wel w
all
thic
kn
ess
(0
: <
3 m
m, 1:
3–4
mm
, 2: >
4 m
m),
rela
tive e
nh
an
cem
en
t (0
: <
70
%, 1:
70
–10
0%
, 2: >
100
%),
m
oti
lity (
0: n
orm
al,
1: r
ed
uced
, 2: ab
sen
t), p
erc
en
tag
e s
ten
osi
s (0
: ≤6
0%
, 1:
>6
0%
), b
ow
el w
all
ed
em
a (
0: ab
sen
t, 1
: p
rese
nt)
, m
uco
sal ab
no
rmalit
ies
(0: ab
sen
t, 1
: p
rese
nt)
, ly
mp
h n
od
es
(0:
ab
sen
t, 1
: p
rese
nt)
, fi
stu
lae o
r si
nu
s tr
acts
(0
: ab
sen
t, 1
: p
rese
nt)
, in
flam
mato
ry m
ass
es
(0: ab
sen
t, 1
: p
rese
nt)
.b
. 0
–2 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
bo
wel w
all
thic
kn
ess
(0
: <
3 m
m, 1:
3–4
mm
, 2: >
4 m
m),
wall-
co
ntr
ast
en
han
cem
en
t (0
: <
70
%, 1:
70
–10
0%
, 2:
>10
0%
), p
erc
en
tag
e s
ten
osi
s (0
: <
50
%, 1:
50
–80
%, 2: >
80
%),
mu
co
sal ab
no
rmalit
ies
(0: ab
sen
t, 1
: p
rese
nt)
, la
yere
d w
all
en
han
cem
en
t (0
: ab
sen
t, 1
: p
rese
nt)
, p
eri
stals
is (
0: p
rese
nt,
1: ab
-se
nt)
, d
iste
nsi
bili
ty (
0: p
rese
nt,
1: ab
sen
t), m
ese
nte
ric invo
lvem
en
t (0
: ab
sen
t, 1
: p
rese
nt)
, p
ath
olo
gic
lym
ph
no
des
(n>
3)
(0: ab
sen
t, 1
: p
rese
nt)
, fi
stu
lae o
r si
nu
s tr
acts
(0
: ab
sen
t, 1
: p
rese
nt)
, in
flam
mato
ry m
ass
es
(0: ab
sen
t, 1
: p
rese
nt)
.
Proefschrift2018-new2.indb 36 13/9/18 10:06
37
Grading CD using CT, MRI, US and scintigraphy
Stud
yN
one
Mild
Seve
re
Flo
rie e
t al.
MR
I sc
ore
(s
ub
jecti
ve)
[21]
No
dis
ease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
Sh
oen
ut
et
al.
MR
I sc
ore
[20
]-
Mild
: ≤7
0%
co
ntr
ast
-en
han
cem
en
t in
th
e m
ost
dis
ease
d
seg
men
t (b
y w
all
thic
kn
ess
an
d len
gth
)M
od
era
te: 7
1–11
9%
co
ntr
ast
-en
-h
an
cem
en
t in
th
e m
ost
dis
ease
d
seg
men
t (b
y w
all
thic
kn
ess
an
d
len
gth
)
Severe
: ≥1
20
% c
on
trast
-en
-h
an
cem
en
t in
th
e m
ost
dis
ease
d
seg
men
t (b
y w
all
thic
kn
ess
an
d
len
gth
)
Sh
oen
ut
et
al.
MR
I sc
ore
[19
]-
Mild
: L
en
gth
of
dis
ease
d s
eg
men
t <
5 c
m, b
ow
el w
all
thic
kn
ess
< 5
mm
, co
ntr
ast
-en
han
cem
en
t <
50
%M
od
era
te: L
en
gth
of
dis
ease
d
seg
men
t >
5 c
m, b
ow
el w
all
thic
kn
ess
0.5
-1 c
m, co
ntr
ast
-en
-h
an
cem
en
t <
10
0%
Severe
: L
en
gth
of
dis
ease
d
seg
men
t >
5 c
m, b
ow
el w
all
thic
kn
ess
> 1
cm
, co
ntr
ast
-en
-h
an
cem
en
t >
10
0%
Sch
reyer
et
al.
MR
I sc
ore
[22]
0: N
o c
rite
ria
1: O
ne o
f th
e f
ollo
win
g c
rite
ria: b
ow
el w
all
thic
ken
-in
g, b
ow
el st
en
osi
s, in
cre
ase
d c
on
trast
med
ia u
pta
ke,
en
larg
ed
lo
cal ly
mp
h n
od
es
an
d lo
cal in
jecti
on
fo
r in
flam
-m
ati
on
ass
ess
men
t
2: Tw
o o
r m
ore
of
the f
ollo
win
g c
rite
ria: b
ow
el w
all
thic
ken
ing
, b
ow
el
sten
osi
s, in
cre
ase
d c
on
trast
med
ia u
pta
ke, en
larg
ed
lo
cal ly
mp
h
no
des
an
d lo
cal in
jecti
on
fo
r in
flam
mati
on
ass
ess
men
t
Sch
reyer
et
al.
MR
I sc
ore
[23
]0
: N
o c
rite
ria
1: O
ne o
f th
e f
ollo
win
g c
rite
ria: b
ow
el w
all
thic
ken
ing
wit
h
co
ntr
ast
en
han
cem
en
t en
larg
ed
lo
cal ly
mp
h n
od
es
an
d
mese
nte
ric in
jecti
on
2: Tw
o o
r m
ore
of
the f
ollo
win
g c
rite
ria: b
ow
el w
all
thic
ken
ing
wit
h
co
ntr
ast
en
han
cem
en
t en
larg
ed
lo
cal ly
mp
h n
od
es
an
d m
ese
nte
ric
inje
cti
on
Dre
ws
et
al.
US
sco
re
[32]
0: B
ow
el w
all
thic
kn
ess
3–4
m
m w
ith
pre
serv
ed
fiv
e-l
ayer
stru
ctu
re, n
o in
cre
ase
d
vasc
ula
rity
1: B
ow
el w
all
thic
kn
ess
>
4 m
m, n
o in
cre
ase
d
vasc
ula
rity
2: G
rad
e 1
plu
s sh
ort
st
retc
hes
of
incre
ase
d
vasc
ula
rity
3: G
rad
e 2
plu
s lo
ng
er
stre
tch
es
of
incre
ase
d v
asc
ula
rity
4: G
rad
e 3
plu
s vasc
ula
rity
exte
nd
ing
in
to s
urr
ou
nd
ing
m
ese
nte
ry
Neye e
t al.
US
sco
re
[31]
1: 0
vess
els
/cm
2 a
nd
bo
wel
wall
thic
kn
ess
< 5
mm
2: 0
vess
els
/cm
2 +
bo
wel w
all
thic
kn
ess
> 5
mm
or
1-2
vess
els
/cm
2 +
bo
wel w
all
thic
kn
ess
< 5
mm
3: 1–
2 v
ess
els
/cm
2 +
bo
wel w
all
thic
kn
ess
> 5
mm
or
> 2
vess
els
/cm
2 +
bo
wel w
all
thic
kn
ess
<
5 m
m
4: >
2 v
ess
els
/cm
2 +
bo
wel w
all
thic
kn
ess
> 5
mm
Bo
zku
rt e
t al.
US
sco
re
[30
] 0
: N
orm
al b
ow
el w
all
wit
h a
n
ech
o-p
oo
r la
yer
of ≤4
mm
w
ith
a s
mo
oth
bo
un
dary
. O
nly
th
e v
en
tral w
all
vis
ualiz
-ab
le t
o g
ase
ou
s d
iste
nti
on
1: B
ow
el w
all
thic
kn
ess
> 4
mm
wit
h in
div
idu
al la
yers
vis
ible
2: B
ow
el w
all
thic
kn
ess
wit
h p
oo
rly d
efi
ned
in
div
idu
al la
yers
an
d
decre
ase
d e
ch
og
en
icit
y
Bia
nco
ne e
t al.
scin
ti-
gra
ph
ic s
co
re [
34
]0
: N
o lab
elin
g1:
Less
th
an
bo
ne m
arr
ow
2: M
ore
th
an
bo
ne m
arr
ow
, le
ss
than
liv
er
3: E
qu
al o
r m
ore
th
an
liv
er
Scia
rrett
a e
t al.
his
tolo
gic
al sc
ore
(s
ub
jecti
ve)
[33
]
0: N
o u
pta
ke
1: L
ess
th
an
bo
ne m
arr
ow
2: M
ore
th
an
bo
ne m
arr
ow
, le
ss
than
liv
er
3: E
qu
al o
r m
ore
th
an
liv
er
a. 0
–2 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
bo
wel w
all
thic
kn
ess
(0
: <
3 m
m, 1:
3–4
mm
, 2: >
4 m
m),
rela
tive e
nh
an
cem
en
t (0
: <
70
%, 1:
70
–10
0%
, 2: >
100
%),
m
oti
lity (
0: n
orm
al,
1: r
ed
uced
, 2: ab
sen
t), p
erc
en
tag
e s
ten
osi
s (0
: ≤6
0%
, 1:
>6
0%
), b
ow
el w
all
ed
em
a (
0: ab
sen
t, 1
: p
rese
nt)
, m
uco
sal ab
no
rmalit
ies
(0: ab
sen
t, 1
: p
rese
nt)
, ly
mp
h n
od
es
(0:
ab
sen
t, 1
: p
rese
nt)
, fi
stu
lae o
r si
nu
s tr
acts
(0
: ab
sen
t, 1
: p
rese
nt)
, in
flam
mato
ry m
ass
es
(0: ab
sen
t, 1
: p
rese
nt)
.b
. 0
–2 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
bo
wel w
all
thic
kn
ess
(0
: <
3 m
m, 1:
3–4
mm
, 2: >
4 m
m),
wall-
co
ntr
ast
en
han
cem
en
t (0
: <
70
%, 1:
70
–10
0%
, 2:
>10
0%
), p
erc
en
tag
e s
ten
osi
s (0
: <
50
%, 1:
50
–80
%, 2: >
80
%),
mu
co
sal ab
no
rmalit
ies
(0: ab
sen
t, 1
: p
rese
nt)
, la
yere
d w
all
en
han
cem
en
t (0
: ab
sen
t, 1
: p
rese
nt)
, p
eri
stals
is (
0: p
rese
nt,
1: ab
-se
nt)
, d
iste
nsi
bili
ty (
0: p
rese
nt,
1: ab
sen
t), m
ese
nte
ric invo
lvem
en
t (0
: ab
sen
t, 1
: p
rese
nt)
, p
ath
olo
gic
lym
ph
no
des
(n>
3)
(0: ab
sen
t, 1
: p
rese
nt)
, fi
stu
lae o
r si
nu
s tr
acts
(0
: ab
sen
t, 1
: p
rese
nt)
, in
flam
mato
ry m
ass
es
(0: ab
sen
t, 1
: p
rese
nt)
.
Proefschrift2018-new2.indb 37 13/9/18 10:06
37
Grading CD using CT, MRI, US and scintigraphy
Stud
yN
one
Mild
Seve
re
Flo
rie e
t al.
MR
I sc
ore
(s
ub
jecti
ve)
[21]
No
dis
ease
Mild
dis
ease
Mo
dera
te d
isease
Severe
dis
ease
Sh
oen
ut
et
al.
MR
I sc
ore
[20
]-
Mild
: ≤7
0%
co
ntr
ast
-en
han
cem
en
t in
th
e m
ost
dis
ease
d
seg
men
t (b
y w
all
thic
kn
ess
an
d len
gth
)M
od
era
te: 7
1–11
9%
co
ntr
ast
-en
-h
an
cem
en
t in
th
e m
ost
dis
ease
d
seg
men
t (b
y w
all
thic
kn
ess
an
d
len
gth
)
Severe
: ≥1
20
% c
on
trast
-en
-h
an
cem
en
t in
th
e m
ost
dis
ease
d
seg
men
t (b
y w
all
thic
kn
ess
an
d
len
gth
)
Sh
oen
ut
et
al.
MR
I sc
ore
[19
]-
Mild
: L
en
gth
of
dis
ease
d s
eg
men
t <
5 c
m, b
ow
el w
all
thic
kn
ess
< 5
mm
, co
ntr
ast
-en
han
cem
en
t <
50
%M
od
era
te: L
en
gth
of
dis
ease
d
seg
men
t >
5 c
m, b
ow
el w
all
thic
kn
ess
0.5
-1 c
m, co
ntr
ast
-en
-h
an
cem
en
t <
10
0%
Severe
: L
en
gth
of
dis
ease
d
seg
men
t >
5 c
m, b
ow
el w
all
thic
kn
ess
> 1
cm
, co
ntr
ast
-en
-h
an
cem
en
t >
10
0%
Sch
reyer
et
al.
MR
I sc
ore
[22]
0: N
o c
rite
ria
1: O
ne o
f th
e f
ollo
win
g c
rite
ria: b
ow
el w
all
thic
ken
-in
g, b
ow
el st
en
osi
s, in
cre
ase
d c
on
trast
med
ia u
pta
ke,
en
larg
ed
lo
cal ly
mp
h n
od
es
an
d lo
cal in
jecti
on
fo
r in
flam
-m
ati
on
ass
ess
men
t
2: Tw
o o
r m
ore
of
the f
ollo
win
g c
rite
ria: b
ow
el w
all
thic
ken
ing
, b
ow
el
sten
osi
s, in
cre
ase
d c
on
trast
med
ia u
pta
ke, en
larg
ed
lo
cal ly
mp
h
no
des
an
d lo
cal in
jecti
on
fo
r in
flam
mati
on
ass
ess
men
t
Sch
reyer
et
al.
MR
I sc
ore
[23
]0
: N
o c
rite
ria
1: O
ne o
f th
e f
ollo
win
g c
rite
ria: b
ow
el w
all
thic
ken
ing
wit
h
co
ntr
ast
en
han
cem
en
t en
larg
ed
lo
cal ly
mp
h n
od
es
an
d
mese
nte
ric in
jecti
on
2: Tw
o o
r m
ore
of
the f
ollo
win
g c
rite
ria: b
ow
el w
all
thic
ken
ing
wit
h
co
ntr
ast
en
han
cem
en
t en
larg
ed
lo
cal ly
mp
h n
od
es
an
d m
ese
nte
ric
inje
cti
on
Dre
ws
et
al.
US
sco
re
[32]
0: B
ow
el w
all
thic
kn
ess
3–4
m
m w
ith
pre
serv
ed
fiv
e-l
ayer
stru
ctu
re, n
o in
cre
ase
d
vasc
ula
rity
1: B
ow
el w
all
thic
kn
ess
>
4 m
m, n
o in
cre
ase
d
vasc
ula
rity
2: G
rad
e 1
plu
s sh
ort
st
retc
hes
of
incre
ase
d
vasc
ula
rity
3: G
rad
e 2
plu
s lo
ng
er
stre
tch
es
of
incre
ase
d v
asc
ula
rity
4: G
rad
e 3
plu
s vasc
ula
rity
exte
nd
ing
in
to s
urr
ou
nd
ing
m
ese
nte
ry
Neye e
t al.
US
sco
re
[31]
1: 0
vess
els
/cm
2 a
nd
bo
wel
wall
thic
kn
ess
< 5
mm
2: 0
vess
els
/cm
2 +
bo
wel w
all
thic
kn
ess
> 5
mm
or
1-2
vess
els
/cm
2 +
bo
wel w
all
thic
kn
ess
< 5
mm
3: 1–
2 v
ess
els
/cm
2 +
bo
wel w
all
thic
kn
ess
> 5
mm
or
> 2
vess
els
/cm
2 +
bo
wel w
all
thic
kn
ess
<
5 m
m
4: >
2 v
ess
els
/cm
2 +
bo
wel w
all
thic
kn
ess
> 5
mm
Bo
zku
rt e
t al.
US
sco
re
[30
] 0
: N
orm
al b
ow
el w
all
wit
h a
n
ech
o-p
oo
r la
yer
of ≤4
mm
w
ith
a s
mo
oth
bo
un
dary
. O
nly
th
e v
en
tral w
all
vis
ualiz
-ab
le t
o g
ase
ou
s d
iste
nti
on
1: B
ow
el w
all
thic
kn
ess
> 4
mm
wit
h in
div
idu
al la
yers
vis
ible
2: B
ow
el w
all
thic
kn
ess
wit
h p
oo
rly d
efi
ned
in
div
idu
al la
yers
an
d
decre
ase
d e
ch
og
en
icit
y
Bia
nco
ne e
t al.
scin
ti-
gra
ph
ic s
co
re [
34
]0
: N
o lab
elin
g1:
Less
th
an
bo
ne m
arr
ow
2: M
ore
th
an
bo
ne m
arr
ow
, le
ss
than
liv
er
3: E
qu
al o
r m
ore
th
an
liv
er
Scia
rrett
a e
t al.
his
tolo
gic
al sc
ore
(s
ub
jecti
ve)
[33
]
0: N
o u
pta
ke
1: L
ess
th
an
bo
ne m
arr
ow
2: M
ore
th
an
bo
ne m
arr
ow
, le
ss
than
liv
er
3: E
qu
al o
r m
ore
th
an
liv
er
a. 0
–2 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
bo
wel w
all
thic
kn
ess
(0
: <
3 m
m, 1:
3–4
mm
, 2: >
4 m
m),
rela
tive e
nh
an
cem
en
t (0
: <
70
%, 1:
70
–10
0%
, 2: >
100
%),
m
oti
lity (
0: n
orm
al,
1: r
ed
uced
, 2: ab
sen
t), p
erc
en
tag
e s
ten
osi
s (0
: ≤6
0%
, 1:
>6
0%
), b
ow
el w
all
ed
em
a (
0: ab
sen
t, 1
: p
rese
nt)
, m
uco
sal ab
no
rmalit
ies
(0: ab
sen
t, 1
: p
rese
nt)
, ly
mp
h n
od
es
(0:
ab
sen
t, 1
: p
rese
nt)
, fi
stu
lae o
r si
nu
s tr
acts
(0
: ab
sen
t, 1
: p
rese
nt)
, in
flam
mato
ry m
ass
es
(0: ab
sen
t, 1
: p
rese
nt)
.b
. 0
–2 p
oin
ts a
re g
iven
fo
r th
e f
ollo
win
g a
re g
iven
to
fo
llow
ing
featu
res:
bo
wel w
all
thic
kn
ess
(0
: <
3 m
m, 1:
3–4
mm
, 2: >
4 m
m),
wall-
co
ntr
ast
en
han
cem
en
t (0
: <
70
%, 1:
70
–10
0%
, 2:
>10
0%
), p
erc
en
tag
e s
ten
osi
s (0
: <
50
%, 1:
50
–80
%, 2: >
80
%),
mu
co
sal ab
no
rmalit
ies
(0: ab
sen
t, 1
: p
rese
nt)
, la
yere
d w
all
en
han
cem
en
t (0
: ab
sen
t, 1
: p
rese
nt)
, p
eri
stals
is (
0: p
rese
nt,
1: ab
-se
nt)
, d
iste
nsi
bili
ty (
0: p
rese
nt,
1: ab
sen
t), m
ese
nte
ric invo
lvem
en
t (0
: ab
sen
t, 1
: p
rese
nt)
, p
ath
olo
gic
lym
ph
no
des
(n>
3)
(0: ab
sen
t, 1
: p
rese
nt)
, fi
stu
lae o
r si
nu
s tr
acts
(0
: ab
sen
t, 1
: p
rese
nt)
, in
flam
mato
ry m
ass
es
(0: ab
sen
t, 1
: p
rese
nt)
.
Proefschrift2018-new2.indb 37 13/9/18 10:06
38
Chapter 2
Table 10. Comparison table with results for imaging tests from the 3x3 data-analysis and
corresponding P-values
Accurate grading Overgrading Undergrading
Per-patient (13 data-sets)
CT (n=2) vs MRI (n=9) 0.86 vs 0.84 (P=0.8) 0.10 vs 0.09 (P=0.8) 0.03 vs 0.06 (P=0.5)
CT (n=2) vs US (n=1) 0.86 vs 0.44 (P=0.0001) 0.10 vs 0.25 (P=0.07) 0.03 vs 0.31 (P=0.002)
CT (n=2) vs SG (n=1) 0.86 vs 0.40 (P=0.003) 0.10 vs 0.10 (P=1.0) 0.03 vs 0.50 (P=0.0005)
MRI (n=9) vs US (n=1) 0.84 vs 0.44 (P=0.001) 0.09 vs 0.25 (P=0.03) 0.06 vs 0.31 (P=0.003)
MRI (n=9) vs SG (n=1) 0.84 vs 0.40 (P=0.01) 0.09 vs 0.10 (P=0.9) 0.06 vs 0.50 (P=0.001)
US (n=1) vs SG (n=1) 0.44 vs 0.40 (P=0.8) 0.25 vs 0.10 (P=0.3) 0.31 vs 0.50 (P=0.3)
Per-segment (3 data-sets) a
CT (n=1) vs SG (n=2) 0.87 vs 0.86 (P=0.8) 0.00 vs 0.04 (P=0.2) 0.13 vs 0.10 (P=0.5)
a. Data on MRI and US was not pooled and included in the comparison, as the data was too heteroge-neous (I2 > 75%).
Figure 4. Accurate grading, over- and under-grading per study on a per-patient and per-
segment basis
Study Index Accurate grading Overgrading Undergrading Accurate grading Overgrading Undergrading test Per-patient
Mao (14) CT 78% (61–89%) 19% (9–36%) 3% (0–19%) Mohamed (13) CT 96% (77–99%) 0% (0–28%) 4% (1–23%) Schil (26) MRI 99% (91–100%) 0% (0–10%) 1% (0–9%)Gallego (25) MRI 77% (65–86%) 16% (9–28%) 7% (3–16%)Koilakou (24) MRI 92% (74–98%) 8% (2–26%) 0% (0–28%)Horsthuis (23): ob1 MRI 47% (24–71%) 7% (1–35%) 47% (24–71%)Horsthuis (23): ob2 MRI 27% (10–53%) 7% (1–35%) 67% (41–85%)Horsthuis (23): ob3 MRI 67% (41–85%) 13% (3–41%) 20% (7–47%) Girometti (22) MRI 91% (79–97%) 4% (1–16%) 4% (1–16%)Horsthuis (21): ob1 MRI 55% (34–75%) 30% (14–53%) 15% (5–38%)Horsthuis (21): ob2 MRI 75% (52–89%) 10% (3–32%) 15% (5–38%)Florie (18): ob1 MRI 61% (43–77%) 26% (13–44%) 13% (5–30%)Florie (18): ob2 MRI 58% (40–74%) 26% (13–44%) 16% (7–33%)Shoenut (17) MRI 100% (58–100%) 0% (0–42%) 0% (0–42%)Shoenut (16) MRI 74% (50–89%) 21% (8–45%) 5% (1–29%)Drews (29) US 44% (28–61%) 25% (13–43%) 31% (18–49%)Biancone (31) SG 40% (16–70%) 10% (1–47%) 50% (22–78%)
Per-segment
Kolkman (15) CT 87% (77–93%) 0% (0–10%) 13% (7–23%)Schreyer (19) MRI 82% (76–87%) 2% (1–5%) 16% (11–22%)Schreyer (20) MRI 67% (56–76%) 0% (0–9%) 33% (24–44%)Neye (28) US 75% (67–82%) 8% (4–14%) 17% (11–24%)Bozkurt (27) US 56% (49–63%) 33% (27–40%) 10% (7–16%)Sciarretta (30) SG 86% (78–92%) 5% (2–11%) 9% (5–17%)Kolkman (15) SG 86% (75–92%) 3% (1–11%) 11% (6–21%)
CT = Computed Tomography, MRI = Magnetic Resonance Imaging, US = Ultrasound, SG = Scintigraphy, ob = observer
0 20 40 60 80 100 %
0 20 40 60 80 100 %
0 20 40 60 80 100 %
Proefschrift2018-new2.indb 38 13/9/18 10:06
38
Chapter 2
Table 10. Comparison table with results for imaging tests from the 3x3 data-analysis and
corresponding P-values
Accurate grading Overgrading Undergrading
Per-patient (13 data-sets)
CT (n=2) vs MRI (n=9) 0.86 vs 0.84 (P=0.8) 0.10 vs 0.09 (P=0.8) 0.03 vs 0.06 (P=0.5)
CT (n=2) vs US (n=1) 0.86 vs 0.44 (P=0.0001) 0.10 vs 0.25 (P=0.07) 0.03 vs 0.31 (P=0.002)
CT (n=2) vs SG (n=1) 0.86 vs 0.40 (P=0.003) 0.10 vs 0.10 (P=1.0) 0.03 vs 0.50 (P=0.0005)
MRI (n=9) vs US (n=1) 0.84 vs 0.44 (P=0.001) 0.09 vs 0.25 (P=0.03) 0.06 vs 0.31 (P=0.003)
MRI (n=9) vs SG (n=1) 0.84 vs 0.40 (P=0.01) 0.09 vs 0.10 (P=0.9) 0.06 vs 0.50 (P=0.001)
US (n=1) vs SG (n=1) 0.44 vs 0.40 (P=0.8) 0.25 vs 0.10 (P=0.3) 0.31 vs 0.50 (P=0.3)
Per-segment (3 data-sets) a
CT (n=1) vs SG (n=2) 0.87 vs 0.86 (P=0.8) 0.00 vs 0.04 (P=0.2) 0.13 vs 0.10 (P=0.5)
a. Data on MRI and US was not pooled and included in the comparison, as the data was too heteroge-neous (I2 > 75%).
Figure 4. Accurate grading, over- and under-grading per study on a per-patient and per-
segment basis
Study Index Accurate grading Overgrading Undergrading Accurate grading Overgrading Undergrading test Per-patient
Mao (14) CT 78% (61–89%) 19% (9–36%) 3% (0–19%) Mohamed (13) CT 96% (77–99%) 0% (0–28%) 4% (1–23%) Schil (26) MRI 99% (91–100%) 0% (0–10%) 1% (0–9%)Gallego (25) MRI 77% (65–86%) 16% (9–28%) 7% (3–16%)Koilakou (24) MRI 92% (74–98%) 8% (2–26%) 0% (0–28%)Horsthuis (23): ob1 MRI 47% (24–71%) 7% (1–35%) 47% (24–71%)Horsthuis (23): ob2 MRI 27% (10–53%) 7% (1–35%) 67% (41–85%)Horsthuis (23): ob3 MRI 67% (41–85%) 13% (3–41%) 20% (7–47%) Girometti (22) MRI 91% (79–97%) 4% (1–16%) 4% (1–16%)Horsthuis (21): ob1 MRI 55% (34–75%) 30% (14–53%) 15% (5–38%)Horsthuis (21): ob2 MRI 75% (52–89%) 10% (3–32%) 15% (5–38%)Florie (18): ob1 MRI 61% (43–77%) 26% (13–44%) 13% (5–30%)Florie (18): ob2 MRI 58% (40–74%) 26% (13–44%) 16% (7–33%)Shoenut (17) MRI 100% (58–100%) 0% (0–42%) 0% (0–42%)Shoenut (16) MRI 74% (50–89%) 21% (8–45%) 5% (1–29%)Drews (29) US 44% (28–61%) 25% (13–43%) 31% (18–49%)Biancone (31) SG 40% (16–70%) 10% (1–47%) 50% (22–78%)
Per-segment
Kolkman (15) CT 87% (77–93%) 0% (0–10%) 13% (7–23%)Schreyer (19) MRI 82% (76–87%) 2% (1–5%) 16% (11–22%)Schreyer (20) MRI 67% (56–76%) 0% (0–9%) 33% (24–44%)Neye (28) US 75% (67–82%) 8% (4–14%) 17% (11–24%)Bozkurt (27) US 56% (49–63%) 33% (27–40%) 10% (7–16%)Sciarretta (30) SG 86% (78–92%) 5% (2–11%) 9% (5–17%)Kolkman (15) SG 86% (75–92%) 3% (1–11%) 11% (6–21%)
CT = Computed Tomography, MRI = Magnetic Resonance Imaging, US = Ultrasound, SG = Scintigraphy, ob = observer
0 20 40 60 80 100 %
0 20 40 60 80 100 %
0 20 40 60 80 100 %
Proefschrift2018-new2.indb 38 13/9/18 10:06
39
Grading CD using CT, MRI, US and scintigraphy
DISCUSSION
In this study, we have shown that MRI and CT are highly accurate for grading
Crohn’s disease activity. These findings are important, as cross-sectional imaging
plays an increasing role in the assessment of Crohn’s disease activity, and there has
been ongoing debate regarding the modality that should be the preferred choice
[35–37]. Several studies have compared two or more modalities in the same patient
group [38–41], but they have had relatively small sample sizes or only evaluated the
terminal ileum.
CT and MRI showed similar accuracy in grading Crohn’s disease activity (86% and
84% on a per-patient basis, respectively), and no significant differences in accuracy
were seen between these two modalities. Data on over- and under- grading
showed similar results for CT and MRI, further strengthening our conclusion of their
comparability. Scintigraphy showed high accuracy of 86% and 86% for the studies
using per-segment data, while accuracy of 40% was reported in per-patient data.
However, per-patient data for scintigraphy was reported in only one study, and
with a small sample size (n = 10) [34]. Furthermore, scintigraphy had the least
number of included patients (n = 58) in our meta-analysis. US showed low accuracy
of 44% in the per-patient data and 75% and 56% for studies in the per-segment
data. However, a relatively small number of patients (n = 86) were included. In
addition, no eligible studies evaluated luminal or intravenous contrast medium
for US. The use of intravenous contrast appears to be a particularly promising
technique, and may increase the accuracy of US. However, no robust reference
standard or appropriate grading scale were used in these studies. We considered
the possibility of performing subgroup and covariate analyses on the differences
in technique, imaging criteria, reference methods and methodological criteria, but
the results of these analyses would not be meaningful given the limited amount of
available data. We examined MRI imaging features in three studies with the highest
accuracy values. The following MRI features were used by at least two of these
studies: bowel wall thickness, T1 enhancement and pattern, T2 mural signal intensity,
mucosal abnormalities, presence of inflammatory mass, stenosis (with pre-stenotic
dilatation), lymph nodes, abscesses, and fistulas [25, 27, 29].
The observed heterogeneity of the grading criteria for the index and reference tests
in the studies that we included, our adjustment to construct 3 x 3 tables, and the
Proefschrift2018-new2.indb 39 13/9/18 10:06
39
Grading CD using CT, MRI, US and scintigraphy
DISCUSSION
In this study, we have shown that MRI and CT are highly accurate for grading
Crohn’s disease activity. These findings are important, as cross-sectional imaging
plays an increasing role in the assessment of Crohn’s disease activity, and there has
been ongoing debate regarding the modality that should be the preferred choice
[35–37]. Several studies have compared two or more modalities in the same patient
group [38–41], but they have had relatively small sample sizes or only evaluated the
terminal ileum.
CT and MRI showed similar accuracy in grading Crohn’s disease activity (86% and
84% on a per-patient basis, respectively), and no significant differences in accuracy
were seen between these two modalities. Data on over- and under- grading
showed similar results for CT and MRI, further strengthening our conclusion of their
comparability. Scintigraphy showed high accuracy of 86% and 86% for the studies
using per-segment data, while accuracy of 40% was reported in per-patient data.
However, per-patient data for scintigraphy was reported in only one study, and
with a small sample size (n = 10) [34]. Furthermore, scintigraphy had the least
number of included patients (n = 58) in our meta-analysis. US showed low accuracy
of 44% in the per-patient data and 75% and 56% for studies in the per-segment
data. However, a relatively small number of patients (n = 86) were included. In
addition, no eligible studies evaluated luminal or intravenous contrast medium
for US. The use of intravenous contrast appears to be a particularly promising
technique, and may increase the accuracy of US. However, no robust reference
standard or appropriate grading scale were used in these studies. We considered
the possibility of performing subgroup and covariate analyses on the differences
in technique, imaging criteria, reference methods and methodological criteria, but
the results of these analyses would not be meaningful given the limited amount of
available data. We examined MRI imaging features in three studies with the highest
accuracy values. The following MRI features were used by at least two of these
studies: bowel wall thickness, T1 enhancement and pattern, T2 mural signal intensity,
mucosal abnormalities, presence of inflammatory mass, stenosis (with pre-stenotic
dilatation), lymph nodes, abscesses, and fistulas [25, 27, 29].
The observed heterogeneity of the grading criteria for the index and reference tests
in the studies that we included, our adjustment to construct 3 x 3 tables, and the
Proefschrift2018-new2.indb 39 13/9/18 10:06
40
Chapter 2
differences in available data between imaging modalities were the major limitations
of this meta-analysis. Although the grading criteria for index and reference tests
differed by study, and different imaging features were used, the studies included
showed considerable overlap in the use of imaging features and grading criteria.
No generally accepted scoring systems exist for imaging of Crohn’s disease. To
construct 3 x 3 tables from original 4 x 4 data, we merged moderate and severe
disease into one group. Our decision to merge these grades was based on five
articles [22, 23, 25, 28, 30] that had originally used 3 x 3 tables; two of these studies
explicitly stated that their highest grade represented moderate and severe disease
combined [25, 28]. The remaining three studies [22, 23, 30] used similar grading
criteria. Another limitation was the heterogeneity of grading results, which we
examined using I2 statistics. Following those results, some of the datasets could not
be pooled. In our conclusions, we took into account the greater availability of data
for MRI compared to CT, US and scintigraphy. Furthermore, US and scintigraphy
studies showed varying results, hampering our ability to arrive at a firm conclusion.
There was only one head-to-head comparison study, which compared CT and
scintigraphy in 17 patients [18].
We selected three reference standards for this meta-analysis [35]. Intraoperative
findings served as the gold standard for assessing Crohn’s disease. We also included
endoscopy and endoscopic biopsies as reference standards, although they are not
ideal, as they are incapable of assessing proximal ileum, jejunum and extraluminal
disease, which could have led to incorrect estimation of disease activity. On the
other hand, surgery is performed only in select patients, whereas endoscopy is
applied across a wider spectrum. For our analysis, we gave precedence to results
from biopsies over endoscopic results, but we recognize that this was a controversial
choice, as there is no widespread consensus on which is the better reference
standard. The number of studies included could have been increased if VCE and/
or double-balloon enteroscopy (DBE) were also used as a reference standard. We
chose not to include these studies because interpretation of VCE and DBE has
not yet been standardized, and so this would further increase heterogeneity in our
study.
A growing number of studies are using correlative statistics to examine quantitative
scoring systems [42]. Because we used an ordinal outcome measure, we could not
include these studies. Nevertheless, a meta-analysis focused on this type of data
Proefschrift2018-new2.indb 40 13/9/18 10:06
40
Chapter 2
differences in available data between imaging modalities were the major limitations
of this meta-analysis. Although the grading criteria for index and reference tests
differed by study, and different imaging features were used, the studies included
showed considerable overlap in the use of imaging features and grading criteria.
No generally accepted scoring systems exist for imaging of Crohn’s disease. To
construct 3 x 3 tables from original 4 x 4 data, we merged moderate and severe
disease into one group. Our decision to merge these grades was based on five
articles [22, 23, 25, 28, 30] that had originally used 3 x 3 tables; two of these studies
explicitly stated that their highest grade represented moderate and severe disease
combined [25, 28]. The remaining three studies [22, 23, 30] used similar grading
criteria. Another limitation was the heterogeneity of grading results, which we
examined using I2 statistics. Following those results, some of the datasets could not
be pooled. In our conclusions, we took into account the greater availability of data
for MRI compared to CT, US and scintigraphy. Furthermore, US and scintigraphy
studies showed varying results, hampering our ability to arrive at a firm conclusion.
There was only one head-to-head comparison study, which compared CT and
scintigraphy in 17 patients [18].
We selected three reference standards for this meta-analysis [35]. Intraoperative
findings served as the gold standard for assessing Crohn’s disease. We also included
endoscopy and endoscopic biopsies as reference standards, although they are not
ideal, as they are incapable of assessing proximal ileum, jejunum and extraluminal
disease, which could have led to incorrect estimation of disease activity. On the
other hand, surgery is performed only in select patients, whereas endoscopy is
applied across a wider spectrum. For our analysis, we gave precedence to results
from biopsies over endoscopic results, but we recognize that this was a controversial
choice, as there is no widespread consensus on which is the better reference
standard. The number of studies included could have been increased if VCE and/
or double-balloon enteroscopy (DBE) were also used as a reference standard. We
chose not to include these studies because interpretation of VCE and DBE has
not yet been standardized, and so this would further increase heterogeneity in our
study.
A growing number of studies are using correlative statistics to examine quantitative
scoring systems [42]. Because we used an ordinal outcome measure, we could not
include these studies. Nevertheless, a meta-analysis focused on this type of data
Proefschrift2018-new2.indb 40 13/9/18 10:06
41
Grading CD using CT, MRI, US and scintigraphy
would be very useful. Finally, only patients with suspected IBD or known Crohn’s
disease were included, possibly introducing observer bias, leading to overgrading
of disease activity.
Assessment of study quality using the QUADAS tool showed overall moderate
quality of the studies included in this meta-analysis. The domains of reference
test and patient flow showed the highest risk of bias, while patient selection and
index test domains showed the lowest. Concern about the applicability of patient
selection and index and reference tests was generally low.
Recently, Vermeire et al. stated that MR enterography had become the reference
standard for assessing small and large bowel disease activity [43]. Based on our
results, we can agree with this statement. Considering the radiation exposure from
CT, it is not appropriate for repeated examinations, even with present-day reduced
ionizing radiation exposure per examination, although it still has an important role
in the acute setting [44]. Compared to endoscopy, MRI is non-invasive and able to
investigate trans- and extramural disease, making it possible to evaluate both the
small bowel and colon in one examination. Steps are being taken to come to a more
uniform evaluation of MRI in Crohn’s disease, which may improve accuracy [42,
45]. Furthermore, the versatility of MRI may be advantageous with new sequences
being studied.
In conclusion, CT and MRI can both be used to grade disease activity in Crohn’s
disease, while no conclusions can be made on US and scintigraphy due to the
limited and inconsistent data.
Proefschrift2018-new2.indb 41 13/9/18 10:06
41
Grading CD using CT, MRI, US and scintigraphy
would be very useful. Finally, only patients with suspected IBD or known Crohn’s
disease were included, possibly introducing observer bias, leading to overgrading
of disease activity.
Assessment of study quality using the QUADAS tool showed overall moderate
quality of the studies included in this meta-analysis. The domains of reference
test and patient flow showed the highest risk of bias, while patient selection and
index test domains showed the lowest. Concern about the applicability of patient
selection and index and reference tests was generally low.
Recently, Vermeire et al. stated that MR enterography had become the reference
standard for assessing small and large bowel disease activity [43]. Based on our
results, we can agree with this statement. Considering the radiation exposure from
CT, it is not appropriate for repeated examinations, even with present-day reduced
ionizing radiation exposure per examination, although it still has an important role
in the acute setting [44]. Compared to endoscopy, MRI is non-invasive and able to
investigate trans- and extramural disease, making it possible to evaluate both the
small bowel and colon in one examination. Steps are being taken to come to a more
uniform evaluation of MRI in Crohn’s disease, which may improve accuracy [42,
45]. Furthermore, the versatility of MRI may be advantageous with new sequences
being studied.
In conclusion, CT and MRI can both be used to grade disease activity in Crohn’s
disease, while no conclusions can be made on US and scintigraphy due to the
limited and inconsistent data.
Proefschrift2018-new2.indb 41 13/9/18 10:06
42
Chapter 2
REFERENCES1. Horsthuis K, Bipat S, Bennink RJ, Stoker J (2008) Inflammatory bowel disease diagnosed
with US, MR, scintigraphy, and CT: meta-analysis of prospective studies. Radiology
247:64–79
2. Panes J, Bouzas R, Chaparro M et al (2011) Systematic review: the use of ultrasonography,
computed tomography and magnetic resonance imaging for the diagnosis, assessment
of activity and abdominal complications of Crohn’s disease. Aliment Pharmacol Ther
34:125–145
3. Travis SP, Stange EF, Lemann M et al (2006) European evidence based consensus on the
diagnosis and management of Crohn’s dis- ease: current management. Gut 55:i16–i35
4. Hommes DW, van Deventer SJ (2004) Endoscopy in inflammatory bowel diseases.
Gastroenterology 126:1561–1573
5. Fletcher JG, Fidler JL, Bruining DH, Huprich JE (2011) New con- cepts in intestinal imaging
for inflammatory bowel diseases. Gastroenterology 140:1795–1806
6. Rimola J, Ordas I, Rodriguez S, Ricart E, Panes J (2012) Imaging indexes of activity and
severity for Crohn’s disease: current status and future trends. Abdom Imaging 37:958–966
7. Horsthuis K, Bipat S, Stokkers PC, Stoker J (2009) Magnetic res- onance imaging for
evaluation of disease activity in Crohn’s disease: a systematic review. Eur Radiol 19:1450–
1460
8. Moher D, Liberati A, Tetzlaff J, Altman DG, The PG (2009) Preferred reporting items for
systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 6, e1000097
9. Whiting PF, Rutjes AW, Westwood ME et al (2011) QUADAS-2: a revised tool for the quality
assessment of diagnostic accuracy stud- ies. Ann Intern Med 155:529–536
10. Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J (2003) The development of
QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in
systematic reviews. BMC Med Res Methodol 3:25
11. Deeks JJ, Macaskill P, Irwig L (2005) The performance of tests of publication bias and
other sample size effects in systematic reviews of diagnostic test accuracy was assessed.
J Clin Epidemiol 58:882–893
12. Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Measuring inconsistency in meta-
analyses. BMJ 327:557–560
13. Akaike H (1974) A new look at the statistical model identification. IEEE Trans Autom
Control 19:716–723
14. Bipat S, Zwinderman AH, Bossuyt PM, Stoker J (2007) Multivariate random-effects
approach: for meta-analysis of cancer staging studies. Acad Radiol 14:974–984
15. Glas AS, Lijmer JG, Prins MH, Bonsel GJ, Bossuyt PMM (2003) The diagnostic odds ratio:
a single indicator of test performance. J Clin Epidemiol 56:1129–1135
16. Mohamed AM, Amin SK, El-Shinnawy MA, Elfouly A, Baki AH (2012) Role of CT
enterography in assessment of Crohn’s disease activity: correlation with histopathologic
diagnosis. Egypt J Radiol Nucl Med 43:353–359
17. Mao R, Gao X, Zhu ZH et al (2013) CT enterography in evaluating postoperative recurrence
of Crohn’s disease after ileocolic resec- tion: complementary role to endoscopy. Inflamm
Bowel Dis 19: 977–982
18. Kolkman JJ, Falke TH, Roos JC et al (1996) Computed tomogra- phy and granulocyte
scintigraphy in active inflammatory bowel disease. Comparison with endoscopy and
operative findings. Dig Dis Sci 41:641–650
Proefschrift2018-new2.indb 42 13/9/18 10:06
42
Chapter 2
REFERENCES1. Horsthuis K, Bipat S, Bennink RJ, Stoker J (2008) Inflammatory bowel disease diagnosed
with US, MR, scintigraphy, and CT: meta-analysis of prospective studies. Radiology
247:64–79
2. Panes J, Bouzas R, Chaparro M et al (2011) Systematic review: the use of ultrasonography,
computed tomography and magnetic resonance imaging for the diagnosis, assessment
of activity and abdominal complications of Crohn’s disease. Aliment Pharmacol Ther
34:125–145
3. Travis SP, Stange EF, Lemann M et al (2006) European evidence based consensus on the
diagnosis and management of Crohn’s dis- ease: current management. Gut 55:i16–i35
4. Hommes DW, van Deventer SJ (2004) Endoscopy in inflammatory bowel diseases.
Gastroenterology 126:1561–1573
5. Fletcher JG, Fidler JL, Bruining DH, Huprich JE (2011) New con- cepts in intestinal imaging
for inflammatory bowel diseases. Gastroenterology 140:1795–1806
6. Rimola J, Ordas I, Rodriguez S, Ricart E, Panes J (2012) Imaging indexes of activity and
severity for Crohn’s disease: current status and future trends. Abdom Imaging 37:958–966
7. Horsthuis K, Bipat S, Stokkers PC, Stoker J (2009) Magnetic res- onance imaging for
evaluation of disease activity in Crohn’s disease: a systematic review. Eur Radiol 19:1450–
1460
8. Moher D, Liberati A, Tetzlaff J, Altman DG, The PG (2009) Preferred reporting items for
systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 6, e1000097
9. Whiting PF, Rutjes AW, Westwood ME et al (2011) QUADAS-2: a revised tool for the quality
assessment of diagnostic accuracy stud- ies. Ann Intern Med 155:529–536
10. Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J (2003) The development of
QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in
systematic reviews. BMC Med Res Methodol 3:25
11. Deeks JJ, Macaskill P, Irwig L (2005) The performance of tests of publication bias and
other sample size effects in systematic reviews of diagnostic test accuracy was assessed.
J Clin Epidemiol 58:882–893
12. Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Measuring inconsistency in meta-
analyses. BMJ 327:557–560
13. Akaike H (1974) A new look at the statistical model identification. IEEE Trans Autom
Control 19:716–723
14. Bipat S, Zwinderman AH, Bossuyt PM, Stoker J (2007) Multivariate random-effects
approach: for meta-analysis of cancer staging studies. Acad Radiol 14:974–984
15. Glas AS, Lijmer JG, Prins MH, Bonsel GJ, Bossuyt PMM (2003) The diagnostic odds ratio:
a single indicator of test performance. J Clin Epidemiol 56:1129–1135
16. Mohamed AM, Amin SK, El-Shinnawy MA, Elfouly A, Baki AH (2012) Role of CT
enterography in assessment of Crohn’s disease activity: correlation with histopathologic
diagnosis. Egypt J Radiol Nucl Med 43:353–359
17. Mao R, Gao X, Zhu ZH et al (2013) CT enterography in evaluating postoperative recurrence
of Crohn’s disease after ileocolic resec- tion: complementary role to endoscopy. Inflamm
Bowel Dis 19: 977–982
18. Kolkman JJ, Falke TH, Roos JC et al (1996) Computed tomogra- phy and granulocyte
scintigraphy in active inflammatory bowel disease. Comparison with endoscopy and
operative findings. Dig Dis Sci 41:641–650
Proefschrift2018-new2.indb 42 13/9/18 10:06
43
Grading CD using CT, MRI, US and scintigraphy
19. Shoenut JP, Semelka RC, Silverman R, Yaffe CS, Micflikier AB (1993) Magnetic resonance
imaging in inflammatory bowel dis- ease. J Clin Gastroenterol 17:73–78
20. Shoenut JP, Semelka RC, Magro CM, Silverman R, Yaffe CS, Micflikier AB (1994) Comparison
of magnetic resonance imaging and endoscopy in distinguishing the type and severity of
inflamma- tory bowel disease. J Clin Gastroenterol 19:31–35
21. Florie J, Horsthuis K, Hommes DW et al (2005) Magnetic reso- nance imaging compared
with ileocolonoscopy in evaluating dis- ease severity in Crohn’s disease. Clin Gastroenterol
Hepatol 3: 1221–1228
22. Schreyer AG, Golder S, Scheibl K et al (2005) Dark lumen mag- netic resonance
enteroclysis in combination with MRI colonography for whole bowel assessment in
patients with Crohn’s disease: first clinical experience. Inflamm Bowel Dis 11: 388–394
23. Schreyer AG, Rath HC, Kikinis R et al (2005) Comparison of magnetic resonance
imaging colonography with conventional co- lonoscopy for the assessment of intestinal
inflammation in patients with inflammatory bowel disease: a feasibility study. Gut 54:
250–256
24. van Gemert-Horsthuis K, Florie J, Hommes DW et al (2006) Feasibility of evaluating
Crohn’s disease activity at 3.0 Tesla. J Magn Reson Imaging 24:340–348
25. Girometti R, Zuiani C, Toso F et al (2008) MRI scoring system including dynamic motility
evaluation in assessing the activity of Crohn’s disease of the terminal ileum. Acad Radiol
15:153–164
26. Horsthuis K, de Ridder L, Smets AM et al (2010) Magnetic reso- nance enterography for
suspected inflammatory bowel disease in a pediatric population. J Pediatr Gastroenterol
Nutr 51:603–609
27. Koilakou S, Sailer J, Peloschek P et al (2010) Endoscopy and MR enteroclysis: equivalent
tools in predicting clinical recurrence in patients with Crohn’s disease after ileocolic
resection. Inflamm Bowel Dis 16:198–203
28. Gallego JC, Echarri AI, Porta A, Ollero V (2011) Ileal Crohn’s disease: MRI with endoscopic
correlation. Eur J Radiol 80:e8–e12
29. Schill G, Iesalnieks I, Haimerl M et al (2013) Assessment of disease behavior in patients
with Crohn’s disease by MR enterography. Inflamm Bowel Dis 19:983–990
30. Bozkurt T, Rommel T, Stabenow-Lohbauer U, Langer M, Schmiegelow P, Lux G (1996)
Sonographic bowel wall morphol- ogy correlates with clinical and endoscopic activity in
Crohn’s disease and ulcerative colitis. Eur J Ultrasound 4:27–33
31. Neye H, Voderholzer W, Rickes S, Weber J, Wermke W, Lochs H (2004) Evaluation of
criteria for the activity of Crohn’s disease by power Doppler sonography. Dig Dis 22:67–72
32. Drews BH, Barth TF, Hanle MM et al (2009) Comparison of sonographically measured
bowel wall vascularity, histology, and disease activity in Crohn’s disease. Eur Radiol
19:1379–1386
33. Sciarretta G, Furno A, Mazzoni M, Basile C, Malaguti P (1993) Technetium-99 m hexamethyl
propylene amine oxime granulocyte scintigraphy in Crohn’s disease: diagnostic and
clinical relevance. Gut 34:1364–1369
34. Biancone L, Scopinaro F, Ierardi M et al (1997) 99mTc-HMPAO granulocyte scintigraphy
in the early detection of postoperative asymptomatic recurrence in Crohn’s disease. Dig
Dis Sci 42: 1549–1556
35. Panes J, Bouhnik Y, Reinisch W et al (2013) Imaging techniques for assessment of
inflammatory bowel disease: joint ECCO and ESGAR evidence-based consensus
guidelines. J Crohns Colitis 7: 556–585
Proefschrift2018-new2.indb 43 13/9/18 10:06
43
Grading CD using CT, MRI, US and scintigraphy
19. Shoenut JP, Semelka RC, Silverman R, Yaffe CS, Micflikier AB (1993) Magnetic resonance
imaging in inflammatory bowel dis- ease. J Clin Gastroenterol 17:73–78
20. Shoenut JP, Semelka RC, Magro CM, Silverman R, Yaffe CS, Micflikier AB (1994) Comparison
of magnetic resonance imaging and endoscopy in distinguishing the type and severity of
inflamma- tory bowel disease. J Clin Gastroenterol 19:31–35
21. Florie J, Horsthuis K, Hommes DW et al (2005) Magnetic reso- nance imaging compared
with ileocolonoscopy in evaluating dis- ease severity in Crohn’s disease. Clin Gastroenterol
Hepatol 3: 1221–1228
22. Schreyer AG, Golder S, Scheibl K et al (2005) Dark lumen mag- netic resonance
enteroclysis in combination with MRI colonography for whole bowel assessment in
patients with Crohn’s disease: first clinical experience. Inflamm Bowel Dis 11: 388–394
23. Schreyer AG, Rath HC, Kikinis R et al (2005) Comparison of magnetic resonance
imaging colonography with conventional co- lonoscopy for the assessment of intestinal
inflammation in patients with inflammatory bowel disease: a feasibility study. Gut 54:
250–256
24. van Gemert-Horsthuis K, Florie J, Hommes DW et al (2006) Feasibility of evaluating
Crohn’s disease activity at 3.0 Tesla. J Magn Reson Imaging 24:340–348
25. Girometti R, Zuiani C, Toso F et al (2008) MRI scoring system including dynamic motility
evaluation in assessing the activity of Crohn’s disease of the terminal ileum. Acad Radiol
15:153–164
26. Horsthuis K, de Ridder L, Smets AM et al (2010) Magnetic reso- nance enterography for
suspected inflammatory bowel disease in a pediatric population. J Pediatr Gastroenterol
Nutr 51:603–609
27. Koilakou S, Sailer J, Peloschek P et al (2010) Endoscopy and MR enteroclysis: equivalent
tools in predicting clinical recurrence in patients with Crohn’s disease after ileocolic
resection. Inflamm Bowel Dis 16:198–203
28. Gallego JC, Echarri AI, Porta A, Ollero V (2011) Ileal Crohn’s disease: MRI with endoscopic
correlation. Eur J Radiol 80:e8–e12
29. Schill G, Iesalnieks I, Haimerl M et al (2013) Assessment of disease behavior in patients
with Crohn’s disease by MR enterography. Inflamm Bowel Dis 19:983–990
30. Bozkurt T, Rommel T, Stabenow-Lohbauer U, Langer M, Schmiegelow P, Lux G (1996)
Sonographic bowel wall morphol- ogy correlates with clinical and endoscopic activity in
Crohn’s disease and ulcerative colitis. Eur J Ultrasound 4:27–33
31. Neye H, Voderholzer W, Rickes S, Weber J, Wermke W, Lochs H (2004) Evaluation of
criteria for the activity of Crohn’s disease by power Doppler sonography. Dig Dis 22:67–72
32. Drews BH, Barth TF, Hanle MM et al (2009) Comparison of sonographically measured
bowel wall vascularity, histology, and disease activity in Crohn’s disease. Eur Radiol
19:1379–1386
33. Sciarretta G, Furno A, Mazzoni M, Basile C, Malaguti P (1993) Technetium-99 m hexamethyl
propylene amine oxime granulocyte scintigraphy in Crohn’s disease: diagnostic and
clinical relevance. Gut 34:1364–1369
34. Biancone L, Scopinaro F, Ierardi M et al (1997) 99mTc-HMPAO granulocyte scintigraphy
in the early detection of postoperative asymptomatic recurrence in Crohn’s disease. Dig
Dis Sci 42: 1549–1556
35. Panes J, Bouhnik Y, Reinisch W et al (2013) Imaging techniques for assessment of
inflammatory bowel disease: joint ECCO and ESGAR evidence-based consensus
guidelines. J Crohns Colitis 7: 556–585
Proefschrift2018-new2.indb 43 13/9/18 10:06
44
Chapter 2
36. Masselli G, Gualdi G (2013) CT and MR enterography in evaluat- ing small bowel diseases:
when to use which modality? Abdom Imaging 38:249–259
37. Grand DJ, Harris A, Loftus EV Jr (2012) Imaging for luminal disease and complications:
CT enterography, MR enterography, small-bowel follow-through, and ultrasound.
Gastroenterol Clin N Am 41:497–512
38. Siddiki HA, Fidler JL, Fletcher JG et al (2009) Prospective com- parison of state-of-the-
art MR enterography and CT enterography in small-bowel Crohn’s disease. AJR Am J
Roentgenol 193:113–121
39. Lee SS, Kim AY, Yang SK et al (2009) Crohn disease of the small bowel: comparison
of CT enterography, MR enterography, and small-bowel follow-through as diagnostic
techniques. Radiology 251:751–761
40. Jensen MD, Nathan T, Rafaelsen SR, Kjeldsen J (2011) Diagnostic accuracy of capsule
endoscopy for small bowel Crohn’s disease is superior to that of MR enterography or CT
enterography. Clin Gastroenterol Hepatol 9:124–129
41. Jensen MD, Kjeldsen J, Rafaelsen SR, Nathan T (2011) Diagnostic accuracies of MR
enterography and CT enterography in symptom- atic Crohn’s disease. Scand J
Gastroenterol 46:1449–1457
42. Rimola J, Rodriguez S, Garcia-Bosch O et al (2009) Magnetic resonance for assessment
of disease activity and severity in ileocolonic Crohn’s disease. Gut 58:1113–1120
43. Vermeire S, Ferrante M, Rutgeerts P (2013) Recent advances: personalised use of current
Crohn’s disease therapeutic options. Gut 62:1511–1515
44. Peloquin JM, Pardi DS, Sandborn WJ et al (2008) Diagnostic ion- izing radiation exposure
in a population-based cohort of patients with inflammatory bowel disease. Am J
Gastroenterol 103: 2015–2022
45. Steward MJ, Punwani S, Proctor I et al (2012) Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: deri- vation and histopathological validation of an MR-
based activity index. Eur J Radiol 81:2080–2088
Proefschrift2018-new2.indb 44 13/9/18 10:06
44
Chapter 2
36. Masselli G, Gualdi G (2013) CT and MR enterography in evaluat- ing small bowel diseases:
when to use which modality? Abdom Imaging 38:249–259
37. Grand DJ, Harris A, Loftus EV Jr (2012) Imaging for luminal disease and complications:
CT enterography, MR enterography, small-bowel follow-through, and ultrasound.
Gastroenterol Clin N Am 41:497–512
38. Siddiki HA, Fidler JL, Fletcher JG et al (2009) Prospective com- parison of state-of-the-
art MR enterography and CT enterography in small-bowel Crohn’s disease. AJR Am J
Roentgenol 193:113–121
39. Lee SS, Kim AY, Yang SK et al (2009) Crohn disease of the small bowel: comparison
of CT enterography, MR enterography, and small-bowel follow-through as diagnostic
techniques. Radiology 251:751–761
40. Jensen MD, Nathan T, Rafaelsen SR, Kjeldsen J (2011) Diagnostic accuracy of capsule
endoscopy for small bowel Crohn’s disease is superior to that of MR enterography or CT
enterography. Clin Gastroenterol Hepatol 9:124–129
41. Jensen MD, Kjeldsen J, Rafaelsen SR, Nathan T (2011) Diagnostic accuracies of MR
enterography and CT enterography in symptom- atic Crohn’s disease. Scand J
Gastroenterol 46:1449–1457
42. Rimola J, Rodriguez S, Garcia-Bosch O et al (2009) Magnetic resonance for assessment
of disease activity and severity in ileocolonic Crohn’s disease. Gut 58:1113–1120
43. Vermeire S, Ferrante M, Rutgeerts P (2013) Recent advances: personalised use of current
Crohn’s disease therapeutic options. Gut 62:1511–1515
44. Peloquin JM, Pardi DS, Sandborn WJ et al (2008) Diagnostic ion- izing radiation exposure
in a population-based cohort of patients with inflammatory bowel disease. Am J
Gastroenterol 103: 2015–2022
45. Steward MJ, Punwani S, Proctor I et al (2012) Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: deri- vation and histopathological validation of an MR-
based activity index. Eur J Radiol 81:2080–2088
Proefschrift2018-new2.indb 44 13/9/18 10:06
45
Grading CD using CT, MRI, US and scintigraphy
SUPPLEMENTARY MATERIALS
Appendix 1. Electronic search strategy for Medline, Embase and Cochrane databases (time
period: January 1983 –March 2014)
Medline search1 Crohn’s disease
2 Crohn [tiab]
3 Inflammatory bowel disease
4 1 OR 2 OR 3
5 Computed tomography
6 CT [tiab]
7 MRI
8 “Magnetic resonance” [All fields] OR (“magnetic” [All fields] AND “resonance” [All
9 Ultrasound
10 Scintigraphy
11 Emission computed tomography
12 5 OR 6 OR 7 OR 8 OR 9 OR 10 OR 11
13 4 AND 12
Embase search1 Crohn’s disease.ab,ti,sh,kw
2 Inflammatory bowel disease.ab,ti,sh,kw
3 1 OR 2
4 Computer Assisted Tomography.ab,ti,sh,kw
5 Exp Computer Assisted Tomography/
6 Nuclear magnetic resonance imaging.ab,ti,sh,kw
7 Exp Nuclear magnetic resonance imaging/
8 Echography.ab,ti,sh,kw
9 Exp echography/
10 Scintigraphy.ab,ti,sh,kw
11 Exp scintigraphy/
12 Positron emission tomography.ab,ti,sh,kw
13 Exp positron emission tomography/
14 4 OR 5 OR 6 OR 7 OR 8 OR 9 OR 10 OR 11 OR 12 OR 13
15 3 AND 14
Cochrane search1 Crohn disease [Mesh]
2 Inflammatory bowel disease [Mesh]
3 1 OR 2
4 Diagnostic techniques and procedures [Mesh]
5 3 AND 4
Proefschrift2018-new2.indb 45 13/9/18 10:06
45
Grading CD using CT, MRI, US and scintigraphy
SUPPLEMENTARY MATERIALS
Appendix 1. Electronic search strategy for Medline, Embase and Cochrane databases (time
period: January 1983 –March 2014)
Medline search1 Crohn’s disease
2 Crohn [tiab]
3 Inflammatory bowel disease
4 1 OR 2 OR 3
5 Computed tomography
6 CT [tiab]
7 MRI
8 “Magnetic resonance” [All fields] OR (“magnetic” [All fields] AND “resonance” [All
9 Ultrasound
10 Scintigraphy
11 Emission computed tomography
12 5 OR 6 OR 7 OR 8 OR 9 OR 10 OR 11
13 4 AND 12
Embase search1 Crohn’s disease.ab,ti,sh,kw
2 Inflammatory bowel disease.ab,ti,sh,kw
3 1 OR 2
4 Computer Assisted Tomography.ab,ti,sh,kw
5 Exp Computer Assisted Tomography/
6 Nuclear magnetic resonance imaging.ab,ti,sh,kw
7 Exp Nuclear magnetic resonance imaging/
8 Echography.ab,ti,sh,kw
9 Exp echography/
10 Scintigraphy.ab,ti,sh,kw
11 Exp scintigraphy/
12 Positron emission tomography.ab,ti,sh,kw
13 Exp positron emission tomography/
14 4 OR 5 OR 6 OR 7 OR 8 OR 9 OR 10 OR 11 OR 12 OR 13
15 3 AND 14
Cochrane search1 Crohn disease [Mesh]
2 Inflammatory bowel disease [Mesh]
3 1 OR 2
4 Diagnostic techniques and procedures [Mesh]
5 3 AND 4
Proefschrift2018-new2.indb 45 13/9/18 10:06
46
Chapter 2
Appendix 2. Exclusions after full-text evaluation (n=130)
No imaging features used for grading disease activity defined (n=58)1. Adamek HE, Schantzen W, Rinas U, Goyen M, Ajaj W, Esser C (2012) Ultra-high-field
magnetic resonance enterography in the diagnosis of ileitis (Neo-)terminalis: a prospective study. J Clin Gastroenterol 46:311-316
2. Bettenworth D, Reuter S, Hermann S et al (2013) Translational 18F-FDG PET/CT Imaging to Monitor Lesion Activity in Intestinal Inflammation. J Nucl Med:-
3. Biancone L, Calabrese E, Petruzziello C et al (2007) Wireless capsule endoscopy and small intestine contrast ultrasonography in recurrence of Crohn’s disease. Inflamm Bowel Dis 13:1256-1265
4. Borghi P, Armocida C, Rigo GP et al (1998) Advantages of the echographic staging of intestinal Crohn’s disease. Correlations of echographic patterns and histological findings. La Radiologia medica 95:338-343
5. Borthne AS, Abdelnoor M, Rugtveit J, Perminow G, Reiseter T, Klow NE (2006) Bowel magnetic resonance imaging of pediatric patients with oral mannitol MRI compared to endoscopy and intestinal ultrasound. Eur Radiol 16:207-214
6. Bremner AR, Griffiths M, Argent JD, Fairhurst JJ, Beattie RM (2006) Sonographic evaluation of inflammatory bowel disease: a prospective, blinded, comparative study. Pediatr Radiol 36:-
7. Bru C, Sans M, Defelitto MM et al (2001) Hydrocolonic sonography for evaluating inflammatory bowel disease. AJR Am J Roentgenol 177:99-105
8. C Calabrese E, Petruzziello C, Onali S et al (2009) Severity of postoperative recurrence in Crohn’s disease: correlation between endoscopic and sonographic findings. Inflamm Bowel Dis 15:1635-1642
9. Castiglione F, Bucci L, Pesce G et al (2008) Oral contrast-enhanced sonography for the diagnosis and grading of postsurgical recurrence of Crohn’s disease. Inflamm Bowel Dis 14:1240-1245
10. Castiglione F, Mainenti PP, De Palma GD et al (2013) Noninvasive diagnosis of small bowel Crohn’s disease: direct comparison of bowel sonography and magnetic resonance enterography. Inflamm Bowel Dis 19:991-998
11. Charron M, Del Rosario F, Kocoshis S (1999) Assessment of terminal ileal and colonic inflammation in Crohn’s disease with 99mTc-WBC. Acta Paediatr 88:193-198
12. Charron M, del Rosario FJ, Kocoshis SA (1999) Pediatric inflammatory bowel disease: assessment with scintigraphy with 99mTc white blood cells. Radiology 212:507-513
13. Charron M, Di Lorenzo C, Kocoshis S (1999) Gastric and small bowel Crohn’s disease assessed with leukocytes-Tc(99m) scintigraphy. Pediatr Surg Int 15:500-504
14. Charron M, Di Lorenzo C, Kocoshis S (2000) Are 99mTc leukocyte scintigraphy and SBFT studies useful in children suspected of having inflammatory bowel disease? Am J Gastroenterol 95:1208-1212
15. Choi D, Jin Lee S, Ah Cho Y et al (2003) Bowel wall thickening in patients with Crohn’s disease: CT patterns and correlation with inflammatory activity. Clin Radiol 58:68-74
16. Datz FL, Anderson CE, Ahluwalia R et al (1994) The efficacy of indium-111-polyclonal IgG for the detection of infection and inflammation. J Nucl Med 35:-
17. Daumal J, Martin-Comin J, Gasull MA et al (1989) [Gammagraphy with 111In-labelled leukocytes in an acute outbreak of inflammatory intestinal disease. Evaluation of the localization, extension and degree of activity]. Med Clin (Barc) 93:325-330
18. De Franco A, Di Veronica A, Armuzzi A et al (2012) Ileal Crohn disease: mural microvascularity quantified with contrast-enhanced US correlates with disease activity. Radiology 262:-
19. de Lima Ramos PA, Martin-Comin J, Prats E et al (1998) [Scintigraphic assessment of the severity of inflammatory bowel disease using Tc 99m exametazime-labeled leukocytes]. Rev Esp Med Nucl 17:351-357
20. Del Vescovo R, Sansoni I, Caviglia R et al (2008) Dynamic contrast enhanced magnetic resonance imaging of the terminal ileum: differentiation of activity of Crohn’s disease. Abdom Imaging 33:417-424
21. Fitzgerald PG, Topp TJ, Walton JM, Jackson JR, Gillis DA (1992) The use of indium 111 leukocyte scans in children with inflammatory bowel disease. J Pediatr Surg 27:1298-1300
22. Gee MS, Nimkin K, Hsu M et al (2011) Prospective evaluation of MR enterography as the primary imaging modality for pediatric Crohn disease assessment. AJR Am J Roentgenol 197:224-231
23. Girlich C, Jung EM, Huber E et al (2011) Comparison between preoperative quantitative assessment of bowel wall vascularization by contrast-enhanced ultrasound and operative macroscopic findings and results of histopathological scoring in Crohn’s disease. Ultraschall Med 32:-
24. Haber HP, Busch A, Ziebach R, Dette S, Ruck P, Stern M (2002) Ultrasonographic findings correspond to clinical, endoscopic, and histologic findings in inflammatory bowel disease and other enterocolitides. J Ultrasound Med 21:375-382
25. Hara AK, Alam S, Heigh RI, Gurudu SR, Hentz JG, Leighton JA (2008) Using CT enterography to monitor Crohn’s disease activity: a preliminary study. AJR Am J
Proefschrift2018-new2.indb 46 13/9/18 10:06
46
Chapter 2
Appendix 2. Exclusions after full-text evaluation (n=130)
No imaging features used for grading disease activity defined (n=58)1. Adamek HE, Schantzen W, Rinas U, Goyen M, Ajaj W, Esser C (2012) Ultra-high-field
magnetic resonance enterography in the diagnosis of ileitis (Neo-)terminalis: a prospective study. J Clin Gastroenterol 46:311-316
2. Bettenworth D, Reuter S, Hermann S et al (2013) Translational 18F-FDG PET/CT Imaging to Monitor Lesion Activity in Intestinal Inflammation. J Nucl Med:-
3. Biancone L, Calabrese E, Petruzziello C et al (2007) Wireless capsule endoscopy and small intestine contrast ultrasonography in recurrence of Crohn’s disease. Inflamm Bowel Dis 13:1256-1265
4. Borghi P, Armocida C, Rigo GP et al (1998) Advantages of the echographic staging of intestinal Crohn’s disease. Correlations of echographic patterns and histological findings. La Radiologia medica 95:338-343
5. Borthne AS, Abdelnoor M, Rugtveit J, Perminow G, Reiseter T, Klow NE (2006) Bowel magnetic resonance imaging of pediatric patients with oral mannitol MRI compared to endoscopy and intestinal ultrasound. Eur Radiol 16:207-214
6. Bremner AR, Griffiths M, Argent JD, Fairhurst JJ, Beattie RM (2006) Sonographic evaluation of inflammatory bowel disease: a prospective, blinded, comparative study. Pediatr Radiol 36:-
7. Bru C, Sans M, Defelitto MM et al (2001) Hydrocolonic sonography for evaluating inflammatory bowel disease. AJR Am J Roentgenol 177:99-105
8. C Calabrese E, Petruzziello C, Onali S et al (2009) Severity of postoperative recurrence in Crohn’s disease: correlation between endoscopic and sonographic findings. Inflamm Bowel Dis 15:1635-1642
9. Castiglione F, Bucci L, Pesce G et al (2008) Oral contrast-enhanced sonography for the diagnosis and grading of postsurgical recurrence of Crohn’s disease. Inflamm Bowel Dis 14:1240-1245
10. Castiglione F, Mainenti PP, De Palma GD et al (2013) Noninvasive diagnosis of small bowel Crohn’s disease: direct comparison of bowel sonography and magnetic resonance enterography. Inflamm Bowel Dis 19:991-998
11. Charron M, Del Rosario F, Kocoshis S (1999) Assessment of terminal ileal and colonic inflammation in Crohn’s disease with 99mTc-WBC. Acta Paediatr 88:193-198
12. Charron M, del Rosario FJ, Kocoshis SA (1999) Pediatric inflammatory bowel disease: assessment with scintigraphy with 99mTc white blood cells. Radiology 212:507-513
13. Charron M, Di Lorenzo C, Kocoshis S (1999) Gastric and small bowel Crohn’s disease assessed with leukocytes-Tc(99m) scintigraphy. Pediatr Surg Int 15:500-504
14. Charron M, Di Lorenzo C, Kocoshis S (2000) Are 99mTc leukocyte scintigraphy and SBFT studies useful in children suspected of having inflammatory bowel disease? Am J Gastroenterol 95:1208-1212
15. Choi D, Jin Lee S, Ah Cho Y et al (2003) Bowel wall thickening in patients with Crohn’s disease: CT patterns and correlation with inflammatory activity. Clin Radiol 58:68-74
16. Datz FL, Anderson CE, Ahluwalia R et al (1994) The efficacy of indium-111-polyclonal IgG for the detection of infection and inflammation. J Nucl Med 35:-
17. Daumal J, Martin-Comin J, Gasull MA et al (1989) [Gammagraphy with 111In-labelled leukocytes in an acute outbreak of inflammatory intestinal disease. Evaluation of the localization, extension and degree of activity]. Med Clin (Barc) 93:325-330
18. De Franco A, Di Veronica A, Armuzzi A et al (2012) Ileal Crohn disease: mural microvascularity quantified with contrast-enhanced US correlates with disease activity. Radiology 262:-
19. de Lima Ramos PA, Martin-Comin J, Prats E et al (1998) [Scintigraphic assessment of the severity of inflammatory bowel disease using Tc 99m exametazime-labeled leukocytes]. Rev Esp Med Nucl 17:351-357
20. Del Vescovo R, Sansoni I, Caviglia R et al (2008) Dynamic contrast enhanced magnetic resonance imaging of the terminal ileum: differentiation of activity of Crohn’s disease. Abdom Imaging 33:417-424
21. Fitzgerald PG, Topp TJ, Walton JM, Jackson JR, Gillis DA (1992) The use of indium 111 leukocyte scans in children with inflammatory bowel disease. J Pediatr Surg 27:1298-1300
22. Gee MS, Nimkin K, Hsu M et al (2011) Prospective evaluation of MR enterography as the primary imaging modality for pediatric Crohn disease assessment. AJR Am J Roentgenol 197:224-231
23. Girlich C, Jung EM, Huber E et al (2011) Comparison between preoperative quantitative assessment of bowel wall vascularization by contrast-enhanced ultrasound and operative macroscopic findings and results of histopathological scoring in Crohn’s disease. Ultraschall Med 32:-
24. Haber HP, Busch A, Ziebach R, Dette S, Ruck P, Stern M (2002) Ultrasonographic findings correspond to clinical, endoscopic, and histologic findings in inflammatory bowel disease and other enterocolitides. J Ultrasound Med 21:375-382
25. Hara AK, Alam S, Heigh RI, Gurudu SR, Hentz JG, Leighton JA (2008) Using CT enterography to monitor Crohn’s disease activity: a preliminary study. AJR Am J
Proefschrift2018-new2.indb 46 13/9/18 10:06
47
Grading CD using CT, MRI, US and scintigraphy
Roentgenol 190:1512-151626. Hara AK, Leighton JA, Heigh RI et al (2006) Crohn disease of the small bowel: preliminary
comparison among CT enterography, capsule endoscopy, small-bowel follow-through, and ileoscopy. Radiology 238:-
27. Hassan C, Cerro P, Zullo A, Spina C, Morini S (2003) Computed tomography enteroclysis in comparison with ileoscopy in patients with Crohn’s disease. Int J Colorectal Dis 18:121-125
28. Heresbach D, Bretagne JF, Raoul JL et al (1993) Indium scanning in assessment of acute Crohn’s disease. A prospective study of sensitivity and correlation with severity of mucosal damage. Dig Dis Sci 38:1601-1607
29. Holtmann MH, Uenzen M, Helisch A et al (2012) 18F-Fluorodeoxyglucose positron-emission tomography (PET) can be used to assess inflammation non-invasively in Crohn’s disease. Dig Dis Sci 57:-
30. Hotze A, Briele B, Wolf F, Biersack HJ, Knapp FF (1988) Localization and activity of inflammatory bowel disease using 111In leukocyte imaging. Nuklearmedizin 27:83-86
31. Jacene HA, Ginsburg P, Kwon J et al (2009) Prediction of the need for surgical intervention in obstructive Crohn’s disease by 18F-FDG PET/CT. J Nucl Med 50:1751-1759
32. Jewell FM, Davies A, Sandhu B, Duncan A, Grier D (1996) Technetium-99m-HMPAO labelled leucocytes in the detection and monitoring of inflammatory bowel disease in children. Br J Radiol 69:508-514
33. Koutroubakis IE, Koukouraki SI, Dimoulios PD, Velidaki AA, Karkavitsas NS, Kouroumalis EA (2003) Active inflammatory bowel disease: evaluation with 99mTc (V) DMSA scintigraphy. Radiology 229:70-74
34. Loffler M, Weckesser M, Franzius C, Schober O, Zimmer KP (2006) High diagnostic value of 18F-FDG-PET in pediatric patients with chronic inflammatory bowel disease. Ann N Y Acad Sci 1072:379-385
35. Louis E, Ancion G, Colard A, Spote V, Belaiche J, Hustinx R (2007) Noninvasive assessment of Crohn’s disease intestinal lesions with (18)F-FDG PET/CT. J Nucl Med 48:1053-1059
36. Low RN, Francis IR, Politoske D, Bennett M (2000) Crohn’s disease evaluation: comparison of contrast-enhanced MR imaging and single-phase helical CT scanning. J Magn Reson Imaging 11:127-135
37. Madsen SM, Thomsen HS, Munkholm P et al (2002) Inflammatory bowel disease evaluated by low-field magnetic resonance imaging. Comparison with endoscopy, 99mTc-HMPAO leucocyte scintigraphy, conventional radiography and surgery. Scand J Gastroenterol 37:307-316
38. Malgras B, Soyer P, Boudiaf M et al (2012) Accuracy of imaging for predicting operative approach in Crohn’s disease. Br J Surg 99:1011-1020
39. Onali S, Calabrese E, Petruzziello C et al (2010) Endoscopic vs ultrasonographic findings related to Crohn’s disease recurrence: a prospective longitudinal study at 3 years. J Crohns Colitis 4:319-328
40. Pallotta N, Giovannone M, Pezzotti P et al (2010) Ultrasonographic detection and assessment of the severity of Crohn’s disease recurrence after ileal resection. BMC Gastroenterol 10:69
41. Papos M, Nagy F, Lang J, Csernay L (1993) Technetium-99m hexamethylpropylene amine oxime labelled leucocyte scintigraphy in ulcerative colitis and Crohn’s disease. Eur J Nucl Med 20:766-769
42. Papos M, Nagy F, Narai G et al (1996) Anti-granulocyte immunoscintigraphy and [99mTc]hexamethylpropyleneamine-oxime-labeled leukocyte scintigraphy in inflammatory bowel disease. Dig Dis Sci 41:412-420
43. Piekkala M, Kalajoki-Helmio T, Martelius L, Pakarinen M, Rintala R, Kolho KL (2012) Magnetic resonance enterography guiding treatment in children with Crohn’s jejunoileitis. Acta Paediatr 101:631-636
44. Pilleul F, Godefroy C, Yzebe-Beziat D, Dugougeat-Pilleul F, Lachaux A, Valette PJ (2005) Magnetic resonance imaging in Crohn’s disease. Gastroenterol Clin Biol 29:803-808
45. Rispo A, Bucci L, Pesce G et al (2006) Bowel sonography for the diagnosis and grading of postsurgical recurrence of Crohn’s disease. Inflamm Bowel Dis 12:486-490
46. Rottgen R, Herzog H, Lopez-Haninnen E, Felix R (2006) Bowel wall enhancement in magnetic resonance colonography for assessing activity in Crohn’s disease. Clin Imaging 30:27-31
47. Sauer CG, Middleton JP, Alazraki A et al (2012) Comparison of magnetic resonance enterography with endoscopy, histopathology, and laboratory evaluation in pediatric Crohn disease. J Pediatr Gastroenterol Nutr 55:178-184
48. Solem CA, Loftus EV, Jr., Fletcher JG et al (2008) Small-bowel imaging in Crohn’s disease: a prospective, blinded, 4-way comparison trial. Gastrointest Endosc 68:255-266
49. Thomson M, Rao P, Berger L, Rawat D (2012) Graded compression and power Doppler ultrasonography versus endoscopy to assess paediatric Crohn disease activity pre- and posttreatment. J Pediatr Gastroenterol Nutr 54:404-408
50. Tretjak Z, Fettich J, Batagelj I (1990) Technetium-99m-sucralfate scintigraphy: poor correlation with endoscopy and radiology in inflammatory bowel disease. Gastrointest Endosc 36:-
51. Vilien M, Nielsen SL, Jorgensen M et al (1992) Leucocyte scintigraphy to localize
Proefschrift2018-new2.indb 47 13/9/18 10:06
47
Grading CD using CT, MRI, US and scintigraphy
Roentgenol 190:1512-151626. Hara AK, Leighton JA, Heigh RI et al (2006) Crohn disease of the small bowel: preliminary
comparison among CT enterography, capsule endoscopy, small-bowel follow-through, and ileoscopy. Radiology 238:-
27. Hassan C, Cerro P, Zullo A, Spina C, Morini S (2003) Computed tomography enteroclysis in comparison with ileoscopy in patients with Crohn’s disease. Int J Colorectal Dis 18:121-125
28. Heresbach D, Bretagne JF, Raoul JL et al (1993) Indium scanning in assessment of acute Crohn’s disease. A prospective study of sensitivity and correlation with severity of mucosal damage. Dig Dis Sci 38:1601-1607
29. Holtmann MH, Uenzen M, Helisch A et al (2012) 18F-Fluorodeoxyglucose positron-emission tomography (PET) can be used to assess inflammation non-invasively in Crohn’s disease. Dig Dis Sci 57:-
30. Hotze A, Briele B, Wolf F, Biersack HJ, Knapp FF (1988) Localization and activity of inflammatory bowel disease using 111In leukocyte imaging. Nuklearmedizin 27:83-86
31. Jacene HA, Ginsburg P, Kwon J et al (2009) Prediction of the need for surgical intervention in obstructive Crohn’s disease by 18F-FDG PET/CT. J Nucl Med 50:1751-1759
32. Jewell FM, Davies A, Sandhu B, Duncan A, Grier D (1996) Technetium-99m-HMPAO labelled leucocytes in the detection and monitoring of inflammatory bowel disease in children. Br J Radiol 69:508-514
33. Koutroubakis IE, Koukouraki SI, Dimoulios PD, Velidaki AA, Karkavitsas NS, Kouroumalis EA (2003) Active inflammatory bowel disease: evaluation with 99mTc (V) DMSA scintigraphy. Radiology 229:70-74
34. Loffler M, Weckesser M, Franzius C, Schober O, Zimmer KP (2006) High diagnostic value of 18F-FDG-PET in pediatric patients with chronic inflammatory bowel disease. Ann N Y Acad Sci 1072:379-385
35. Louis E, Ancion G, Colard A, Spote V, Belaiche J, Hustinx R (2007) Noninvasive assessment of Crohn’s disease intestinal lesions with (18)F-FDG PET/CT. J Nucl Med 48:1053-1059
36. Low RN, Francis IR, Politoske D, Bennett M (2000) Crohn’s disease evaluation: comparison of contrast-enhanced MR imaging and single-phase helical CT scanning. J Magn Reson Imaging 11:127-135
37. Madsen SM, Thomsen HS, Munkholm P et al (2002) Inflammatory bowel disease evaluated by low-field magnetic resonance imaging. Comparison with endoscopy, 99mTc-HMPAO leucocyte scintigraphy, conventional radiography and surgery. Scand J Gastroenterol 37:307-316
38. Malgras B, Soyer P, Boudiaf M et al (2012) Accuracy of imaging for predicting operative approach in Crohn’s disease. Br J Surg 99:1011-1020
39. Onali S, Calabrese E, Petruzziello C et al (2010) Endoscopic vs ultrasonographic findings related to Crohn’s disease recurrence: a prospective longitudinal study at 3 years. J Crohns Colitis 4:319-328
40. Pallotta N, Giovannone M, Pezzotti P et al (2010) Ultrasonographic detection and assessment of the severity of Crohn’s disease recurrence after ileal resection. BMC Gastroenterol 10:69
41. Papos M, Nagy F, Lang J, Csernay L (1993) Technetium-99m hexamethylpropylene amine oxime labelled leucocyte scintigraphy in ulcerative colitis and Crohn’s disease. Eur J Nucl Med 20:766-769
42. Papos M, Nagy F, Narai G et al (1996) Anti-granulocyte immunoscintigraphy and [99mTc]hexamethylpropyleneamine-oxime-labeled leukocyte scintigraphy in inflammatory bowel disease. Dig Dis Sci 41:412-420
43. Piekkala M, Kalajoki-Helmio T, Martelius L, Pakarinen M, Rintala R, Kolho KL (2012) Magnetic resonance enterography guiding treatment in children with Crohn’s jejunoileitis. Acta Paediatr 101:631-636
44. Pilleul F, Godefroy C, Yzebe-Beziat D, Dugougeat-Pilleul F, Lachaux A, Valette PJ (2005) Magnetic resonance imaging in Crohn’s disease. Gastroenterol Clin Biol 29:803-808
45. Rispo A, Bucci L, Pesce G et al (2006) Bowel sonography for the diagnosis and grading of postsurgical recurrence of Crohn’s disease. Inflamm Bowel Dis 12:486-490
46. Rottgen R, Herzog H, Lopez-Haninnen E, Felix R (2006) Bowel wall enhancement in magnetic resonance colonography for assessing activity in Crohn’s disease. Clin Imaging 30:27-31
47. Sauer CG, Middleton JP, Alazraki A et al (2012) Comparison of magnetic resonance enterography with endoscopy, histopathology, and laboratory evaluation in pediatric Crohn disease. J Pediatr Gastroenterol Nutr 55:178-184
48. Solem CA, Loftus EV, Jr., Fletcher JG et al (2008) Small-bowel imaging in Crohn’s disease: a prospective, blinded, 4-way comparison trial. Gastrointest Endosc 68:255-266
49. Thomson M, Rao P, Berger L, Rawat D (2012) Graded compression and power Doppler ultrasonography versus endoscopy to assess paediatric Crohn disease activity pre- and posttreatment. J Pediatr Gastroenterol Nutr 54:404-408
50. Tretjak Z, Fettich J, Batagelj I (1990) Technetium-99m-sucralfate scintigraphy: poor correlation with endoscopy and radiology in inflammatory bowel disease. Gastrointest Endosc 36:-
51. Vilien M, Nielsen SL, Jorgensen M et al (1992) Leucocyte scintigraphy to localize
Proefschrift2018-new2.indb 47 13/9/18 10:06
48
Chapter 2
inflammatory activity in ulcerative colitis and Crohn’s disease. Scand J Gastroenterol 27:582-586
52. Vogel J, da Luz Moreira A, Baker M et al (2007) CT enterography for Crohn’s disease: accurate preoperative diagnostic imaging. Dis Colon Rectum 50:-
53. Weldon MJ (1994) Assessment of inflammatory bowel disease activity using 99mTc-HMPAO single-photon emission computerized tomography imaging. Scand J Gastroenterol Suppl 203:61-68
54. Weldon MJ, Masoomi AM, Britten AJ et al (1995) Quantification of inflammatory bowel disease activity using technetium-99m HMPAO labelled leucocyte single photon emission computerised tomography (SPECT). Gut 36:243-250
55. Wu YW, Tang YH, Hao NX, Tang CY, Miao F (2012) Crohn’s disease: CT enterography manifestations before and after treatment. Eur J Radiol 81:52-59
56. Zappa M, Stefanescu C, Cazals-Hatem D et al (2011) Which magnetic resonance imaging findings accurately evaluate inflammation in small bowel Crohn’s disease? A retrospective comparison with surgical pathologic analysis. Inflamm Bowel Dis 17:984-993
57. Ziech ML, Bipat S, Roelofs JJ et al (2011) Retrospective comparison of magnetic resonance imaging features and histopathology in Crohn’s disease patients. Eur J Radiol 80:e299-305
58. Ziech ML, Lavini C, Caan MW et al (2012) Dynamic contrast-enhanced MRI in patients with luminal Crohn’s disease. Eur J Radiol 81:3019-3027
No raw data available for 3 x 3 or 4 x 4 tables (n=36)1. Adler J, Punglia DR, Dillman JR et al (2012) Computed tomography enterography findings
correlate with tissue inflammation, not fibrosis in resected small bowel Crohn’s disease. Inflamm Bowel Dis 18:849-856
2. Afifi AH, Kassem MI (2012) Crohn’s disease: Activity, complications and treatment. Evaluation using MDCT enterography and endoscopy. Egyptian Journal of Radiology and Nuclear Medicine 43:507-517
3. Ajaj WM, Lauenstein TC, Pelster G et al (2005) Magnetic resonance colonography for the detection of inflammatory diseases of the large bowel: quantifying the inflammatory activity. Gut 54:257-263
4. Alberini JL, Badran A, Freneaux E et al (2001) Technetium-99m HMPAO-labeled leukocyte imaging compared with endoscopy, ultrasonography, and contrast radiology in children with inflammatory bowel disease. J Pediatr Gastroenterol Nutr 32:278-286
5. Bhargava SA, Orenstein SR, Charron M. Technetium-99m hexamethylpropyleneamine-oxime-labeled leukocyte scintigraphy in inflammatory bowel disease in children. The Journal of pediatrics. 1994;125(2):213-217.
6. Biancone L, Schillaci O, Capoccetti F et al (2005) Technetium-99m-HMPAO labeled leukocyte single photon emission computerized tomography (SPECT) for assessing Crohn’s disease extent and intestinal infiltration. Am J Gastroenterol 100:344-354
7. Cucchiara S, Celentano L, de Magistris TM, Montisci A, Iula VD, Fecarotta S (1999) Colonoscopy and technetium-99m white cell scan in children with suspected inflammatory bowel disease. J Pediatr 135:727-732
8. Friedrich C, Fajfar A, Pawlik M et al (2012) Magnetic resonance enterography with and without biphasic contrast agent enema compared to conventional ileocolonoscopy in patients with Crohn’s disease. Inflamm Bowel Dis 18:1842-1848
9. Gourtsoyiannis N, Papanikolaou N, Grammatikakis J, Papamastorakis G, Prassopoulos P, Roussomoustakaki M (2004) Assessment of Crohn’s disease activity in the small bowel with MR and conventional enteroclysis: preliminary results. Eur Radiol 14:1017-1024
10. Grand DJ, Kampalath V, Harris A et al (2012) MR enterography correlates highly with colonoscopy and histology for both distal ileal and colonic Crohn’s disease in 310 patients. Eur J Radiol 81:e763-769
11. Hyun SB, Kitazume Y, Nagahori M et al (2011) Magnetic resonance enterocolonography is useful for simultaneous evaluation of small and large intestinal lesions in Crohn’s disease. Inflamm Bowel Dis 17:1063-1072
12. Johnson KT, Hara AK, Johnson CD (2009) Evaluation of colitis: usefulness of CT enterography technique. Emerg Radiol 16:277-282
13. Langhorst J, Kuhle CA, Ajaj W et al (2007) MR colonography without bowel purgation for the assessment of inflammatory bowel diseases: diagnostic accuracy and patient acceptance. Inflamm Bowel Dis 13:1001-1008
14. Lapp RT, Spier BJ, Perlman SB, Jaskowiak CJ, Reichelderfer M (2011) Clinical utility of positron emission tomography/computed tomography in inflammatory bowel disease. Mol Imaging Biol 13:573-576
15. Lasocki A, Pitman A, Williams R, Lui B, Kalade AV, Farish S (2011) Relative efficacy of different MRI signs in diagnosing active Crohn’s disease, compared against a histological gold standard. J Med Imaging Radiat Oncol 55:11-19
16. Lawrance IC, Welman CJ, Shipman P, Murray K (2009) Correlation of MRI-determined small bowel Crohn’s disease categories with medical response and surgical pathology. World J Gastroenterol 15:3367-3375
Proefschrift2018-new2.indb 48 13/9/18 10:06
48
Chapter 2
inflammatory activity in ulcerative colitis and Crohn’s disease. Scand J Gastroenterol 27:582-586
52. Vogel J, da Luz Moreira A, Baker M et al (2007) CT enterography for Crohn’s disease: accurate preoperative diagnostic imaging. Dis Colon Rectum 50:-
53. Weldon MJ (1994) Assessment of inflammatory bowel disease activity using 99mTc-HMPAO single-photon emission computerized tomography imaging. Scand J Gastroenterol Suppl 203:61-68
54. Weldon MJ, Masoomi AM, Britten AJ et al (1995) Quantification of inflammatory bowel disease activity using technetium-99m HMPAO labelled leucocyte single photon emission computerised tomography (SPECT). Gut 36:243-250
55. Wu YW, Tang YH, Hao NX, Tang CY, Miao F (2012) Crohn’s disease: CT enterography manifestations before and after treatment. Eur J Radiol 81:52-59
56. Zappa M, Stefanescu C, Cazals-Hatem D et al (2011) Which magnetic resonance imaging findings accurately evaluate inflammation in small bowel Crohn’s disease? A retrospective comparison with surgical pathologic analysis. Inflamm Bowel Dis 17:984-993
57. Ziech ML, Bipat S, Roelofs JJ et al (2011) Retrospective comparison of magnetic resonance imaging features and histopathology in Crohn’s disease patients. Eur J Radiol 80:e299-305
58. Ziech ML, Lavini C, Caan MW et al (2012) Dynamic contrast-enhanced MRI in patients with luminal Crohn’s disease. Eur J Radiol 81:3019-3027
No raw data available for 3 x 3 or 4 x 4 tables (n=36)1. Adler J, Punglia DR, Dillman JR et al (2012) Computed tomography enterography findings
correlate with tissue inflammation, not fibrosis in resected small bowel Crohn’s disease. Inflamm Bowel Dis 18:849-856
2. Afifi AH, Kassem MI (2012) Crohn’s disease: Activity, complications and treatment. Evaluation using MDCT enterography and endoscopy. Egyptian Journal of Radiology and Nuclear Medicine 43:507-517
3. Ajaj WM, Lauenstein TC, Pelster G et al (2005) Magnetic resonance colonography for the detection of inflammatory diseases of the large bowel: quantifying the inflammatory activity. Gut 54:257-263
4. Alberini JL, Badran A, Freneaux E et al (2001) Technetium-99m HMPAO-labeled leukocyte imaging compared with endoscopy, ultrasonography, and contrast radiology in children with inflammatory bowel disease. J Pediatr Gastroenterol Nutr 32:278-286
5. Bhargava SA, Orenstein SR, Charron M. Technetium-99m hexamethylpropyleneamine-oxime-labeled leukocyte scintigraphy in inflammatory bowel disease in children. The Journal of pediatrics. 1994;125(2):213-217.
6. Biancone L, Schillaci O, Capoccetti F et al (2005) Technetium-99m-HMPAO labeled leukocyte single photon emission computerized tomography (SPECT) for assessing Crohn’s disease extent and intestinal infiltration. Am J Gastroenterol 100:344-354
7. Cucchiara S, Celentano L, de Magistris TM, Montisci A, Iula VD, Fecarotta S (1999) Colonoscopy and technetium-99m white cell scan in children with suspected inflammatory bowel disease. J Pediatr 135:727-732
8. Friedrich C, Fajfar A, Pawlik M et al (2012) Magnetic resonance enterography with and without biphasic contrast agent enema compared to conventional ileocolonoscopy in patients with Crohn’s disease. Inflamm Bowel Dis 18:1842-1848
9. Gourtsoyiannis N, Papanikolaou N, Grammatikakis J, Papamastorakis G, Prassopoulos P, Roussomoustakaki M (2004) Assessment of Crohn’s disease activity in the small bowel with MR and conventional enteroclysis: preliminary results. Eur Radiol 14:1017-1024
10. Grand DJ, Kampalath V, Harris A et al (2012) MR enterography correlates highly with colonoscopy and histology for both distal ileal and colonic Crohn’s disease in 310 patients. Eur J Radiol 81:e763-769
11. Hyun SB, Kitazume Y, Nagahori M et al (2011) Magnetic resonance enterocolonography is useful for simultaneous evaluation of small and large intestinal lesions in Crohn’s disease. Inflamm Bowel Dis 17:1063-1072
12. Johnson KT, Hara AK, Johnson CD (2009) Evaluation of colitis: usefulness of CT enterography technique. Emerg Radiol 16:277-282
13. Langhorst J, Kuhle CA, Ajaj W et al (2007) MR colonography without bowel purgation for the assessment of inflammatory bowel diseases: diagnostic accuracy and patient acceptance. Inflamm Bowel Dis 13:1001-1008
14. Lapp RT, Spier BJ, Perlman SB, Jaskowiak CJ, Reichelderfer M (2011) Clinical utility of positron emission tomography/computed tomography in inflammatory bowel disease. Mol Imaging Biol 13:573-576
15. Lasocki A, Pitman A, Williams R, Lui B, Kalade AV, Farish S (2011) Relative efficacy of different MRI signs in diagnosing active Crohn’s disease, compared against a histological gold standard. J Med Imaging Radiat Oncol 55:11-19
16. Lawrance IC, Welman CJ, Shipman P, Murray K (2009) Correlation of MRI-determined small bowel Crohn’s disease categories with medical response and surgical pathology. World J Gastroenterol 15:3367-3375
Proefschrift2018-new2.indb 48 13/9/18 10:06
49
Grading CD using CT, MRI, US and scintigraphy
17. Maccioni F, Viola F, Carrozzo F et al (2012) Differences in the location and activity of intestinal Crohn’s disease lesions between adult and paediatric patients detected with MRI. Eur Radiol 22:2465-2477
18. Marcos HB, Semelka RC (2000) Evaluation of Crohn’s disease using half-fourier RARE and gadolinium-enhanced SGE sequences: initial results. Magn Reson Imaging 18:263-268
19. Menys A, Atkinson D, Odille F et al (2012) Quantified terminal ileal motility during MR enterography as a potential biomarker of Crohn’s disease activity: a preliminary study. Eur Radiol 22:-
20. Minordi LM, Vecchioli A, Poloni G, Guidi L, De Vitis I, Bonomo L (2009) Enteroclysis CT and PEG-CT in patients with previous small-bowel surgical resection for Crohn’s disease: CT findings and correlation with endoscopy. Eur Radiol 19:-
21. Oussalah A, Laurent V, Bruot O et al (2010) Diffusion-weighted magnetic resonance without bowel preparation for detecting colonic inflammation in inflammatory bowel disease. Gut 59:1056-1065
22. Paredes JM, Ripolles T, Cortes X et al (2013) Contrast-enhanced ultrasonography: usefulness in the assessment of postoperative recurrence of Crohn’s disease. J Crohns Colitis 7:192-201
23. Paredes JM, Ripolles T, Cortes X et al (2010) Non-invasive diagnosis and grading of postsurgical endoscopic recurrence in Crohn’s disease: usefulness of abdominal ultrasonography and (99m)Tc-hexamethylpropylene amineoxime-labelled leucocyte scintigraphy. J Crohns Colitis 4:537-545
24. Parisinos CA, McIntyre VE, Heron T et al (2010) Magnetic resonance follow-through imaging for evaluation of disease activity in ileal Crohn’s disease: an observational, retrospective cohort study. Inflamm Bowel Dis 16:1219-1226
25. Pascu M, Roznowski AB, Muller HP, Adler A, Wiedenmann B, Dignass AU (2004) Clinical relevance of transabdominal ultrasonography and magnetic resonance imaging in patients with inflammatory bowel disease of the terminal ileum and large bowel. Inflamm Bowel Dis 10:373-382
26. Pullman WE, Sullivan PJ, Barratt PJ, Lising J, Booth JA, Doe WF (1988) Assessment of inflammatory bowel disease activity by technetium 99m phagocyte scanning. Gastroenterology 95:989-996
27. Rigazio C, Ercole E, Laudi C et al (2009) Abdominal bowel ultrasound can predict the risk of surgery in Crohn’s disease: proposal of an ultrasonographic score. Scand J Gastroenterol 44:585-593
28. Rimola J, Ordas I, Rodriguez S et al (2011) Magnetic resonance imaging for evaluation of Crohn’s disease: validation of parameters of severity and quantitative index of activity. Inflamm Bowel Dis 17:1759-1768
29. Rimola J, Rodriguez S, Garcia-Bosch O et al (2009) Magnetic resonance for assessment of disease activity and severity in ileocolonic Crohn’s disease. Gut 58:1113-1120
30. Ripolles T, Martinez MJ, Paredes JM, Blanc E, Flors L, Delgado F (2009) Crohn disease: correlation of findings at contrast-enhanced US with severity at endoscopy. Radiology 253:241-248
31. Ripolles T, Rausell N, Paredes JM, Grau E, Martinez MJ, Vizuete J (2013) Effectiveness of contrast-enhanced ultrasound for characterisation of intestinal inflammation in Crohn’s disease: a comparison with surgical histopathology analysis. J Crohns Colitis 7:120-128
32. Sailer J, Peloschek P, Reinisch W, Vogelsang H, Turetschek K, Schima W (2008) Anastomotic recurrence of Crohn’s disease after ileocolic resection: comparison of MR enteroclysis with endoscopy. Eur Radiol 18:2512-2521
33. Shyn PB, Mortele KJ, Britz-Cunningham SH et al (2010) Low-dose 18F-FDG PET/CT enterography: improving on CT enterography assessment of patients with Crohn disease. J Nucl Med 51:1841-1848
34. Silverstein J, Grand D, Kawatu D, Shah SA, Steinkeler J, LeLeiko N (2012) Feasibility of using MR enterography for the assessment of terminal ileitis and inflammatory activity in children with Crohn disease. J Pediatr Gastroenterol Nutr 55:173-177
35. Stathaki MI, Koutroubakis IE, Koukouraki SI et al (2008) Active inflammatory bowel disease: head-to-head comparison between 99mTc-hexamethylpropylene amine oxime white blood cells and 99mTc(V)-dimercaptosuccinic acid scintigraphy. Nucl Med Commun 29:27-32
36. Steward MJ, Punwani S, Proctor I et al (2012) Non-perforating small bowel Crohn’s disease assessed by MRI enterography: derivation and histopathological validation of an MR-based activity index. Eur J Radiol 81:2080-2088
No endoscopy, biopsies or intra-operative findings (n=25)1. Arndt JW, Grootscholten MI, van Hogezand RA, Griffioen G, Lamers CB, Pauwels EK
(1997) Inflammatory bowel disease activity assessment using technetium-99m-HMPAO leukocytes. Dig Dis Sci 42:387-393
2. Becker W, Fischbach W, Jenett M, Reiners C, Borner W (1986) 111In-oxine-labelled white blood cells in the diagnosis and follow-up of Crohn’s disease. Klin Wochenschr 64:141-148
3. Brouwers AH, De Jong DJ, Dams ET et al (2000) Tc-99m-PEG-Liposomes for the evaluation of colitis in Crohn’s disease. J Drug Target 8:225-233
Proefschrift2018-new2.indb 49 13/9/18 10:06
49
Grading CD using CT, MRI, US and scintigraphy
17. Maccioni F, Viola F, Carrozzo F et al (2012) Differences in the location and activity of intestinal Crohn’s disease lesions between adult and paediatric patients detected with MRI. Eur Radiol 22:2465-2477
18. Marcos HB, Semelka RC (2000) Evaluation of Crohn’s disease using half-fourier RARE and gadolinium-enhanced SGE sequences: initial results. Magn Reson Imaging 18:263-268
19. Menys A, Atkinson D, Odille F et al (2012) Quantified terminal ileal motility during MR enterography as a potential biomarker of Crohn’s disease activity: a preliminary study. Eur Radiol 22:-
20. Minordi LM, Vecchioli A, Poloni G, Guidi L, De Vitis I, Bonomo L (2009) Enteroclysis CT and PEG-CT in patients with previous small-bowel surgical resection for Crohn’s disease: CT findings and correlation with endoscopy. Eur Radiol 19:-
21. Oussalah A, Laurent V, Bruot O et al (2010) Diffusion-weighted magnetic resonance without bowel preparation for detecting colonic inflammation in inflammatory bowel disease. Gut 59:1056-1065
22. Paredes JM, Ripolles T, Cortes X et al (2013) Contrast-enhanced ultrasonography: usefulness in the assessment of postoperative recurrence of Crohn’s disease. J Crohns Colitis 7:192-201
23. Paredes JM, Ripolles T, Cortes X et al (2010) Non-invasive diagnosis and grading of postsurgical endoscopic recurrence in Crohn’s disease: usefulness of abdominal ultrasonography and (99m)Tc-hexamethylpropylene amineoxime-labelled leucocyte scintigraphy. J Crohns Colitis 4:537-545
24. Parisinos CA, McIntyre VE, Heron T et al (2010) Magnetic resonance follow-through imaging for evaluation of disease activity in ileal Crohn’s disease: an observational, retrospective cohort study. Inflamm Bowel Dis 16:1219-1226
25. Pascu M, Roznowski AB, Muller HP, Adler A, Wiedenmann B, Dignass AU (2004) Clinical relevance of transabdominal ultrasonography and magnetic resonance imaging in patients with inflammatory bowel disease of the terminal ileum and large bowel. Inflamm Bowel Dis 10:373-382
26. Pullman WE, Sullivan PJ, Barratt PJ, Lising J, Booth JA, Doe WF (1988) Assessment of inflammatory bowel disease activity by technetium 99m phagocyte scanning. Gastroenterology 95:989-996
27. Rigazio C, Ercole E, Laudi C et al (2009) Abdominal bowel ultrasound can predict the risk of surgery in Crohn’s disease: proposal of an ultrasonographic score. Scand J Gastroenterol 44:585-593
28. Rimola J, Ordas I, Rodriguez S et al (2011) Magnetic resonance imaging for evaluation of Crohn’s disease: validation of parameters of severity and quantitative index of activity. Inflamm Bowel Dis 17:1759-1768
29. Rimola J, Rodriguez S, Garcia-Bosch O et al (2009) Magnetic resonance for assessment of disease activity and severity in ileocolonic Crohn’s disease. Gut 58:1113-1120
30. Ripolles T, Martinez MJ, Paredes JM, Blanc E, Flors L, Delgado F (2009) Crohn disease: correlation of findings at contrast-enhanced US with severity at endoscopy. Radiology 253:241-248
31. Ripolles T, Rausell N, Paredes JM, Grau E, Martinez MJ, Vizuete J (2013) Effectiveness of contrast-enhanced ultrasound for characterisation of intestinal inflammation in Crohn’s disease: a comparison with surgical histopathology analysis. J Crohns Colitis 7:120-128
32. Sailer J, Peloschek P, Reinisch W, Vogelsang H, Turetschek K, Schima W (2008) Anastomotic recurrence of Crohn’s disease after ileocolic resection: comparison of MR enteroclysis with endoscopy. Eur Radiol 18:2512-2521
33. Shyn PB, Mortele KJ, Britz-Cunningham SH et al (2010) Low-dose 18F-FDG PET/CT enterography: improving on CT enterography assessment of patients with Crohn disease. J Nucl Med 51:1841-1848
34. Silverstein J, Grand D, Kawatu D, Shah SA, Steinkeler J, LeLeiko N (2012) Feasibility of using MR enterography for the assessment of terminal ileitis and inflammatory activity in children with Crohn disease. J Pediatr Gastroenterol Nutr 55:173-177
35. Stathaki MI, Koutroubakis IE, Koukouraki SI et al (2008) Active inflammatory bowel disease: head-to-head comparison between 99mTc-hexamethylpropylene amine oxime white blood cells and 99mTc(V)-dimercaptosuccinic acid scintigraphy. Nucl Med Commun 29:27-32
36. Steward MJ, Punwani S, Proctor I et al (2012) Non-perforating small bowel Crohn’s disease assessed by MRI enterography: derivation and histopathological validation of an MR-based activity index. Eur J Radiol 81:2080-2088
No endoscopy, biopsies or intra-operative findings (n=25)1. Arndt JW, Grootscholten MI, van Hogezand RA, Griffioen G, Lamers CB, Pauwels EK
(1997) Inflammatory bowel disease activity assessment using technetium-99m-HMPAO leukocytes. Dig Dis Sci 42:387-393
2. Becker W, Fischbach W, Jenett M, Reiners C, Borner W (1986) 111In-oxine-labelled white blood cells in the diagnosis and follow-up of Crohn’s disease. Klin Wochenschr 64:141-148
3. Brouwers AH, De Jong DJ, Dams ET et al (2000) Tc-99m-PEG-Liposomes for the evaluation of colitis in Crohn’s disease. J Drug Target 8:225-233
Proefschrift2018-new2.indb 49 13/9/18 10:06
50
Chapter 2
4. Darnault P, Bretagne JF, Moisan A et al (1985) [Scintigraphy with indium 111-labeled polynuclears: correlation with the extensiveness and activity of cryptogenetic enterocolitis]. Gastroenterol Clin Biol 9:690-696
5. Dhote R, Beades E, Le Dinh T et al (1995) [Scintigraphy using leukocytes marked with 99m-Tc-HMPAO in Crohn’s disease]. Acta Gastroenterol Belg 58:353-363
6. Florie J, Wasser MN, Arts-Cieslik K, Akkerman EM, Siersema PD, Stoker J (2006) Dynamic contrast-enhanced MRI of the bowel wall for assessment of disease activity in Crohn’s disease. AJR Am J Roentgenol 186:1384-1392
7. Futagami Y, Haruma K, Hata J et al (1999) Development and validation of an ultrasonographic activity index of Crohn’s disease. Eur J Gastroenterol Hepatol 11:1007-1012
8. Grieser C, Denecke T, Steffen IG et al (2012) Magnetic resonance enteroclysis in patients with Crohn’s disease: fat saturated T2-weighted sequences for evaluation of inflammatory activity. J Crohns Colitis 6:294-301
9. Hata J, Haruma K, Suenaga K et al (1992) Ultrasonographic assessment of inflammatory bowel disease. Am J Gastroenterol 87:443-447
10. Hata J, Haruma K, Yamanaka H et al (1994) Ultrasonographic evaluation of the bowel wall in inflammatory bowel disease: comparison of in vivo and in vitro studies. Abdom Imaging 19:395-399
11. Limberg B (1986) Diagnosis of inflammatory and tumorous changes of the large intestine using colonic sonography. Deutsche medizinische Wochenschrift 111:1273-1276
12. Martinez MJ, Ripolles T, Paredes JM, Blanc E, Marti-Bonmati L (2009) Assessment of the extension and the inflammatory activity in Crohn’s disease: comparison of ultrasound and MRI. Abdom Imaging 34:141-148
13. Migaleddu V, Scanu AM, Quaia E et al (2009) Contrast-enhanced ultrasonographic evaluation of inflammatory activity in Crohn’s disease. Gastroenterology 137:43-52
14. Molnar T, Papos M, Gyulai C et al (2001) Clinical value of technetium-99m-HMPAO-labeled leukocyte scintigraphy and spiral computed tomography in active Crohn’s disease. Am J Gastroenterol 96:1517-1521
15. Poitras P, Carrier L, Chartrand R et al (1987) Indium-111 leukocyte scanning of the abdomen. Analysis of its value for diagnosis and management of inflammatory bowel disease. J Clin Gastroenterol 9:-
16. Pradel JA, David XR, Taourel P, Djafari M, Veyrac M, Bruel JM (1997) Sonographic assessment of the normal and abnormal bowel wall in nondiverticular ileitis and colitis. Abdom Imaging 22:167-172
17. Pullman W, Hanna R, Sullivan P, Booth JA, Lomas F, Doe WF (1986) Technetium-99m autologous phagocyte scanning: a new imaging technique for inflammatory bowel disease. Br Med J (Clin Res Ed) 293:171-174
18. Saverymuttu SH, Camilleri M, Rees H, Lavender JP, Hodgson HJ, Chadwick VS (1986) Indium 111-granulocyte scanning in the assessment of disease extent and disease activity in inflammatory bowel disease. A comparison with colonoscopy, histology, and fecal indium 111-granulocyte excretion. Gastroenterology 90:1121-1128
19. Scholmerich J, Schmidt E, Schumichen C, Billmann P, Schmidt H, Gerok W (1988) Scintigraphic assessment of bowel involvement and disease activity in Crohn’s disease using technetium 99m-hexamethyl propylene amine oxine as leukocyte label. Gastroenterology 95:1287-1293
20. Sopena F, Nerin JM, Prats E et al (1991) [Gammagraphy with labeled leukocytes as an activity and extension index of Crohn disease]. Rev Esp Enferm Dig 79:387-392
21. Spinelli F, Milella M, Sara R et al (1991) The 99mTc-HMPAO leukocyte scan: an alternative to radiology and endoscopy in evaluating the extent and the activity of inflammatory bowel disease. J Nucl Biol Med 35:82-87
22. Tarjan Z, Toth G, Gyorke T, Mester A, Karlinger K, Mako EK (2000) Ultrasound in Crohn’s disease of the small bowel. Eur J Radiol 35:176-182
23. Tillack C, Seiderer J, Brand S et al (2008) Correlation of magnetic resonance enteroclysis (MRE) and wireless capsule endoscopy (CE) in the diagnosis of small bowel lesions in Crohn’s disease. Inflamm Bowel Dis 14:1219-1228
24. Tong JL, Feng Q, Shen J et al (2013) Computed tomography enterography versus balloon-assisted enteroscopy for evaluation of small bowel lesions in Crohn’s disease. J Gastroenterol Hepatol. 10.1111/jgh.12231
25. Verdu Rico J, Juste Ruiz M, Jover R et al (2006) [99mTc-HMPAO-leukocyte-labeled scintigraphy in the detection and follow-up of inflammatory bowel disease]. An Pediatr (Barc) 64:457-463
Less than 10 included patients (n=6)1. Alvarez Beltran M, Barber Martinez de la Torre I, Segarra Canton O, Redecillas Ferreiro
S, Castellote Alonso A, Infante Pina D (2013) [MRI enterography in the assessment of paediatric Crohn’s disease]. An Pediatr (Barc) 78:314-320
2. Charron M, Di Lorenzo C, Kocoshis SA et al (2001) (99m)Tc antigranulocyte monoclonal antibody imaging for the detection and assessment of inflammatory bowel disease newly
Proefschrift2018-new2.indb 50 13/9/18 10:06
50
Chapter 2
4. Darnault P, Bretagne JF, Moisan A et al (1985) [Scintigraphy with indium 111-labeled polynuclears: correlation with the extensiveness and activity of cryptogenetic enterocolitis]. Gastroenterol Clin Biol 9:690-696
5. Dhote R, Beades E, Le Dinh T et al (1995) [Scintigraphy using leukocytes marked with 99m-Tc-HMPAO in Crohn’s disease]. Acta Gastroenterol Belg 58:353-363
6. Florie J, Wasser MN, Arts-Cieslik K, Akkerman EM, Siersema PD, Stoker J (2006) Dynamic contrast-enhanced MRI of the bowel wall for assessment of disease activity in Crohn’s disease. AJR Am J Roentgenol 186:1384-1392
7. Futagami Y, Haruma K, Hata J et al (1999) Development and validation of an ultrasonographic activity index of Crohn’s disease. Eur J Gastroenterol Hepatol 11:1007-1012
8. Grieser C, Denecke T, Steffen IG et al (2012) Magnetic resonance enteroclysis in patients with Crohn’s disease: fat saturated T2-weighted sequences for evaluation of inflammatory activity. J Crohns Colitis 6:294-301
9. Hata J, Haruma K, Suenaga K et al (1992) Ultrasonographic assessment of inflammatory bowel disease. Am J Gastroenterol 87:443-447
10. Hata J, Haruma K, Yamanaka H et al (1994) Ultrasonographic evaluation of the bowel wall in inflammatory bowel disease: comparison of in vivo and in vitro studies. Abdom Imaging 19:395-399
11. Limberg B (1986) Diagnosis of inflammatory and tumorous changes of the large intestine using colonic sonography. Deutsche medizinische Wochenschrift 111:1273-1276
12. Martinez MJ, Ripolles T, Paredes JM, Blanc E, Marti-Bonmati L (2009) Assessment of the extension and the inflammatory activity in Crohn’s disease: comparison of ultrasound and MRI. Abdom Imaging 34:141-148
13. Migaleddu V, Scanu AM, Quaia E et al (2009) Contrast-enhanced ultrasonographic evaluation of inflammatory activity in Crohn’s disease. Gastroenterology 137:43-52
14. Molnar T, Papos M, Gyulai C et al (2001) Clinical value of technetium-99m-HMPAO-labeled leukocyte scintigraphy and spiral computed tomography in active Crohn’s disease. Am J Gastroenterol 96:1517-1521
15. Poitras P, Carrier L, Chartrand R et al (1987) Indium-111 leukocyte scanning of the abdomen. Analysis of its value for diagnosis and management of inflammatory bowel disease. J Clin Gastroenterol 9:-
16. Pradel JA, David XR, Taourel P, Djafari M, Veyrac M, Bruel JM (1997) Sonographic assessment of the normal and abnormal bowel wall in nondiverticular ileitis and colitis. Abdom Imaging 22:167-172
17. Pullman W, Hanna R, Sullivan P, Booth JA, Lomas F, Doe WF (1986) Technetium-99m autologous phagocyte scanning: a new imaging technique for inflammatory bowel disease. Br Med J (Clin Res Ed) 293:171-174
18. Saverymuttu SH, Camilleri M, Rees H, Lavender JP, Hodgson HJ, Chadwick VS (1986) Indium 111-granulocyte scanning in the assessment of disease extent and disease activity in inflammatory bowel disease. A comparison with colonoscopy, histology, and fecal indium 111-granulocyte excretion. Gastroenterology 90:1121-1128
19. Scholmerich J, Schmidt E, Schumichen C, Billmann P, Schmidt H, Gerok W (1988) Scintigraphic assessment of bowel involvement and disease activity in Crohn’s disease using technetium 99m-hexamethyl propylene amine oxine as leukocyte label. Gastroenterology 95:1287-1293
20. Sopena F, Nerin JM, Prats E et al (1991) [Gammagraphy with labeled leukocytes as an activity and extension index of Crohn disease]. Rev Esp Enferm Dig 79:387-392
21. Spinelli F, Milella M, Sara R et al (1991) The 99mTc-HMPAO leukocyte scan: an alternative to radiology and endoscopy in evaluating the extent and the activity of inflammatory bowel disease. J Nucl Biol Med 35:82-87
22. Tarjan Z, Toth G, Gyorke T, Mester A, Karlinger K, Mako EK (2000) Ultrasound in Crohn’s disease of the small bowel. Eur J Radiol 35:176-182
23. Tillack C, Seiderer J, Brand S et al (2008) Correlation of magnetic resonance enteroclysis (MRE) and wireless capsule endoscopy (CE) in the diagnosis of small bowel lesions in Crohn’s disease. Inflamm Bowel Dis 14:1219-1228
24. Tong JL, Feng Q, Shen J et al (2013) Computed tomography enterography versus balloon-assisted enteroscopy for evaluation of small bowel lesions in Crohn’s disease. J Gastroenterol Hepatol. 10.1111/jgh.12231
25. Verdu Rico J, Juste Ruiz M, Jover R et al (2006) [99mTc-HMPAO-leukocyte-labeled scintigraphy in the detection and follow-up of inflammatory bowel disease]. An Pediatr (Barc) 64:457-463
Less than 10 included patients (n=6)1. Alvarez Beltran M, Barber Martinez de la Torre I, Segarra Canton O, Redecillas Ferreiro
S, Castellote Alonso A, Infante Pina D (2013) [MRI enterography in the assessment of paediatric Crohn’s disease]. An Pediatr (Barc) 78:314-320
2. Charron M, Di Lorenzo C, Kocoshis SA et al (2001) (99m)Tc antigranulocyte monoclonal antibody imaging for the detection and assessment of inflammatory bowel disease newly
Proefschrift2018-new2.indb 50 13/9/18 10:06
51
Grading CD using CT, MRI, US and scintigraphy
diagnosed by colonoscopy in children. Pediatr Radiol 31:796-8003. Durno CA, Sherman P, Williams T, Shuckett B, Dupuis A, Griffiths AM (2000) Magnetic
resonance imaging to distinguish the type and severity of pediatric inflammatory bowel diseases. J Pediatr Gastroenterol Nutr 30:-
4. Lantto E, Jarvi K, Krekela I et al (1992) Technetium-99m hexamethyl propylene amine oxine leucocytes in the assessment of disease activity in inflammatory bowel disease. Eur J Nucl Med 19:14-18
5. Papos M, Varkonyi A, Lang J et al (1996) HM-PAO-labeled leukocyte scintigraphy in pediatric patients with inflammatory bowel disease. J Pediatr Gastroenterol Nutr 23:547-552
6. Stahlberg D, Veress B, Mare K et al (1997) Leukocyte migration in acute colonic inflammatory bowel disease: comparison of histological assessment and Tc-99m-HMPAO labeled leukocyte scan. Am J Gastroenterol 92:283-288
Separate analysis CD patients not possible (n=2)1. Caobelli F, Panarotto MB, Andreoli F, Ravelli A, De Agostini A, Giubbini R (2011) Is 99mTc-
HMPAO granulocyte scan an alternative to endoscopy in pediatric chronic inflammatory bowel disease (IBD)? Eur J Pediatr 170:51-57
2. Laghi A, Borrelli O, Paolantonio P et al (2003) Contrast enhanced magnetic resonance imaging of the terminal ileum in children with Crohn’s disease. Gut 52:393-397
Duplicate publication (n=1)1. Gallego Ojea JC, Echarri Piudo AI, Porta Vila A (2011) [Crohn’s disease: the usefulness of
MR enterography in the detection of recurrence after surgery]. Radiologia 53:552-559
Incorrect type of work (n=1)1. Gandolfi L (1996) Comparison of magnetic resonance imaging and endoscopy in
distinguishing the type and severity of inflammatory bowel disease. Gastrointest Endosc 43:-
Article not in English, German, French, Italian, Spanish or Dutch (n=1)1. Li WJ, Jiang WY, Zhang XF, Liu XS, Shi RH, Zhang HJ (2013) Value of CT enteroclysis in the
diagnosis of Crohn’s disease. World Chinese Journal of Digestology 21:220-225
Appendix 3. Three-by-three contingency tables
Table 1. Computed tomography (CT)
Reference test None Mild Frank
CT None Mild Frank None Mild Frank None Mild Frank
Study
Per-patient
Mao [16] 3 1 0 1 3 5 0 0 19
Mohamed [17] 0 0 0 0 0 0 0 1 25
Per-segment
Kolkman [18] 51 0 0 5 2 0 1 3 8
Proefschrift2018-new2.indb 51 13/9/18 10:06
51
Grading CD using CT, MRI, US and scintigraphy
diagnosed by colonoscopy in children. Pediatr Radiol 31:796-8003. Durno CA, Sherman P, Williams T, Shuckett B, Dupuis A, Griffiths AM (2000) Magnetic
resonance imaging to distinguish the type and severity of pediatric inflammatory bowel diseases. J Pediatr Gastroenterol Nutr 30:-
4. Lantto E, Jarvi K, Krekela I et al (1992) Technetium-99m hexamethyl propylene amine oxine leucocytes in the assessment of disease activity in inflammatory bowel disease. Eur J Nucl Med 19:14-18
5. Papos M, Varkonyi A, Lang J et al (1996) HM-PAO-labeled leukocyte scintigraphy in pediatric patients with inflammatory bowel disease. J Pediatr Gastroenterol Nutr 23:547-552
6. Stahlberg D, Veress B, Mare K et al (1997) Leukocyte migration in acute colonic inflammatory bowel disease: comparison of histological assessment and Tc-99m-HMPAO labeled leukocyte scan. Am J Gastroenterol 92:283-288
Separate analysis CD patients not possible (n=2)1. Caobelli F, Panarotto MB, Andreoli F, Ravelli A, De Agostini A, Giubbini R (2011) Is 99mTc-
HMPAO granulocyte scan an alternative to endoscopy in pediatric chronic inflammatory bowel disease (IBD)? Eur J Pediatr 170:51-57
2. Laghi A, Borrelli O, Paolantonio P et al (2003) Contrast enhanced magnetic resonance imaging of the terminal ileum in children with Crohn’s disease. Gut 52:393-397
Duplicate publication (n=1)1. Gallego Ojea JC, Echarri Piudo AI, Porta Vila A (2011) [Crohn’s disease: the usefulness of
MR enterography in the detection of recurrence after surgery]. Radiologia 53:552-559
Incorrect type of work (n=1)1. Gandolfi L (1996) Comparison of magnetic resonance imaging and endoscopy in
distinguishing the type and severity of inflammatory bowel disease. Gastrointest Endosc 43:-
Article not in English, German, French, Italian, Spanish or Dutch (n=1)1. Li WJ, Jiang WY, Zhang XF, Liu XS, Shi RH, Zhang HJ (2013) Value of CT enteroclysis in the
diagnosis of Crohn’s disease. World Chinese Journal of Digestology 21:220-225
Appendix 3. Three-by-three contingency tables
Table 1. Computed tomography (CT)
Reference test None Mild Frank
CT None Mild Frank None Mild Frank None Mild Frank
Study
Per-patient
Mao [16] 3 1 0 1 3 5 0 0 19
Mohamed [17] 0 0 0 0 0 0 0 1 25
Per-segment
Kolkman [18] 51 0 0 5 2 0 1 3 8
Proefschrift2018-new2.indb 51 13/9/18 10:06
52
Chapter 2
Table 2. Magnetic resonance imaging (MRI)
Reference standard None Mild Frank
MRI None Mild Frank None Mild Frank None Mild Frank
Study
Per-patient
Schill [19] 0 0 0 0 4 0 0 1 71
Gallego [20] 4 1 0 0 24 9 0 4 19
Koilakou [21] 2 1 0 0 5 1 0 0 17
Horsthuis [22] Ob1 0 0 0 3 1 1 2 2 6
Horsthuis [22] Ob2 0 0 0 4 0 1 1 5 4
Horsthuis [22] Ob3 0 0 0 0 3 2 0 3 7
Girometti [23] 14 2 0 2 12 0 0 0 15
Horsthuis [24] Ob1 0 5 0 0 8 1 0 3 3
Horsthuis [24] Ob2 4 1 0 1 7 1 0 2 4
Florie [25] Ob1 6 3 1 0 4 4 0 4 9
Florie [25] Ob2 5 4 1 2 3 3 1 2 10
Shoenut [26] 0 0 0 0 1 0 0 0 11
Shoenut [27] 0 0 0 0 0 4 0 1 14
Per-segment
Schreyer [28] 112 1 1 11 3 1 14 3 30
Schreyer [29] 46 0 0 25 2 0 1 1 7
Ob1, observer 1; Ob2, observer 2
Table 3. Ultrasound (US)
Reference standard None Mild Frank
US None Mild Frank None Mild Frank None Mild Frank
Study
Per-patient
Drews [30] 5 1 2 5 5 5 0 5 4
Per-segment
Neye [31] 54 4 0 12 7 6 3 6 34
Bozkurt [32] 51 26 3 9 29 35 0 11 28
Table 4. Scintigraphy
Reference standard None Mild Frank
Scintigraphy None Mild Frank None Mild Frank None Mild Frank
Study
Per-patient
Biancone [33] 1 0 0 2 3 1 1 2 0
Per-segment
Kolkman [18] 50 1 0 4 2 1 1 3 8
Scarrietta [34] 10 0 0 4 24 4 0 4 42
Proefschrift2018-new2.indb 52 13/9/18 10:06
52
Chapter 2
Table 2. Magnetic resonance imaging (MRI)
Reference standard None Mild Frank
MRI None Mild Frank None Mild Frank None Mild Frank
Study
Per-patient
Schill [19] 0 0 0 0 4 0 0 1 71
Gallego [20] 4 1 0 0 24 9 0 4 19
Koilakou [21] 2 1 0 0 5 1 0 0 17
Horsthuis [22] Ob1 0 0 0 3 1 1 2 2 6
Horsthuis [22] Ob2 0 0 0 4 0 1 1 5 4
Horsthuis [22] Ob3 0 0 0 0 3 2 0 3 7
Girometti [23] 14 2 0 2 12 0 0 0 15
Horsthuis [24] Ob1 0 5 0 0 8 1 0 3 3
Horsthuis [24] Ob2 4 1 0 1 7 1 0 2 4
Florie [25] Ob1 6 3 1 0 4 4 0 4 9
Florie [25] Ob2 5 4 1 2 3 3 1 2 10
Shoenut [26] 0 0 0 0 1 0 0 0 11
Shoenut [27] 0 0 0 0 0 4 0 1 14
Per-segment
Schreyer [28] 112 1 1 11 3 1 14 3 30
Schreyer [29] 46 0 0 25 2 0 1 1 7
Ob1, observer 1; Ob2, observer 2
Table 3. Ultrasound (US)
Reference standard None Mild Frank
US None Mild Frank None Mild Frank None Mild Frank
Study
Per-patient
Drews [30] 5 1 2 5 5 5 0 5 4
Per-segment
Neye [31] 54 4 0 12 7 6 3 6 34
Bozkurt [32] 51 26 3 9 29 35 0 11 28
Table 4. Scintigraphy
Reference standard None Mild Frank
Scintigraphy None Mild Frank None Mild Frank None Mild Frank
Study
Per-patient
Biancone [33] 1 0 0 2 3 1 1 2 0
Per-segment
Kolkman [18] 50 1 0 4 2 1 1 3 8
Scarrietta [34] 10 0 0 4 24 4 0 4 42
Proefschrift2018-new2.indb 52 13/9/18 10:06
Proefschrift2018-new2.indb 53 13/9/18 10:06 Proefschrift2018-new2.indb 53 13/9/18 10:06
Proefschrift2018-new2.indb 54 13/9/18 10:06 Proefschrift2018-new2.indb 54 13/9/18 10:06
CHAPTER 3
Comparison of MRI activity scoring systems and features for the terminal ileum in Crohn’s disease patients
Carl. A.J. Puylaert*, Charlotte J. Tutein Nolthenius*, Jeroen A.W. Tielbeek, Jesica C.
Makanyanga, C. Yung Nio, Doug A. Pendse, Cyriel Y. Ponsioen, Frans M. Vos, Stuart
A. Taylor, Jaap Stoker
*Both authors contributed equally
Proefschrift2018-new2.indb 55 13/9/18 10:06
CHAPTER 3
Comparison of MRI activity scoring systems and features for the terminal ileum in Crohn’s disease patients
Carl. A.J. Puylaert*, Charlotte J. Tutein Nolthenius*, Jeroen A.W. Tielbeek, Jesica C.
Makanyanga, C. Yung Nio, Doug A. Pendse, Cyriel Y. Ponsioen, Frans M. Vos, Stuart
A. Taylor, Jaap Stoker
*Both authors contributed equally
Proefschrift2018-new2.indb 55 13/9/18 10:06
56
Chapter 3
ABSTRACT
Background
To evaluate four previously validated magnetic resonance imaging (MRI) activity
scoring systems for diagnosis and grading of Crohn’s disease (CD) in the terminal
ileum (TI) against an endoscopic and histopathological reference standard.
Methods
Ethics approval and written informed consent were obtained. Subjects with
known or suspected CD were prospectively recruited between December 2011 and
August 2014. Each patient underwent MRI and ileocolonoscopy with TI biopsies.
Four MRI scoring systems (CDMI, London, MaRIA and Clermont) and component
features were applied by two observers and correlated to the Crohn’s Disease
Endoscopic Index of Severity (CDEIS, 0–44) and histopathological eAIS score (0–
6). Interobserver agreement (weighted kappa and intraclass coefficient (ICC)) and
diagnostic accuracy for active and ulcerating endoscopic/histopathologic disease
were evaluated.
Results
98 patients (median age 32 years; female 56%) were included. All four scoring
systems showed good interobserver agreement (ICC: 0.70–0.78), moderate-to-
strong correlation to CDEIS (r=0.57–0.67) and weak-to-moderate correlation to
eAIS (r=0.38–0.49). Scoring systems diagnostic accuracy for active and ulcerating
endoscopic disease ranged from 73–78% and 71–76%, respectively, while for active
histopathologic disease this ranged from 65–72%. Between the scoring systems,
no significant differences were found for both observers regarding interobserver
agreement, correlation coefficients and diagnostic accuracy.
Conclusions
All scoring systems were comparable in terms of interobserver agreement,
correlation to the endoscopic and histopathological reference standard and
diagnostic accuracy. The London score, MaRIA and Clermont score have the
additional benefit of having validated cut-off values for both active and ulcerating
endoscopic disease.
Proefschrift2018-new2.indb 56 13/9/18 10:06
56
Chapter 3
ABSTRACT
Background
To evaluate four previously validated magnetic resonance imaging (MRI) activity
scoring systems for diagnosis and grading of Crohn’s disease (CD) in the terminal
ileum (TI) against an endoscopic and histopathological reference standard.
Methods
Ethics approval and written informed consent were obtained. Subjects with
known or suspected CD were prospectively recruited between December 2011 and
August 2014. Each patient underwent MRI and ileocolonoscopy with TI biopsies.
Four MRI scoring systems (CDMI, London, MaRIA and Clermont) and component
features were applied by two observers and correlated to the Crohn’s Disease
Endoscopic Index of Severity (CDEIS, 0–44) and histopathological eAIS score (0–
6). Interobserver agreement (weighted kappa and intraclass coefficient (ICC)) and
diagnostic accuracy for active and ulcerating endoscopic/histopathologic disease
were evaluated.
Results
98 patients (median age 32 years; female 56%) were included. All four scoring
systems showed good interobserver agreement (ICC: 0.70–0.78), moderate-to-
strong correlation to CDEIS (r=0.57–0.67) and weak-to-moderate correlation to
eAIS (r=0.38–0.49). Scoring systems diagnostic accuracy for active and ulcerating
endoscopic disease ranged from 73–78% and 71–76%, respectively, while for active
histopathologic disease this ranged from 65–72%. Between the scoring systems,
no significant differences were found for both observers regarding interobserver
agreement, correlation coefficients and diagnostic accuracy.
Conclusions
All scoring systems were comparable in terms of interobserver agreement,
correlation to the endoscopic and histopathological reference standard and
diagnostic accuracy. The London score, MaRIA and Clermont score have the
additional benefit of having validated cut-off values for both active and ulcerating
endoscopic disease.
Proefschrift2018-new2.indb 56 13/9/18 10:06
57
Comparison of MRI scoring systems
INTRODUCTION
Magnetic resonance imaging (MRI) is widely used for evaluation of Crohn’s disease
(1,2). Grading of disease activity is important to determine optimal treatment
strategies and thereafter monitor response to treatment. Extensive research has
yielded a range of MRI features that are useful in assessing disease activity (3–6).
From these features, several activity scoring systems – the Magnetic Resonance
Index of Activity (MaRIA), Clermont score, London score and Crohn’s disease MRI
Index (CDMI) – have been developed and validated (3,5,7–10). Ideally, a scoring
system should be objective, quick and easy to use, reproducible, quantitative and
non-invasive. The available scoring systems incorporate different features and were
developed and validated using different reference standards (endoscopic and
histopathologic scores), which potentially complicate their interchangeability. There
is therefore the clinical need to determine the strengths of each scoring system and
its incorporated features.
The terminal ileum is the most commonly affected part of the gastrointestinal tract
in Crohn’s disease. The ileocecal valve (or ileocolonic anastomosis in post-operative
patients) is an easily identifiable landmark at both at ileocolonoscopy and MRI,
facilitating accurate anatomical co-localisation between the two modalities. The
robustness of correlative studies between MRI and histopathological activity scores
is therefore improved by focusing on the ileocecal valve and immediately adjacent
area. Whilst ileocolonoscopy provides a much fuller overview of disease extent
and activity compared to histopathology, correlative studies against the Crohn’s
Disease Endoscopic Index of Severity (CDEIS) also benefit from robust anatomical
co-localisation, given the difficulties in precise matching of colonic segments
between endoscopy and MRI.
The aim of this prospective study was to compare several scoring systems for
grading of Crohn’s disease activity in the terminal ileum and evaluate individual MRI
parameters using both an endoscopic and histopathological reference standard.
Proefschrift2018-new2.indb 57 13/9/18 10:06
57
Comparison of MRI scoring systems
INTRODUCTION
Magnetic resonance imaging (MRI) is widely used for evaluation of Crohn’s disease
(1,2). Grading of disease activity is important to determine optimal treatment
strategies and thereafter monitor response to treatment. Extensive research has
yielded a range of MRI features that are useful in assessing disease activity (3–6).
From these features, several activity scoring systems – the Magnetic Resonance
Index of Activity (MaRIA), Clermont score, London score and Crohn’s disease MRI
Index (CDMI) – have been developed and validated (3,5,7–10). Ideally, a scoring
system should be objective, quick and easy to use, reproducible, quantitative and
non-invasive. The available scoring systems incorporate different features and were
developed and validated using different reference standards (endoscopic and
histopathologic scores), which potentially complicate their interchangeability. There
is therefore the clinical need to determine the strengths of each scoring system and
its incorporated features.
The terminal ileum is the most commonly affected part of the gastrointestinal tract
in Crohn’s disease. The ileocecal valve (or ileocolonic anastomosis in post-operative
patients) is an easily identifiable landmark at both at ileocolonoscopy and MRI,
facilitating accurate anatomical co-localisation between the two modalities. The
robustness of correlative studies between MRI and histopathological activity scores
is therefore improved by focusing on the ileocecal valve and immediately adjacent
area. Whilst ileocolonoscopy provides a much fuller overview of disease extent
and activity compared to histopathology, correlative studies against the Crohn’s
Disease Endoscopic Index of Severity (CDEIS) also benefit from robust anatomical
co-localisation, given the difficulties in precise matching of colonic segments
between endoscopy and MRI.
The aim of this prospective study was to compare several scoring systems for
grading of Crohn’s disease activity in the terminal ileum and evaluate individual MRI
parameters using both an endoscopic and histopathological reference standard.
Proefschrift2018-new2.indb 57 13/9/18 10:06
58
Chapter 3
MATERIALS AND METHODS
Patients
Between December 2011 until August 2014, patients with suspected or proven Crohn’s
disease (based on endoscopy or histopathology) were consecutively recruited
from two tertiary referral centers (1. Academic Medical Center, Amsterdam, the
Netherlands and 2. University College London Hospitals, London, United Kingdom)
as part of the VIGOR++ study (FP7/2007-2013, 270379). Patients were ≥18 years of
age, and underwent MRI and ileocolonoscopy within two weeks of each other. For
the current study, we included all patients in whom a complete ileocolonoscopy with
terminal ileum biopsies was performed. The following exclusion criteria were used:
general contraindications to MRI (claustrophobia, pregnancy, renal insufficiency,
pacemaker), an incomplete MRI scan protocol (see below), a final diagnosis other
than Crohn’s disease or if no biopsies from the terminal ileum could be taken (e.g.
in case of incomplete colonoscopy due to impassable strictures). Ethical permission
was obtained from the hospitals’ medical ethics committee and all patients
gave written informed consent. Part of our study population was published in a
previous study on semiautomatic MRI assessment of Crohn’s disease (11). Details on
overlapping patient inclusions and exclusions are detailed in the results.
Reference standard
The MaRIA and Clermont score were developed using the endoscopic CDEIS as a
reference standard, whilst the CDMI and London score were developed using the
histopathologic acute inflammation score (eAIS) (3,5,8). Therefore, in the current
study we applied both reference standards in order to strengthen the comparison
between the scoring systems and minimize potential bias induced by using one
particular reference standard over another. Ileocolonoscopy was performed
according to standard quality indicators by either a gastroenterologist or a senior
resident in gastroenterology under direct supervision of a gastroenterologist using
a standard endoscope (model CF-160L, Olympus). Sedation was administered
intravenously at the discretion of the endoscopist. The Crohn’s Disease Endoscopic
Index of Severity (CDEIS) was determined in the last 20 cm of the (neo-)terminal
ileum by the gastroenterologist/supervisor (12). For matching to MRI, two to four
biopsies were taken from a fixed location. In case of disease, the most inflamed area
within the distal 5 cm of the terminal ileum was biopsied, while in the remaining
Proefschrift2018-new2.indb 58 13/9/18 10:06
58
Chapter 3
MATERIALS AND METHODS
Patients
Between December 2011 until August 2014, patients with suspected or proven Crohn’s
disease (based on endoscopy or histopathology) were consecutively recruited
from two tertiary referral centers (1. Academic Medical Center, Amsterdam, the
Netherlands and 2. University College London Hospitals, London, United Kingdom)
as part of the VIGOR++ study (FP7/2007-2013, 270379). Patients were ≥18 years of
age, and underwent MRI and ileocolonoscopy within two weeks of each other. For
the current study, we included all patients in whom a complete ileocolonoscopy with
terminal ileum biopsies was performed. The following exclusion criteria were used:
general contraindications to MRI (claustrophobia, pregnancy, renal insufficiency,
pacemaker), an incomplete MRI scan protocol (see below), a final diagnosis other
than Crohn’s disease or if no biopsies from the terminal ileum could be taken (e.g.
in case of incomplete colonoscopy due to impassable strictures). Ethical permission
was obtained from the hospitals’ medical ethics committee and all patients
gave written informed consent. Part of our study population was published in a
previous study on semiautomatic MRI assessment of Crohn’s disease (11). Details on
overlapping patient inclusions and exclusions are detailed in the results.
Reference standard
The MaRIA and Clermont score were developed using the endoscopic CDEIS as a
reference standard, whilst the CDMI and London score were developed using the
histopathologic acute inflammation score (eAIS) (3,5,8). Therefore, in the current
study we applied both reference standards in order to strengthen the comparison
between the scoring systems and minimize potential bias induced by using one
particular reference standard over another. Ileocolonoscopy was performed
according to standard quality indicators by either a gastroenterologist or a senior
resident in gastroenterology under direct supervision of a gastroenterologist using
a standard endoscope (model CF-160L, Olympus). Sedation was administered
intravenously at the discretion of the endoscopist. The Crohn’s Disease Endoscopic
Index of Severity (CDEIS) was determined in the last 20 cm of the (neo-)terminal
ileum by the gastroenterologist/supervisor (12). For matching to MRI, two to four
biopsies were taken from a fixed location. In case of disease, the most inflamed area
within the distal 5 cm of the terminal ileum was biopsied, while in the remaining
Proefschrift2018-new2.indb 58 13/9/18 10:06
59
Comparison of MRI scoring systems
cases, biopsies were taken approximately 2 cm proximal to the ileocecal valve or
anastomosis (in the case of previous ileocecal resection). Biopsies were examined
in consensus by two experienced pathologists (LB and MRJ), who were unaware of
clinical or radiologic findings. Each biopsy was examined on at least three slices and
graded according to the eAIS (3). This grading system describes the presence of six
histological features (erosion or ulceration, polymorphs in lamina propria, cryptitis,
crypt abscess formation, inflammatory exudates (luminal pus) and granulomas; one
point for each feature) with a maximum total score of six. For each patient, the
biopsy with the highest eAIS was used for analysis.
MRI protocol
For the VIGOR++ study an extended MRI protocol with specialized sequences and
additional bowel preparation for colonic distention was performed. The preparation
consisted of four hours of fasting prior to the examination and oral ingestion of
800 ml Mannitol solution (2.5%; Baxter, Utrecht, the Netherlands; Thetford, United
Kingdom) three hours prior to the examination (to fill the colon), followed by an
additional 1600 ml Mannitol solution (2.5%) in one hour prior to the examination (to
fill the small bowel). All MRI examinations were performed at 3.0 T (Philips Intera or
Ingenia, Philips, Best, the Netherlands) with a 16-channel torso phased-array body
coil. The protocol consisted of coronal balanced gradient-echo (GE) and dynamic
coronal balanced turbo field echo (BTFE) sequences, followed by coronal and axial
T2-weighted single-shot fast spin-echo (SSFSE) sequences, the latter with and
without fat saturation (full details of the MRI protocol are found in Supplementary
data 1). An axial DWI sequence (b values = 0, 300, 600 s/mm2) was then performed,
followed by a coronal fat-saturated 3D T1-weighted spoiled gradient-echo (SPGE)
sequence. Subsequently, a dynamic contrast-enhanced (DCE) coronal 3D T1-
weighted fast SPGE sequence was performed. These DCE images were acquired in
the setting of a different study and not evaluated as part of the current report as
DCE evaluation is not incorporated in any of the scoring systems under study. Sixty
seconds after the start of the dynamic sequence intravenous gadolinium contrast
medium was administered using the standard contrast agent in the participating
centers (Gadovist 1.0 mmol/L, Bayer Schering Pharma, Berlin, Germany; Dotarem
0.5 mmol/L, Guerbet, Paris, France). Subsequently, contrast-enhanced coronal and
axial T1-weighted SPGE images were acquired, for approximately seven minutes
after contrast injection. To reduce bowel peristalsis, three separate doses of 10
Proefschrift2018-new2.indb 59 13/9/18 10:06
59
Comparison of MRI scoring systems
cases, biopsies were taken approximately 2 cm proximal to the ileocecal valve or
anastomosis (in the case of previous ileocecal resection). Biopsies were examined
in consensus by two experienced pathologists (LB and MRJ), who were unaware of
clinical or radiologic findings. Each biopsy was examined on at least three slices and
graded according to the eAIS (3). This grading system describes the presence of six
histological features (erosion or ulceration, polymorphs in lamina propria, cryptitis,
crypt abscess formation, inflammatory exudates (luminal pus) and granulomas; one
point for each feature) with a maximum total score of six. For each patient, the
biopsy with the highest eAIS was used for analysis.
MRI protocol
For the VIGOR++ study an extended MRI protocol with specialized sequences and
additional bowel preparation for colonic distention was performed. The preparation
consisted of four hours of fasting prior to the examination and oral ingestion of
800 ml Mannitol solution (2.5%; Baxter, Utrecht, the Netherlands; Thetford, United
Kingdom) three hours prior to the examination (to fill the colon), followed by an
additional 1600 ml Mannitol solution (2.5%) in one hour prior to the examination (to
fill the small bowel). All MRI examinations were performed at 3.0 T (Philips Intera or
Ingenia, Philips, Best, the Netherlands) with a 16-channel torso phased-array body
coil. The protocol consisted of coronal balanced gradient-echo (GE) and dynamic
coronal balanced turbo field echo (BTFE) sequences, followed by coronal and axial
T2-weighted single-shot fast spin-echo (SSFSE) sequences, the latter with and
without fat saturation (full details of the MRI protocol are found in Supplementary
data 1). An axial DWI sequence (b values = 0, 300, 600 s/mm2) was then performed,
followed by a coronal fat-saturated 3D T1-weighted spoiled gradient-echo (SPGE)
sequence. Subsequently, a dynamic contrast-enhanced (DCE) coronal 3D T1-
weighted fast SPGE sequence was performed. These DCE images were acquired in
the setting of a different study and not evaluated as part of the current report as
DCE evaluation is not incorporated in any of the scoring systems under study. Sixty
seconds after the start of the dynamic sequence intravenous gadolinium contrast
medium was administered using the standard contrast agent in the participating
centers (Gadovist 1.0 mmol/L, Bayer Schering Pharma, Berlin, Germany; Dotarem
0.5 mmol/L, Guerbet, Paris, France). Subsequently, contrast-enhanced coronal and
axial T1-weighted SPGE images were acquired, for approximately seven minutes
after contrast injection. To reduce bowel peristalsis, three separate doses of 10
Proefschrift2018-new2.indb 59 13/9/18 10:06
60
Chapter 3
mg butylscopolamine bromide (Buscopan, Boehringer Ingelheim) were applied at
equal intervals during the examination.
Image analysis
For each examination, first, MRI features were examined in the entire terminal ileum
(20 cm) for comparison to the CDEIS and for evaluation of interobserver agreement
(full definitions of MRI features given in Supplementary Data 2). Subsequently, this
evaluation was repeated for the most severe lesion (if present) in the distal terminal
ileum (last 5 cm) for comparison to the eAIS, as all biopsies were taken form this
area. All examinations were evaluated twice to evaluate interobserver agreement.
Examinations from AMC were individually evaluated by two observers (A and B),
while UCLH examinations were evaluated by a different two observers (C and D).
All radiologists had extensive experience in MR enterography (>1100, >500, >800,
and >1500 examinations, respectively). To evaluate the complete dataset, readers
of AMC and UCLH datasets were paired (Observer 1: A, C; Observer 2: B, D). An
individual hospital/reader analysis was also performed to evaluate accordance with
the combined analysis’ results. Luminal distension, defined as the percentage of
adequate distension of the terminal ileum, was graded on a 5-point scale (<20%,
20–40%, 40–60%, 60–80%, >80%).
Two research fellows (C.T.N., >200 MR enterography examinations; and C.P., >100
MR enterography examinations), guided by the radiologists’ findings, obtained
quantitative apparent diffusion coefficient (ADC) values of the terminal ileum by
drawing a region of interest (ROI) of 0.3–1.0 cm2 in the area of the highest signal
intensity on DWI. To determine the uniformity in ADC value between patients, an
ROI of at least 2.5 cm2 was drawn in the erector spinae muscles. Additionally, wall
signal intensity (SI) was measured on pre- and post-contrast sequences by ROI
placement at the location scored by the readers, exhibiting the most pronounced
enhancement. Relative contrast enhancement (RCE) was calculated from wall SI
values using a modified formula as described by Rimola et al. (5). In our experience,
using the background noise (SD) correction is not suitable for correction of the
signal noise ratio (SNR). We found that SD noise values outside the body are
highly dependent on the distance of the ROI from the body and gave inconsistent
measurements. For all RCE and a subset of ADC measurements, a scanner specific
scaling correction was applied (13). The different MRI scoring systems (CDMI,
Proefschrift2018-new2.indb 60 13/9/18 10:06
60
Chapter 3
mg butylscopolamine bromide (Buscopan, Boehringer Ingelheim) were applied at
equal intervals during the examination.
Image analysis
For each examination, first, MRI features were examined in the entire terminal ileum
(20 cm) for comparison to the CDEIS and for evaluation of interobserver agreement
(full definitions of MRI features given in Supplementary Data 2). Subsequently, this
evaluation was repeated for the most severe lesion (if present) in the distal terminal
ileum (last 5 cm) for comparison to the eAIS, as all biopsies were taken form this
area. All examinations were evaluated twice to evaluate interobserver agreement.
Examinations from AMC were individually evaluated by two observers (A and B),
while UCLH examinations were evaluated by a different two observers (C and D).
All radiologists had extensive experience in MR enterography (>1100, >500, >800,
and >1500 examinations, respectively). To evaluate the complete dataset, readers
of AMC and UCLH datasets were paired (Observer 1: A, C; Observer 2: B, D). An
individual hospital/reader analysis was also performed to evaluate accordance with
the combined analysis’ results. Luminal distension, defined as the percentage of
adequate distension of the terminal ileum, was graded on a 5-point scale (<20%,
20–40%, 40–60%, 60–80%, >80%).
Two research fellows (C.T.N., >200 MR enterography examinations; and C.P., >100
MR enterography examinations), guided by the radiologists’ findings, obtained
quantitative apparent diffusion coefficient (ADC) values of the terminal ileum by
drawing a region of interest (ROI) of 0.3–1.0 cm2 in the area of the highest signal
intensity on DWI. To determine the uniformity in ADC value between patients, an
ROI of at least 2.5 cm2 was drawn in the erector spinae muscles. Additionally, wall
signal intensity (SI) was measured on pre- and post-contrast sequences by ROI
placement at the location scored by the readers, exhibiting the most pronounced
enhancement. Relative contrast enhancement (RCE) was calculated from wall SI
values using a modified formula as described by Rimola et al. (5). In our experience,
using the background noise (SD) correction is not suitable for correction of the
signal noise ratio (SNR). We found that SD noise values outside the body are
highly dependent on the distance of the ROI from the body and gave inconsistent
measurements. For all RCE and a subset of ADC measurements, a scanner specific
scaling correction was applied (13). The different MRI scoring systems (CDMI,
Proefschrift2018-new2.indb 60 13/9/18 10:06
61
Comparison of MRI scoring systems
London, MaRIA and Clermont score) were calculated using the formulas described
in the Supplementary Data 2 (3,5,8).
Statistical analysis
Interobserver agreement was calculated for ordinal data using a weighted
kappa coefficient, for binominal data using a multirater kappa coefficient, and
for continuous data using a multirater intraclass correlation coefficient. Kappa
and intraclass correlation coefficient values were interpreted using the following
criteria: ≤0.20, poor; ≥0.21–0.40, fair; ≥0.41–0.60, moderate; ≥0.61–0.80, good; ≥0.81–
1.00,very good (14). Spearman’s rank correlation for non-normal distributions was
used to test associations between MRI scoring systems/individual features and
the reference standards (eAIS and CDEIS). Non-normality was assessed by visual
interpretation of histograms and descriptive statistics. Interpretation of Spearman’s
correlation coefficient was done as follows: >0–<0.20, very weak; ≥0.20–<0.40,
weak; ≥0.40–<0.60, moderate; ≥0.60–<0.80, strong; ≥0.80–<1.00, very strong. To
evaluate the ability to predict active disease the eAIS was dichotomized using ≥2 as
a cut-off (3). For CDEIS a cut-off of ≥3 was used (15). Ulcerating endoscopic disease
was defined as the presence of superficial or deep ulcers at endoscopy (5). Cut-off
points for active and ulcerating disease for MRI scoring systems were obtained from
the original literature for the MaRIA (7 and 11, respectively), the Clermont score (8.4
and 12.5, respectively) and the CDMI (3 for active disease; no cut-off for ulcerating
disease) (3,5,8). Cut-off points for the London score were obtained from a recent
validation study, as this provided both values for active and ulcerating disease (3.4
and 3.8, respectively) (10). For the CDMI, an optimal cut-off point for ulcerating
disease was obtained using receiver-operating characteristics (ROC) analysis, while
the area under the curve (AUC) was calculated. Steiger Z-test was used to compare
correlations between the different scoring systems and eAIS/CDEIS. Sensitivity,
specificity and accuracy were compared using the McNemar test. Discordant
findings between different scoring systems were assessed by identifying outliers
on score-to-score scatter plots with a linear regression fit and associated 95%
confidence intervals.
A value of p<0.05 was considered statistically significant. All statistical analyses
were performed with IBM SPSS Statistics version 22.0 for Mac (SPSS, Chicago, III,
USA) and Vassarstats.net (Richard Lowry, Poughkeepsie, NY, USA).
Proefschrift2018-new2.indb 61 13/9/18 10:06
61
Comparison of MRI scoring systems
London, MaRIA and Clermont score) were calculated using the formulas described
in the Supplementary Data 2 (3,5,8).
Statistical analysis
Interobserver agreement was calculated for ordinal data using a weighted
kappa coefficient, for binominal data using a multirater kappa coefficient, and
for continuous data using a multirater intraclass correlation coefficient. Kappa
and intraclass correlation coefficient values were interpreted using the following
criteria: ≤0.20, poor; ≥0.21–0.40, fair; ≥0.41–0.60, moderate; ≥0.61–0.80, good; ≥0.81–
1.00,very good (14). Spearman’s rank correlation for non-normal distributions was
used to test associations between MRI scoring systems/individual features and
the reference standards (eAIS and CDEIS). Non-normality was assessed by visual
interpretation of histograms and descriptive statistics. Interpretation of Spearman’s
correlation coefficient was done as follows: >0–<0.20, very weak; ≥0.20–<0.40,
weak; ≥0.40–<0.60, moderate; ≥0.60–<0.80, strong; ≥0.80–<1.00, very strong. To
evaluate the ability to predict active disease the eAIS was dichotomized using ≥2 as
a cut-off (3). For CDEIS a cut-off of ≥3 was used (15). Ulcerating endoscopic disease
was defined as the presence of superficial or deep ulcers at endoscopy (5). Cut-off
points for active and ulcerating disease for MRI scoring systems were obtained from
the original literature for the MaRIA (7 and 11, respectively), the Clermont score (8.4
and 12.5, respectively) and the CDMI (3 for active disease; no cut-off for ulcerating
disease) (3,5,8). Cut-off points for the London score were obtained from a recent
validation study, as this provided both values for active and ulcerating disease (3.4
and 3.8, respectively) (10). For the CDMI, an optimal cut-off point for ulcerating
disease was obtained using receiver-operating characteristics (ROC) analysis, while
the area under the curve (AUC) was calculated. Steiger Z-test was used to compare
correlations between the different scoring systems and eAIS/CDEIS. Sensitivity,
specificity and accuracy were compared using the McNemar test. Discordant
findings between different scoring systems were assessed by identifying outliers
on score-to-score scatter plots with a linear regression fit and associated 95%
confidence intervals.
A value of p<0.05 was considered statistically significant. All statistical analyses
were performed with IBM SPSS Statistics version 22.0 for Mac (SPSS, Chicago, III,
USA) and Vassarstats.net (Richard Lowry, Poughkeepsie, NY, USA).
Proefschrift2018-new2.indb 61 13/9/18 10:06
62
Chapter 3
RESULTS
Study population
Of the 158 patients eligible for participation, 44 patients were excluded from
the study as detailed in Figure 1. Finally, 98 patients with known Crohn’s disease
and complete ileocolonoscopy and biopsies from the distal terminal ileum were
included. All of these patients underwent the complete MRI protocol, except for five
patients, where the ADC map was missing (n=2) or the distal terminal ileum was
not in the FOV of the DWI sequence (n=3). These patients were not excluded for
analyses of the other features. No mismatches between MRI and endoscopy were
identified (i.e. more proximal disease on MRI not seen during endoscopic intubation
of the terminal ileum). Characteristics of all included patients are described in Table
1. As stated earlier, part of our study population has been published before (11). The
total overlap with study consisted of 92 patients. The remaining 6 patients were
excluded from the previous study, due to missing dynamic contrast enhancement
sequences (n=6), which was not an exclusion criterion for the current study. The
eAIS ranged from 0 to 4 with a median of 2 (IQR:0–3) (no biopsies were graded
with score 5 or 6). Segmental CDEIS for the terminal ileum ranged from 0 to 36,
with a median of 8.0 (IQR: 0–16.3). Mean luminal distention values (scale 0–4) for
observer 1 and observer 2 were 3.40 (SD: 0.92) and 3.37 (SD: 0.88), respectively.
Figure 1. Flow diagram detailing patient inclusion and exclusion.
Proefschrift2018-new2.indb 62 13/9/18 10:06
62
Chapter 3
RESULTS
Study population
Of the 158 patients eligible for participation, 44 patients were excluded from
the study as detailed in Figure 1. Finally, 98 patients with known Crohn’s disease
and complete ileocolonoscopy and biopsies from the distal terminal ileum were
included. All of these patients underwent the complete MRI protocol, except for five
patients, where the ADC map was missing (n=2) or the distal terminal ileum was
not in the FOV of the DWI sequence (n=3). These patients were not excluded for
analyses of the other features. No mismatches between MRI and endoscopy were
identified (i.e. more proximal disease on MRI not seen during endoscopic intubation
of the terminal ileum). Characteristics of all included patients are described in Table
1. As stated earlier, part of our study population has been published before (11). The
total overlap with study consisted of 92 patients. The remaining 6 patients were
excluded from the previous study, due to missing dynamic contrast enhancement
sequences (n=6), which was not an exclusion criterion for the current study. The
eAIS ranged from 0 to 4 with a median of 2 (IQR:0–3) (no biopsies were graded
with score 5 or 6). Segmental CDEIS for the terminal ileum ranged from 0 to 36,
with a median of 8.0 (IQR: 0–16.3). Mean luminal distention values (scale 0–4) for
observer 1 and observer 2 were 3.40 (SD: 0.92) and 3.37 (SD: 0.88), respectively.
Figure 1. Flow diagram detailing patient inclusion and exclusion.
Proefschrift2018-new2.indb 62 13/9/18 10:06
63
Comparison of MRI scoring systems
Table 1. Patient characteristics
Variable N [%]
Total no. of patients 98
Gender, female, n [%] 55 [56]
Age at MRI in years, median [IQR] 32 [26–44]
Disease duration in years, median [IQR] 8 [4–15]
Days between ileocolonoscopy and MRI, median [IQR] 4 [1–7]
History of ileocecal resection, n [%] 30 [31]
CDEIS for the terminal ileum, median [IQR] 8.0 [0–16.3]
eAIS for the terminal ileum, median [IQR] 2 [0–3]
Therapy
Anti-tumor necrosis factor, n [%] 27 [27]
Steroids, n [%] 16 [16]
Purine-antagonists, n [%] 16 [16]
5-ASA, n [%] 16 [16]
Methotrexate, n [%] 4 [4]
Harvey-Bradshaw index, median [IQR] 5 [2–8]
Harvey-Bradshaw index >4, n [%] 56 [57]
CRP [mg/L], median [IQR] 3.8 [1.2–11.3]
5-ASA, 5-aminosalicylic acid; CRP, C-reactive protein; eAIS, acute inflammation score; IQR, interquartile range; MRI, magnetic resonance imaging.
MRI scoring systems
All four scoring systems showed good interobserver agreement (ICC range: 0.70–
0.78) (Table 2). All four scoring systems showed moderate-to-strong correlation to
CDEIS and moderate correlation to the eAIS, while no significant differences were
found between the correlation coefficients, except for a higher correlation to CDEIS
for the London score compared to the MaRIA for observer 1 (p=0.04). An individual
analysis for each hospital is detailed in the Supplementary Data 3. AMC data showed
consistently higher correlations to CDEIS (r=0.56–0.73) and eAIS (r=0.48–0.53),
compared to UCLH data (r= 0.36–0.64 and r=0.26–0.30, respectively). However,
similar interobserver agreement was seen in both hospitals (ICC range AMC:
0.69–0.80, vs UCLH: 0.71–0.74). When comparing the disease spectrum of AMC
and UCLH patients, significant differences were found in median CDEIS (9.0 vs 2.5,
p=0.02) and median eAIS scores (2.0 vs 1.0, p=0.003).
Proefschrift2018-new2.indb 63 13/9/18 10:06
63
Comparison of MRI scoring systems
Table 1. Patient characteristics
Variable N [%]
Total no. of patients 98
Gender, female, n [%] 55 [56]
Age at MRI in years, median [IQR] 32 [26–44]
Disease duration in years, median [IQR] 8 [4–15]
Days between ileocolonoscopy and MRI, median [IQR] 4 [1–7]
History of ileocecal resection, n [%] 30 [31]
CDEIS for the terminal ileum, median [IQR] 8.0 [0–16.3]
eAIS for the terminal ileum, median [IQR] 2 [0–3]
Therapy
Anti-tumor necrosis factor, n [%] 27 [27]
Steroids, n [%] 16 [16]
Purine-antagonists, n [%] 16 [16]
5-ASA, n [%] 16 [16]
Methotrexate, n [%] 4 [4]
Harvey-Bradshaw index, median [IQR] 5 [2–8]
Harvey-Bradshaw index >4, n [%] 56 [57]
CRP [mg/L], median [IQR] 3.8 [1.2–11.3]
5-ASA, 5-aminosalicylic acid; CRP, C-reactive protein; eAIS, acute inflammation score; IQR, interquartile range; MRI, magnetic resonance imaging.
MRI scoring systems
All four scoring systems showed good interobserver agreement (ICC range: 0.70–
0.78) (Table 2). All four scoring systems showed moderate-to-strong correlation to
CDEIS and moderate correlation to the eAIS, while no significant differences were
found between the correlation coefficients, except for a higher correlation to CDEIS
for the London score compared to the MaRIA for observer 1 (p=0.04). An individual
analysis for each hospital is detailed in the Supplementary Data 3. AMC data showed
consistently higher correlations to CDEIS (r=0.56–0.73) and eAIS (r=0.48–0.53),
compared to UCLH data (r= 0.36–0.64 and r=0.26–0.30, respectively). However,
similar interobserver agreement was seen in both hospitals (ICC range AMC:
0.69–0.80, vs UCLH: 0.71–0.74). When comparing the disease spectrum of AMC
and UCLH patients, significant differences were found in median CDEIS (9.0 vs 2.5,
p=0.02) and median eAIS scores (2.0 vs 1.0, p=0.003).
Proefschrift2018-new2.indb 63 13/9/18 10:06
64
Chapter 3
Table 2. Interobserver agreement and correlation to eAIS/CDEIS for all MRI scoring systems.
Interobserver agree-ment
ICC (95% CI)
CDMI 0.78 (0.69–0.85)
London score 0.72 (0.61–0.81)
MaRIA 0.70(0.58–0.79)
Clermont score 0.71 (0.59–0.79)
Observer 1 Observer 2
Correlation to eAIS Coefficient p-Value Coefficient p-Value
CDMI 0.48 <0.001 0.41 <0.001
London score 0.49 <0.001 0.39 <0.001
MaRIA 0.42 <0.001 0.38 <0.001
Clermont score 0.46 <0.001 0.42 <0.001
Observer 1 Observer 2
Correlation to CDEIS Coefficient p-Value Coefficient p-Value
CDMI 0.57 <0.001 0.67 <0.001
London score 0.59 <0.001 0.64 <0.001
MaRIA 0.60 <0.001 0.64 <0.001
Clermont score 0.62 <0.001 0.66 <0.001
ICC, intraclass correlation coefficient, CI, confidence interval; eAIS, acute inflammation score; CDMI, Crohn’s disease MRI index; MaRIA, Magnetic Resonance Index of Activity.
Proefschrift2018-new2.indb 64 13/9/18 10:06
64
Chapter 3
Table 2. Interobserver agreement and correlation to eAIS/CDEIS for all MRI scoring systems.
Interobserver agree-ment
ICC (95% CI)
CDMI 0.78 (0.69–0.85)
London score 0.72 (0.61–0.81)
MaRIA 0.70(0.58–0.79)
Clermont score 0.71 (0.59–0.79)
Observer 1 Observer 2
Correlation to eAIS Coefficient p-Value Coefficient p-Value
CDMI 0.48 <0.001 0.41 <0.001
London score 0.49 <0.001 0.39 <0.001
MaRIA 0.42 <0.001 0.38 <0.001
Clermont score 0.46 <0.001 0.42 <0.001
Observer 1 Observer 2
Correlation to CDEIS Coefficient p-Value Coefficient p-Value
CDMI 0.57 <0.001 0.67 <0.001
London score 0.59 <0.001 0.64 <0.001
MaRIA 0.60 <0.001 0.64 <0.001
Clermont score 0.62 <0.001 0.66 <0.001
ICC, intraclass correlation coefficient, CI, confidence interval; eAIS, acute inflammation score; CDMI, Crohn’s disease MRI index; MaRIA, Magnetic Resonance Index of Activity.
Proefschrift2018-new2.indb 64 13/9/18 10:06
65
Comparison of MRI scoring systems
Tab
le 3
. Dia
gn
ost
ic a
ccu
racy f
or
acti
ve h
isto
path
olo
gic
(eA
IS≥2
) o
r en
do
sco
pic
dis
ease
(C
DE
IS≥3
) fo
r all
MR
I sc
ori
ng
syst
em
s.
CD
EIS
Ob
serv
er 1
Ob
serv
er 2
Cut
-off
p
oin
tSe
nsit
ivit
y (%
)Sp
ecifi
city
(%
)P
PV
(%
)N
PV
(%
)A
ccur
acy
(%)*
Sens
itiv
ity
(%)
Spec
ifici
ty
(%)
PP
V (
%)
NP
V (
%)
Acc
urac
y (%
)*
CD
MI
379
(4
6/5
8)
63
(25/4
0)
75
(4
6/6
1)6
8 (
25/3
7)
72 (
71/
98
)74
(4
3/5
8)
55
(22/4
0)
70
(4
3/6
1)5
9 (
22/3
7)
66
(6
5/9
8)
Lo
nd
on
sco
re3
.479
(4
6/5
8)
63
(25/4
0)
75
(4
6/6
1)6
8 (
25/3
7)
72 (
71/
98
)76
(4
4/5
8)
55
(22/4
0)
71
(44
/62)
61
(22/3
6)
67 (
66
/98
)
MaR
IA
78
1 (4
7/5
8)
53
(21/
40
)7
1 (4
7/6
6)
66
(21/
32)
69
(6
8/9
8)
74 (
44
/58
)5
0 (
20
/40
)6
9 (
44
/64
)5
9 (
20
/34
)6
5 (
64
/98
)
Cle
rmo
nt
sco
re8
.48
2 (
46
/56
)5
5 (
21/
38
)73
(4
6/6
3)
68
(21/
31)
71
(67/
94
)76
(4
4/5
6)
47 (
18/3
8)
69
(4
4/6
4)
60
(18
/30
)6
6 (
62/9
4)
eAIS
Ob
serv
er 1
Ob
serv
er 2
Cut
-off
p
oin
tSe
nsit
ivit
y (%
)Sp
ecifi
city
(%
)P
PV
(%
)N
PV
(%
)A
ccur
acy
(%)
Sens
itiv
ity
(%)
Spec
ifici
ty
(%)
PP
V (
%)
NP
V (
%)
Acc
urac
y (%
)
CD
MI
38
2 (
45/5
5)
71
(24
/34
)8
2 (
45/5
5)
71
(24
/34
)78
(6
9/8
9)
77 (
43
/56
)6
8 (
23
/34
)8
0 (
43
/54
)6
4 (
23
/36
)73
(6
6/9
0)
Lo
nd
on
sco
re3
.48
2 (
45/5
5)
71
(24
/34
)8
2 (
45/5
5)
71
(24
/34
)78
(6
9/8
9)
79
(4
4/5
6)
68
(23
/34
)8
0 (
44
/55
)6
6 (
23
/35
)74
(6
7/9
0)
MaR
IA
78
4 (
46
/55
)6
2 (
21/
34
)78
(4
6/5
9)
70
(21/
30
)75
(57/
89
)8
0 (
45/5
6)
65
(22/3
4)
79
(4
5/5
7)
67 (
22/3
3)
74 (
67/
90
)
Cle
rmo
nt
sco
re8
.48
5 (
44
/52)
64
(21/
33
)79
(4
4/5
6)
72 (
21/
29
)76
(5
5/8
5)
83
(4
4/5
3)
64
(21/
33
)79
(4
4/5
6)
70
(21/
30
)76
(6
5/8
6)
CD
MI, C
roh
n's
Dis
ease
MR
I in
dex; M
aR
IA, M
ag
neti
c R
eso
nan
ce In
dex o
f A
cti
vit
y; P
PV
, p
osi
tive p
red
icti
ve v
alu
e; N
PV
, n
eg
ati
ve p
red
icti
ve v
alu
e*
Th
e C
lerm
on
t sc
ore
was
no
t sc
ore
d in
4/9
8 s
eg
men
ts d
ue t
o a
rtefa
cts
on
DW
I. In
th
e 9
0 p
ati
en
ts w
ith
CD
EIS
sco
res
the t
erm
inal ile
um
co
uld
be in
tub
ate
d, w
hile
in
th
e r
em
ain
ing
8
pati
en
ts o
nly
bio
psi
es
co
uld
be r
etr
ieved
th
rou
gh
th
e ile
ocecal valv
e. F
or
co
rrela
tio
n t
o C
DE
IS, o
bse
rver
1 d
eem
ed
on
e s
eg
men
t as
no
n-e
valu
ab
le o
ver
the m
ost
of
it 2
0 c
m len
gth
du
e t
o
art
efa
cts
(th
is s
eg
men
t w
as
sco
red
as
no
acti
vit
y b
y o
bse
rver
2).
Tab
le 4
. Dia
gn
ost
ic a
ccu
racy f
or
severe
en
do
sco
pic
dis
ease
(p
rese
nce o
f u
lcers
at
en
do
sco
py)
for
the L
on
do
n s
co
re, M
aR
IA a
nd
Cle
rmo
nt
sco
re.
Ob
serv
er 1
Ob
serv
er 2
Cut
-off
po
int
Sens
itiv
ity
(%)
Spec
ifici
ty
(%)
PP
V (
%)
NP
V (
%)
Acc
urac
y (%
)*Se
nsit
ivit
y (%
)Sp
ecifi
city
(%
)P
PV
(%
)N
PV
(%
)A
ccur
acy
(%)*
Lo
nd
on
sco
re3
.88
3 (
40
/48
)6
3 (
26
/41)
73
(4
0/5
5)
76
(26
/34
)74
(6
6/8
9)
80
(3
9/4
9)
61
(25/4
1)7
1 (3
9/5
5)
71
(25/3
5)
71
(64
/90
)
MaR
IA
118
3 (
40
/48
)6
8 (
28
/41)
75
(4
0/5
3)
78
(28
/36
)76
(6
8/8
9)
76
(37/
49
)6
6 (
27/
41)
73
(37/
51)
69
(27/
39
)7
1 (6
4/9
0)
Cle
rmo
nt
sco
re12
.58
3 (
38
/46
)6
7(2
6/3
9)
75
(3
8/5
1)76
(26
/34
)75
(6
4/8
5)
79
(37/
47)
67 (
26
/39
)74
(37/
50
)72 (
26
/36
)73
(6
3/9
0)
CD
MI, C
roh
n’s
Dis
ease
MR
I in
dex; M
aR
IA, M
ag
neti
c R
eso
nan
ce In
dex o
f A
cti
vit
y; P
PV
, p
osi
tive p
red
icti
ve v
alu
e; N
PV
, n
eg
ati
ve p
red
icti
ve v
alu
e*
Th
e C
lerm
on
t sc
ore
was
no
t sc
ore
d in
4/9
8 s
eg
men
ts d
ue t
o a
rtefa
cts
on
DW
I. F
or
co
rrela
tio
n t
o C
DE
IS, o
bse
rver
1 d
eem
ed
on
e s
eg
men
t as
no
n-e
valu
ab
le o
ver
the m
ost
of
it 2
0 c
m
len
gth
du
e t
o a
rtefa
cts
(th
is s
eg
men
t w
as
sco
red
as
no
acti
vit
y b
y o
bse
rver
2).
Proefschrift2018-new2.indb 65 13/9/18 10:06
65
Comparison of MRI scoring systems
Tab
le 3
. Dia
gn
ost
ic a
ccu
racy f
or
acti
ve h
isto
path
olo
gic
(eA
IS≥2
) o
r en
do
sco
pic
dis
ease
(C
DE
IS≥3
) fo
r all
MR
I sc
ori
ng
syst
em
s.
CD
EIS
Ob
serv
er 1
Ob
serv
er 2
Cut
-off
p
oin
tSe
nsit
ivit
y (%
)Sp
ecifi
city
(%
)P
PV
(%
)N
PV
(%
)A
ccur
acy
(%)*
Sens
itiv
ity
(%)
Spec
ifici
ty
(%)
PP
V (
%)
NP
V (
%)
Acc
urac
y (%
)*
CD
MI
379
(4
6/5
8)
63
(25/4
0)
75
(4
6/6
1)6
8 (
25/3
7)
72 (
71/
98
)74
(4
3/5
8)
55
(22/4
0)
70
(4
3/6
1)5
9 (
22/3
7)
66
(6
5/9
8)
Lo
nd
on
sco
re3
.479
(4
6/5
8)
63
(25/4
0)
75
(4
6/6
1)6
8 (
25/3
7)
72 (
71/
98
)76
(4
4/5
8)
55
(22/4
0)
71
(44
/62)
61
(22/3
6)
67 (
66
/98
)
MaR
IA
78
1 (4
7/5
8)
53
(21/
40
)7
1 (4
7/6
6)
66
(21/
32)
69
(6
8/9
8)
74 (
44
/58
)5
0 (
20
/40
)6
9 (
44
/64
)5
9 (
20
/34
)6
5 (
64
/98
)
Cle
rmo
nt
sco
re8
.48
2 (
46
/56
)5
5 (
21/
38
)73
(4
6/6
3)
68
(21/
31)
71
(67/
94
)76
(4
4/5
6)
47 (
18/3
8)
69
(4
4/6
4)
60
(18
/30
)6
6 (
62/9
4)
eAIS
Ob
serv
er 1
Ob
serv
er 2
Cut
-off
p
oin
tSe
nsit
ivit
y (%
)Sp
ecifi
city
(%
)P
PV
(%
)N
PV
(%
)A
ccur
acy
(%)
Sens
itiv
ity
(%)
Spec
ifici
ty
(%)
PP
V (
%)
NP
V (
%)
Acc
urac
y (%
)
CD
MI
38
2 (
45/5
5)
71
(24
/34
)8
2 (
45/5
5)
71
(24
/34
)78
(6
9/8
9)
77 (
43
/56
)6
8 (
23
/34
)8
0 (
43
/54
)6
4 (
23
/36
)73
(6
6/9
0)
Lo
nd
on
sco
re3
.48
2 (
45/5
5)
71
(24
/34
)8
2 (
45/5
5)
71
(24
/34
)78
(6
9/8
9)
79
(4
4/5
6)
68
(23
/34
)8
0 (
44
/55
)6
6 (
23
/35
)74
(6
7/9
0)
MaR
IA
78
4 (
46
/55
)6
2 (
21/
34
)78
(4
6/5
9)
70
(21/
30
)75
(57/
89
)8
0 (
45/5
6)
65
(22/3
4)
79
(4
5/5
7)
67 (
22/3
3)
74 (
67/
90
)
Cle
rmo
nt
sco
re8
.48
5 (
44
/52)
64
(21/
33
)79
(4
4/5
6)
72 (
21/
29
)76
(5
5/8
5)
83
(4
4/5
3)
64
(21/
33
)79
(4
4/5
6)
70
(21/
30
)76
(6
5/8
6)
CD
MI, C
roh
n's
Dis
ease
MR
I in
dex; M
aR
IA, M
ag
neti
c R
eso
nan
ce In
dex o
f A
cti
vit
y; P
PV
, p
osi
tive p
red
icti
ve v
alu
e; N
PV
, n
eg
ati
ve p
red
icti
ve v
alu
e*
Th
e C
lerm
on
t sc
ore
was
no
t sc
ore
d in
4/9
8 s
eg
men
ts d
ue t
o a
rtefa
cts
on
DW
I. In
th
e 9
0 p
ati
en
ts w
ith
CD
EIS
sco
res
the t
erm
inal ile
um
co
uld
be in
tub
ate
d, w
hile
in
th
e r
em
ain
ing
8
pati
en
ts o
nly
bio
psi
es
co
uld
be r
etr
ieved
th
rou
gh
th
e ile
ocecal valv
e. F
or
co
rrela
tio
n t
o C
DE
IS, o
bse
rver
1 d
eem
ed
on
e s
eg
men
t as
no
n-e
valu
ab
le o
ver
the m
ost
of
it 2
0 c
m len
gth
du
e t
o
art
efa
cts
(th
is s
eg
men
t w
as
sco
red
as
no
acti
vit
y b
y o
bse
rver
2).
Tab
le 4
. Dia
gn
ost
ic a
ccu
racy f
or
severe
en
do
sco
pic
dis
ease
(p
rese
nce o
f u
lcers
at
en
do
sco
py)
for
the L
on
do
n s
co
re, M
aR
IA a
nd
Cle
rmo
nt
sco
re.
Ob
serv
er 1
Ob
serv
er 2
Cut
-off
po
int
Sens
itiv
ity
(%)
Spec
ifici
ty
(%)
PP
V (
%)
NP
V (
%)
Acc
urac
y (%
)*Se
nsit
ivit
y (%
)Sp
ecifi
city
(%
)P
PV
(%
)N
PV
(%
)A
ccur
acy
(%)*
Lo
nd
on
sco
re3
.88
3 (
40
/48
)6
3 (
26
/41)
73
(4
0/5
5)
76
(26
/34
)74
(6
6/8
9)
80
(3
9/4
9)
61
(25/4
1)7
1 (3
9/5
5)
71
(25/3
5)
71
(64
/90
)
MaR
IA
118
3 (
40
/48
)6
8 (
28
/41)
75
(4
0/5
3)
78
(28
/36
)76
(6
8/8
9)
76
(37/
49
)6
6 (
27/
41)
73
(37/
51)
69
(27/
39
)7
1 (6
4/9
0)
Cle
rmo
nt
sco
re12
.58
3 (
38
/46
)6
7(2
6/3
9)
75
(3
8/5
1)76
(26
/34
)75
(6
4/8
5)
79
(37/
47)
67 (
26
/39
)74
(37/
50
)72 (
26
/36
)73
(6
3/9
0)
CD
MI, C
roh
n’s
Dis
ease
MR
I in
dex; M
aR
IA, M
ag
neti
c R
eso
nan
ce In
dex o
f A
cti
vit
y; P
PV
, p
osi
tive p
red
icti
ve v
alu
e; N
PV
, n
eg
ati
ve p
red
icti
ve v
alu
e*
Th
e C
lerm
on
t sc
ore
was
no
t sc
ore
d in
4/9
8 s
eg
men
ts d
ue t
o a
rtefa
cts
on
DW
I. F
or
co
rrela
tio
n t
o C
DE
IS, o
bse
rver
1 d
eem
ed
on
e s
eg
men
t as
no
n-e
valu
ab
le o
ver
the m
ost
of
it 2
0 c
m
len
gth
du
e t
o a
rtefa
cts
(th
is s
eg
men
t w
as
sco
red
as
no
acti
vit
y b
y o
bse
rver
2).
Proefschrift2018-new2.indb 65 13/9/18 10:06
66
Chapter 3
In the terminal ileum, active endoscopic disease (CDEIS≥3) was present in 56 patients
(57%), while active histopathologic disease (eAIS≥2) was present in 58 patients
(59%). Ulcerating endoscopic disease (presence of ulcers) was seen in the terminal
ileum of 49 patients (50%). Diagnostic accuracy for detection of histopathologic
and endoscopic disease activity are presented in in Tables 3 and 4. No significant
differences in sensitivity, specificity or accuracy were found for diagnosis of either
active or ulcerating disease (p>0.05).
For detection of ulcerating endoscopic disease using the CDMI, an optimal cut-off
value of 5 was determined using receiver-operating characteristics analysis, which
provided an AUC of 0.75, and sensitivity/specificity of 73% / 76% for observer 1 and
76% / 76% for observer 2, respectively.
Very good agreement was seen between the CDMI and London score (ICC Ob1/2:
0.89/0.90) and between the MaRIA and Clermont score (ICC Ob1/2: 0.99/0.99).
In contrast, only fair-to-good agreement when pairing either the CDMI or London
score to the MaRIA or Clermont score (ICC range: 0.44–0.65). These pairings also
accounted for all evident outliers (n=12), which were identified during score-to-
score comparison. All outliers with disproportionally high CDMI or London scores
(n=7) showed high grades (2 of 3) for mural T2 signal, mural thickness and T1
enhancement (CDMI only), while ulcers were absent in all cases. In contrast, outliers
with disproportionally high MaRIA or Clermont scores (n=4) all showed presence
of ulcers. Two moderate outliers found between the CDMI and London score and
additional minor outliers (n=7, i.e. with very close proximity to the 95% confidence
interval) did not provide additional insights. No outliers were found between the
MaRIA and Clermont score.
MRI features
Interobserver agreement and correlation to CDEIS and eAIS of the individual
MRI features incorporated in the different scoring systems are given in the
Supplementary data 4 and 5, respectively. Enhancement and RCE showed good
interobserver agreement (kappa: 0.66 (95% CI: 0.55–77) and ICC: 0.67 (95% CI:
0.55–0.77), respectively), while wall thickness (mm), ADC, mural T2 signal and
perimural T2 signal showed moderate agreement (ICC: 0.54 (95% CI: 0.22–0.72),
ICC: 0.51 (95% CI: 0.32–0.62), kappa: 0.45 (95% CI: 0.33–0.57) and kappa: 0.52 (95%
Proefschrift2018-new2.indb 66 13/9/18 10:06
66
Chapter 3
In the terminal ileum, active endoscopic disease (CDEIS≥3) was present in 56 patients
(57%), while active histopathologic disease (eAIS≥2) was present in 58 patients
(59%). Ulcerating endoscopic disease (presence of ulcers) was seen in the terminal
ileum of 49 patients (50%). Diagnostic accuracy for detection of histopathologic
and endoscopic disease activity are presented in in Tables 3 and 4. No significant
differences in sensitivity, specificity or accuracy were found for diagnosis of either
active or ulcerating disease (p>0.05).
For detection of ulcerating endoscopic disease using the CDMI, an optimal cut-off
value of 5 was determined using receiver-operating characteristics analysis, which
provided an AUC of 0.75, and sensitivity/specificity of 73% / 76% for observer 1 and
76% / 76% for observer 2, respectively.
Very good agreement was seen between the CDMI and London score (ICC Ob1/2:
0.89/0.90) and between the MaRIA and Clermont score (ICC Ob1/2: 0.99/0.99).
In contrast, only fair-to-good agreement when pairing either the CDMI or London
score to the MaRIA or Clermont score (ICC range: 0.44–0.65). These pairings also
accounted for all evident outliers (n=12), which were identified during score-to-
score comparison. All outliers with disproportionally high CDMI or London scores
(n=7) showed high grades (2 of 3) for mural T2 signal, mural thickness and T1
enhancement (CDMI only), while ulcers were absent in all cases. In contrast, outliers
with disproportionally high MaRIA or Clermont scores (n=4) all showed presence
of ulcers. Two moderate outliers found between the CDMI and London score and
additional minor outliers (n=7, i.e. with very close proximity to the 95% confidence
interval) did not provide additional insights. No outliers were found between the
MaRIA and Clermont score.
MRI features
Interobserver agreement and correlation to CDEIS and eAIS of the individual
MRI features incorporated in the different scoring systems are given in the
Supplementary data 4 and 5, respectively. Enhancement and RCE showed good
interobserver agreement (kappa: 0.66 (95% CI: 0.55–77) and ICC: 0.67 (95% CI:
0.55–0.77), respectively), while wall thickness (mm), ADC, mural T2 signal and
perimural T2 signal showed moderate agreement (ICC: 0.54 (95% CI: 0.22–0.72),
ICC: 0.51 (95% CI: 0.32–0.62), kappa: 0.45 (95% CI: 0.33–0.57) and kappa: 0.52 (95%
Proefschrift2018-new2.indb 66 13/9/18 10:06
67
Comparison of MRI scoring systems
CI: 0.33–0.70), respectively) and ulcers showed only fair agreement (kappa: 0.26
(95% CI: 0.06–0.46)). ADC reference measurements in the erector spinae muscle
for observers 1 and 2 showed mean values of 1618 (SD: 133, range: 1194–1893) and
1618 (SD: 209, range: 1216–2491), respectively. Most MRI features showed moderate-
to-strong correlation to CDEIS (r=0.44–0.63), while perimural T2 signal, ulcers and
RCE showed weak-to-moderate correlation to CDEIS (r=0.27–0.49). Most features
showed weak-to-moderate correlation to eAIS for both observers (r=0.31–0.51),
while ulcers and RCE showed very weak-to-weak correlation to eAIS (r=0.10–0.23).
DISCUSSION
In this study we evaluated and compared four validated MRI scoring systems and
their incorporated MRI features for grading of inflammation in Crohn’s disease. All
four MRI scoring systems were comparable in terms of interobserver agreement
(ICC: 0.70–0.78), correlation to endoscopic and histopathological reference
standards (r=0.57–0.67 and r=0.38–0.49, respectively) and diagnostic accuracy for
active and ulcerating endoscopic disease (73–78% and 71–76%, respectively) and
active histopathologic disease (65–72%). The MRI evaluation of ulcers accounted
for almost all cases of discordant findings between the scores incorporating this
feature (MaRIA and Clermont score) and the scores not incorporating ulcers (CDMI
and London score). Enhancement and RCE showed good interobserver agreement,
while other MRI features showed fair-to-moderate agreement. The individual
features showed at best a moderate correlation to the eAIS, but moderate-to-strong
correlation to the CDEIS. There were differences in the strength of correlation to
CDEIS/eAIS between the two recruitment sites, but consistently good interobserver
agreement. It is likely that this observation stems from differences in disease
spectrum between the two sites; we found significantly higher disease activity in
patients recruited from one compared to the other. MRI scoring systems have been
shown to work better in patients with more active disease (16). Such differences
underline the advantages of testing scores across a range of disease severities as
part of a multicenter evaluation, providing more robust and broadly applicable
results. To our knowledge, our study is the first to compare four available MRI
scoring systems and to validate the MaRIA and Clermont score using the eAIS as
the reference standard.
Proefschrift2018-new2.indb 67 13/9/18 10:06
67
Comparison of MRI scoring systems
CI: 0.33–0.70), respectively) and ulcers showed only fair agreement (kappa: 0.26
(95% CI: 0.06–0.46)). ADC reference measurements in the erector spinae muscle
for observers 1 and 2 showed mean values of 1618 (SD: 133, range: 1194–1893) and
1618 (SD: 209, range: 1216–2491), respectively. Most MRI features showed moderate-
to-strong correlation to CDEIS (r=0.44–0.63), while perimural T2 signal, ulcers and
RCE showed weak-to-moderate correlation to CDEIS (r=0.27–0.49). Most features
showed weak-to-moderate correlation to eAIS for both observers (r=0.31–0.51),
while ulcers and RCE showed very weak-to-weak correlation to eAIS (r=0.10–0.23).
DISCUSSION
In this study we evaluated and compared four validated MRI scoring systems and
their incorporated MRI features for grading of inflammation in Crohn’s disease. All
four MRI scoring systems were comparable in terms of interobserver agreement
(ICC: 0.70–0.78), correlation to endoscopic and histopathological reference
standards (r=0.57–0.67 and r=0.38–0.49, respectively) and diagnostic accuracy for
active and ulcerating endoscopic disease (73–78% and 71–76%, respectively) and
active histopathologic disease (65–72%). The MRI evaluation of ulcers accounted
for almost all cases of discordant findings between the scores incorporating this
feature (MaRIA and Clermont score) and the scores not incorporating ulcers (CDMI
and London score). Enhancement and RCE showed good interobserver agreement,
while other MRI features showed fair-to-moderate agreement. The individual
features showed at best a moderate correlation to the eAIS, but moderate-to-strong
correlation to the CDEIS. There were differences in the strength of correlation to
CDEIS/eAIS between the two recruitment sites, but consistently good interobserver
agreement. It is likely that this observation stems from differences in disease
spectrum between the two sites; we found significantly higher disease activity in
patients recruited from one compared to the other. MRI scoring systems have been
shown to work better in patients with more active disease (16). Such differences
underline the advantages of testing scores across a range of disease severities as
part of a multicenter evaluation, providing more robust and broadly applicable
results. To our knowledge, our study is the first to compare four available MRI
scoring systems and to validate the MaRIA and Clermont score using the eAIS as
the reference standard.
Proefschrift2018-new2.indb 67 13/9/18 10:06
68
Chapter 3
Several studies have compared two or more of the reported scoring systems
(10,17–19). A recent study compared the MaRIA, Clermont and London scores,
reporting equal diagnostic accuracy between all three scores, albeit concluding
the MaRIA had superior operational characteristics, due to its high accuracy for
detecting of ulcerating disease (10). Importantly, that study showed a substantially
more active disease spectrum (median SES-CD of 6 (IQR: 2–10), maximum score
12 per segment), compared to ours (median CDEIS of 8 (IQR: 0–16.3), maximum
score 44 per segment). MRI’s limitations in the evaluation of mild disease have
been clearly shown in a previous meta-analysis (20), emphasizing the importance
of including these patients in new studies. Three studies have compared only the
MaRIA and Clermont scores in differing manners (17–19). Hordonneau et al. found a
high correlation between both scores in ileal Crohn’s disease (r=0.99), but less so
in colonic Crohn’s disease (r<0.80) (17). Kopylov et al. found that the MaRIA and
Clermont score correlated equally to the video-capsule endoscopy based Lewis
index (r=0.50 and r=0.53, respectively) (18). Lastly, a study by Buisson et al. found
that high baseline scores for either the MaRIA or Clermont score predicted clinical
remission after 12 weeks of anti-TNF therapy (19).
Compared to our study, higher interobserver agreement was seen in development
studies for the MaRIA, Clermont and London scores (ICC: 0.90–0.99 vs 0.70–0.72
in our study) (3,17,21). Arguably, our study provides a more realistic performance
estimate of performance in daily radiologic practice, due to the use of an independent
cohort, and a multi-center and multi-reader design. However, our reader expertise
most likely is higher than can be expected in general practice. In one study, the
use of exemplary reference (or key) images could have positively influenced their
results (3). Similar correlations to CDEIS were found in previous studies (6,22), with
exception of the original validation cohort for the MaRIA score (r=0.80 vs 0.60–0.64
in our study) (7). Correlations to eAIS have only been reported in the development
study for the CDMI and London scores, which were similar to those found in
our study (0.40 versus 0.49/0.39 in our study) (3). Although our aim in using a
second reference standard was to test the influence of the reference standard on
our outcomes, interestingly all four MRI scoring systems were better correlated to
CDEIS than to the histopathological eAIS. This higher correlation could be related
due to the similarity in macroscopic assessment by ileocolonoscopy and MRI,
compared to the focal microscopic evaluation by histopathology, which may be
confounded by the patchy distribution of inflammation inherent to Crohn’s disease.
Proefschrift2018-new2.indb 68 13/9/18 10:06
68
Chapter 3
Several studies have compared two or more of the reported scoring systems
(10,17–19). A recent study compared the MaRIA, Clermont and London scores,
reporting equal diagnostic accuracy between all three scores, albeit concluding
the MaRIA had superior operational characteristics, due to its high accuracy for
detecting of ulcerating disease (10). Importantly, that study showed a substantially
more active disease spectrum (median SES-CD of 6 (IQR: 2–10), maximum score
12 per segment), compared to ours (median CDEIS of 8 (IQR: 0–16.3), maximum
score 44 per segment). MRI’s limitations in the evaluation of mild disease have
been clearly shown in a previous meta-analysis (20), emphasizing the importance
of including these patients in new studies. Three studies have compared only the
MaRIA and Clermont scores in differing manners (17–19). Hordonneau et al. found a
high correlation between both scores in ileal Crohn’s disease (r=0.99), but less so
in colonic Crohn’s disease (r<0.80) (17). Kopylov et al. found that the MaRIA and
Clermont score correlated equally to the video-capsule endoscopy based Lewis
index (r=0.50 and r=0.53, respectively) (18). Lastly, a study by Buisson et al. found
that high baseline scores for either the MaRIA or Clermont score predicted clinical
remission after 12 weeks of anti-TNF therapy (19).
Compared to our study, higher interobserver agreement was seen in development
studies for the MaRIA, Clermont and London scores (ICC: 0.90–0.99 vs 0.70–0.72
in our study) (3,17,21). Arguably, our study provides a more realistic performance
estimate of performance in daily radiologic practice, due to the use of an independent
cohort, and a multi-center and multi-reader design. However, our reader expertise
most likely is higher than can be expected in general practice. In one study, the
use of exemplary reference (or key) images could have positively influenced their
results (3). Similar correlations to CDEIS were found in previous studies (6,22), with
exception of the original validation cohort for the MaRIA score (r=0.80 vs 0.60–0.64
in our study) (7). Correlations to eAIS have only been reported in the development
study for the CDMI and London scores, which were similar to those found in
our study (0.40 versus 0.49/0.39 in our study) (3). Although our aim in using a
second reference standard was to test the influence of the reference standard on
our outcomes, interestingly all four MRI scoring systems were better correlated to
CDEIS than to the histopathological eAIS. This higher correlation could be related
due to the similarity in macroscopic assessment by ileocolonoscopy and MRI,
compared to the focal microscopic evaluation by histopathology, which may be
confounded by the patchy distribution of inflammation inherent to Crohn’s disease.
Proefschrift2018-new2.indb 68 13/9/18 10:06
69
Comparison of MRI scoring systems
Moreover, the low correlation to the eAIS reveals the difficulties of characterizing
active inflammation on MRI, as has been shown in several studies (23–25). A
critical concept for improving therapeutic guidance, MRI’s ability to characterize
inflammation and fibrosis is moderate at best, and current MRI scores are likely
unsuited for this specific purpose.
In our study, we found high interobserver agreement for RCE, but a very weak
correlation with eAIS, despite the application of a corrected scaling. Scardapane et
al. showed that a modified MaRIA score (without the RCE) was comparable to MaRIA
in the evaluation of disease activity in Crohn’s disease (26). A different modified
MaRIA score was developed by Kim et al., which substituted the ulcer feature for
a subjective DWI grading (27). However, for optimization of both modified scores,
the weightings of incorporated features should be revised by regression analysis.
We agree that replacement of the ulcer feature by a more reliable component is
preferable. Ulcer detection, although valuable in individual cases, is likely too low
and variable to be of value in universal MRI scoring systems.
Current limitations of quantitative ADC measurements, such as scanner and
interscanner variability, different technical factors and methods of measurement,
were addressed by a number of studies (28–30). The variability produced by these
limitations was confirmed in our study by a wide range of ADC values in the erector
spinae muscle. The cut-off values determined for quantitative parameters are
subject to spectrum bias, which means the numerical ADC values are individual
and study-specific and therefore not generalizable (28). We should also note that
the necessary scaling corrections for RCE and ADC on Philips scanner systems
require technical expertise, which, to our knowledge, cannot be done automatically.
Furthermore, multiple ROI placement in the bowel wall is itself time consuming,
compared to scores based on objective observations.
Our study it limited due to its sole evaluation of the terminal ileum. Our goal was to
provide a comprehensive evaluation based on multiple reference standards including
biopsies, which were only taken from the terminal ileum. Also, colonic biopsies
would have provided less accurate location matching due to the absence of visual
landmarks on endoscopy. Our study used post-contrast sequences in the delayed
phase (7 minutes post-injection), instead of sequences in the portal-venous phase
(70 seconds post-injection). As the first peak of enhancement is followed by a long
Proefschrift2018-new2.indb 69 13/9/18 10:06
69
Comparison of MRI scoring systems
Moreover, the low correlation to the eAIS reveals the difficulties of characterizing
active inflammation on MRI, as has been shown in several studies (23–25). A
critical concept for improving therapeutic guidance, MRI’s ability to characterize
inflammation and fibrosis is moderate at best, and current MRI scores are likely
unsuited for this specific purpose.
In our study, we found high interobserver agreement for RCE, but a very weak
correlation with eAIS, despite the application of a corrected scaling. Scardapane et
al. showed that a modified MaRIA score (without the RCE) was comparable to MaRIA
in the evaluation of disease activity in Crohn’s disease (26). A different modified
MaRIA score was developed by Kim et al., which substituted the ulcer feature for
a subjective DWI grading (27). However, for optimization of both modified scores,
the weightings of incorporated features should be revised by regression analysis.
We agree that replacement of the ulcer feature by a more reliable component is
preferable. Ulcer detection, although valuable in individual cases, is likely too low
and variable to be of value in universal MRI scoring systems.
Current limitations of quantitative ADC measurements, such as scanner and
interscanner variability, different technical factors and methods of measurement,
were addressed by a number of studies (28–30). The variability produced by these
limitations was confirmed in our study by a wide range of ADC values in the erector
spinae muscle. The cut-off values determined for quantitative parameters are
subject to spectrum bias, which means the numerical ADC values are individual
and study-specific and therefore not generalizable (28). We should also note that
the necessary scaling corrections for RCE and ADC on Philips scanner systems
require technical expertise, which, to our knowledge, cannot be done automatically.
Furthermore, multiple ROI placement in the bowel wall is itself time consuming,
compared to scores based on objective observations.
Our study it limited due to its sole evaluation of the terminal ileum. Our goal was to
provide a comprehensive evaluation based on multiple reference standards including
biopsies, which were only taken from the terminal ileum. Also, colonic biopsies
would have provided less accurate location matching due to the absence of visual
landmarks on endoscopy. Our study used post-contrast sequences in the delayed
phase (7 minutes post-injection), instead of sequences in the portal-venous phase
(70 seconds post-injection). As the first peak of enhancement is followed by a long
Proefschrift2018-new2.indb 69 13/9/18 10:06
70
Chapter 3
(albeit lowered) plateau phase, we do not expect this to substantially change RCE
performance (31,32). Although a recent study found an increase in enhancement
between 70 seconds and 7 minutes to be associated with fibrosis, this finding is not
corroborated by previous DCE studies, and is yet to be validated (23,24). Evaluation
of DWI and subsequent placement of ADC measurement ROI’s might have been
positively influenced by the concurrent evaluation of contrast-enhanced sequences
in our study. For all scores, except the CDMI, externally developed cut-off values
were available for detection of ulcerating disease. Although we have devised a new
cut-off value for CDMI, this value should be externally validated on an independent
cohort. Lastly, we did not quantify the time needed to derive each of the scores, so
were unable to compare between them.
MRI scoring systems have been increasingly used in research studies and are starting
to be used as outcome measures for clinical trials. In contrast, daily practice has
not yet seen major uptake of MRI scoring systems, despite the promising abilities
for quantification of inflammatory activity and treatment response monitoring.
This limited application can in part be explained by unfamiliarity with the scoring
systems or some of their features, especially ulcer detection, the supposedly added
time-effort and the diversity of available scoring systems without the knowledge of
their strengths and limitations.
For all four scoring systems the interobserver agreement, diagnostic accuracy and
the correlation to the histopathological and endoscopic reference standard were
comparable. The London score, MaRIA and Clermont score have the additional
benefit of a validated cut-off value for ulcerating disease, with a cut off now also
proposed for CDMI.
ACKNOWLEDGMENTS
The authors thank Ernst Harting for the management of the VIGOR++ project and
Costis Kompis for project exploitation, Rado Andriantsimiavona and Laurence
Bourne from Biotronics3D for technical support, Christopher Pawley and Asif Jaffar
for patient recruitment, and Isha Verkaik for database management.
Proefschrift2018-new2.indb 70 13/9/18 10:06
70
Chapter 3
(albeit lowered) plateau phase, we do not expect this to substantially change RCE
performance (31,32). Although a recent study found an increase in enhancement
between 70 seconds and 7 minutes to be associated with fibrosis, this finding is not
corroborated by previous DCE studies, and is yet to be validated (23,24). Evaluation
of DWI and subsequent placement of ADC measurement ROI’s might have been
positively influenced by the concurrent evaluation of contrast-enhanced sequences
in our study. For all scores, except the CDMI, externally developed cut-off values
were available for detection of ulcerating disease. Although we have devised a new
cut-off value for CDMI, this value should be externally validated on an independent
cohort. Lastly, we did not quantify the time needed to derive each of the scores, so
were unable to compare between them.
MRI scoring systems have been increasingly used in research studies and are starting
to be used as outcome measures for clinical trials. In contrast, daily practice has
not yet seen major uptake of MRI scoring systems, despite the promising abilities
for quantification of inflammatory activity and treatment response monitoring.
This limited application can in part be explained by unfamiliarity with the scoring
systems or some of their features, especially ulcer detection, the supposedly added
time-effort and the diversity of available scoring systems without the knowledge of
their strengths and limitations.
For all four scoring systems the interobserver agreement, diagnostic accuracy and
the correlation to the histopathological and endoscopic reference standard were
comparable. The London score, MaRIA and Clermont score have the additional
benefit of a validated cut-off value for ulcerating disease, with a cut off now also
proposed for CDMI.
ACKNOWLEDGMENTS
The authors thank Ernst Harting for the management of the VIGOR++ project and
Costis Kompis for project exploitation, Rado Andriantsimiavona and Laurence
Bourne from Biotronics3D for technical support, Christopher Pawley and Asif Jaffar
for patient recruitment, and Isha Verkaik for database management.
Proefschrift2018-new2.indb 70 13/9/18 10:06
71
Comparison of MRI scoring systems
REFERENCES1. Gourtsoyiannis NC, Grammatikakis J, Papamastorakis G, Koutroumbakis J, Prassopoulos
P, Rousomoustakaki M, et al. Imaging of small intestinal Crohn’s disease: Comparison
between MR enteroclysis and conventional enteroclysis. Eur Radiol. 2006;16(9):1915–25.
2. Panes J, Bouhnik Y, Reinisch W, Stoker J, Taylor SA, Baumgart DC, et al. Imaging techniques
for assessment of inflammatory bowel disease: Joint ECCO and ESGAR evidence-based
consensus guidelines. J Crohn’s Colitis. 2013;7(7):556–85.
3. Steward MJ, Punwani S, Proctor I, Adjei-Gyamfi Y, Chatterjee F, Bloom S, et al. Non-
perforating small bowel Crohn’s disease assessed by MRI enterography: Derivation and
histopathological validation of an MR-based activity index. Eur J Radiol. 2012;81(9):2080–
8.
4. Ziech MLW, Bossuyt PMM, Laghi A, Lauenstein TC, Taylor SA, Stoker J. Grading luminal
Crohn’s disease: Which MRI features are considered as important? Eur J Radiol. 2012;81(4).
5. Rimola J, Rodriguez S, Garcia-Bosch O, Ordas I, Ayala E, Aceituno M, et al. Magnetic
resonance for assessment of disease activity and severity in ileocolonic Crohn’s disease.
Gut. 2009;58:1113–20.
6. Tielbeek JAW, Makanyanga JC, Bipat S, Pendse DA, Nio CY, Vos FM, et al. Grading crohn
disease activity with MRI: Interobserver variability of MRI features, MRI scoring of severity,
and correlation with crohn disease endoscopic index of severity. Am J Roentgenol.
2013;201(6):1220–8.
7. Rimola J, Ordás I, Rodriguez S, García-Bosch O, Aceituno M, Llach J, et al. Magnetic
resonance imaging for evaluation of Crohn’s disease: Validation of parameters of severity
and quantitative index of activity. Inflamm Bowel Dis. 2011;17(8):1759–68.
8. Buisson A, Joubert A, Montoriol PF, Ines DD, Hordonneau C, Pereira B, et al. Diffusion-
weighted magnetic resonance imaging for detecting and assessing ileal inflammation in
Crohn’s disease. Aliment Pharmacol Ther. 2013;37(5):537–45.
9. Hordonneau C, Buisson A, Scanzi J, Goutorbe F, Pereira B, Borderon C, et al. Diffusion-
Weighted Magnetic Resonance Imaging in Ileocolonic Crohn’s Disease: Validation of
Quantitative Index of Activity. Am J Gastroenterol. 2013;109(1):89–98.
10. Rimola J, Alvarez-Cofiño A, Pérez-Jeldres T, Ayuso C, Alfaro I, Rodríguez S, et al.
Comparison of three magnetic resonance enterography indices for grading activity in
Crohn’s disease. J Gastroenterol. 2017;52(5):585–93.
11. Puylaert CAJ, Schüffler PJ, Naziroglu RE, Tielbeek JAW, Li Z, Makanyanga JC, et al.
Semiautomatic Assessment of the Terminal Ileum and Colon in Patients with Crohn
Disease Using MRI (the VIGOR++ Project). Acad Radiol. 2018.
12. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Thérapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut. 1989;30(7):983–9.
13. Chenevert TL, Malyarenko DI, Newitt D, Li X, Jayatilake M, Tudorica A, et al. Errors in
Quantitative Image Analysis due to Platform-Dependent Image Scaling. Transl Oncol.
2014;7(1):65–71.
14. Landis JR, Koch GG. The measurement of observer agreement for categorical data.
Biometrics. 1977 Mar;33(1):159–74.
15. Daperno M, Castiglione F, de Ridder L, Dotan I, Farkkila M, Florholmen J, et al. Results of
the 2nd part Scientific Workshop of the ECCO (II): Measures and markers of prediction
to achieve, detect, and monitor intestinal healing in Inflammatory Bowel Disease. Vol. 5,
Proefschrift2018-new2.indb 71 13/9/18 10:06
71
Comparison of MRI scoring systems
REFERENCES1. Gourtsoyiannis NC, Grammatikakis J, Papamastorakis G, Koutroumbakis J, Prassopoulos
P, Rousomoustakaki M, et al. Imaging of small intestinal Crohn’s disease: Comparison
between MR enteroclysis and conventional enteroclysis. Eur Radiol. 2006;16(9):1915–25.
2. Panes J, Bouhnik Y, Reinisch W, Stoker J, Taylor SA, Baumgart DC, et al. Imaging techniques
for assessment of inflammatory bowel disease: Joint ECCO and ESGAR evidence-based
consensus guidelines. J Crohn’s Colitis. 2013;7(7):556–85.
3. Steward MJ, Punwani S, Proctor I, Adjei-Gyamfi Y, Chatterjee F, Bloom S, et al. Non-
perforating small bowel Crohn’s disease assessed by MRI enterography: Derivation and
histopathological validation of an MR-based activity index. Eur J Radiol. 2012;81(9):2080–
8.
4. Ziech MLW, Bossuyt PMM, Laghi A, Lauenstein TC, Taylor SA, Stoker J. Grading luminal
Crohn’s disease: Which MRI features are considered as important? Eur J Radiol. 2012;81(4).
5. Rimola J, Rodriguez S, Garcia-Bosch O, Ordas I, Ayala E, Aceituno M, et al. Magnetic
resonance for assessment of disease activity and severity in ileocolonic Crohn’s disease.
Gut. 2009;58:1113–20.
6. Tielbeek JAW, Makanyanga JC, Bipat S, Pendse DA, Nio CY, Vos FM, et al. Grading crohn
disease activity with MRI: Interobserver variability of MRI features, MRI scoring of severity,
and correlation with crohn disease endoscopic index of severity. Am J Roentgenol.
2013;201(6):1220–8.
7. Rimola J, Ordás I, Rodriguez S, García-Bosch O, Aceituno M, Llach J, et al. Magnetic
resonance imaging for evaluation of Crohn’s disease: Validation of parameters of severity
and quantitative index of activity. Inflamm Bowel Dis. 2011;17(8):1759–68.
8. Buisson A, Joubert A, Montoriol PF, Ines DD, Hordonneau C, Pereira B, et al. Diffusion-
weighted magnetic resonance imaging for detecting and assessing ileal inflammation in
Crohn’s disease. Aliment Pharmacol Ther. 2013;37(5):537–45.
9. Hordonneau C, Buisson A, Scanzi J, Goutorbe F, Pereira B, Borderon C, et al. Diffusion-
Weighted Magnetic Resonance Imaging in Ileocolonic Crohn’s Disease: Validation of
Quantitative Index of Activity. Am J Gastroenterol. 2013;109(1):89–98.
10. Rimola J, Alvarez-Cofiño A, Pérez-Jeldres T, Ayuso C, Alfaro I, Rodríguez S, et al.
Comparison of three magnetic resonance enterography indices for grading activity in
Crohn’s disease. J Gastroenterol. 2017;52(5):585–93.
11. Puylaert CAJ, Schüffler PJ, Naziroglu RE, Tielbeek JAW, Li Z, Makanyanga JC, et al.
Semiautomatic Assessment of the Terminal Ileum and Colon in Patients with Crohn
Disease Using MRI (the VIGOR++ Project). Acad Radiol. 2018.
12. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Thérapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut. 1989;30(7):983–9.
13. Chenevert TL, Malyarenko DI, Newitt D, Li X, Jayatilake M, Tudorica A, et al. Errors in
Quantitative Image Analysis due to Platform-Dependent Image Scaling. Transl Oncol.
2014;7(1):65–71.
14. Landis JR, Koch GG. The measurement of observer agreement for categorical data.
Biometrics. 1977 Mar;33(1):159–74.
15. Daperno M, Castiglione F, de Ridder L, Dotan I, Farkkila M, Florholmen J, et al. Results of
the 2nd part Scientific Workshop of the ECCO (II): Measures and markers of prediction
to achieve, detect, and monitor intestinal healing in Inflammatory Bowel Disease. Vol. 5,
Proefschrift2018-new2.indb 71 13/9/18 10:06
72
Chapter 3
Journal of Crohn’s and Colitis. 2011. p. 484–98.
16. Klang E, Amitai MM, Lahat A, Yablecovitch D, Avidan B, Neuman S, et al. Capsule endoscopy
validation of the magnetic enterography global score in patients with established Crohn’s
disease. J Crohns Colitis. 2017.
17. Hordonneau C, Buisson A, Scanzi J, Goutorbe F, Pereira B, Borderon C, et al. Diffusion-
weighted magnetic resonance imaging in ileocolonic Crohn’s disease: validation of
quantitative index of activity. Am J Gastroenterol. 2014 Jan;109(1):89–98.
18. Kopylov U, Klang E, Yablecovitch D, Lahat A, Avidan B, Neuman S, et al. Magnetic resonance
enterography versus capsule endoscopy activity indices for quantification of small bowel
inflammation in Crohn’s disease. Therap Adv Gastroenterol. 2016;9(5):655–63.
19. Buisson A, Hordonneau C, Goutte M, Scanzi J, Goutorbe F, Klotz T, et al. Diffusion-
weighted magnetic resonance enterocolonography in predicting remission after anti-
TNF induction therapy in Crohn’s disease. Dig Liver Dis. 2016 Mar;48(3):260–6.
20. Horsthuis K, Bipat S, Stokkers PCF, Stoker J. Magnetic resonance imaging for evaluation of
disease activity in Crohn’s disease: a systematic review. Eur Radiol. 2009;19(6):1450–60.
21. Coimbra AJF, Rimola J, O’Byrne S, Lu TT, Bengtsson T, De Crespigny A, et al. Magnetic
resonance enterography is feasible and reliable in multicenter clinical trials in patients
with Crohn’s disease, and may help select subjects with active inflammation. Aliment
Pharmacol Ther. 2016;43(1):61–72.
22. Buisson A, Hordonneau C, Goutte M, Boyer L, Pereira B, Bommelaer G. Diffusion-weighted
magnetic resonance imaging is effective to detect ileocolonic ulcerations in Crohn’s
disease. Aliment Pharmacol Ther. 2015 Aug;42(4):452–60.
23. Rimola J, Planell N, Rodríguez S, Delgado S, Ordás I, Ramírez-Morros A, et al.
Characterization of Inflammation and Fibrosis in Crohn’s Disease Lesions by Magnetic
Resonance Imaging. Am J Gastroenterol. 2015 Oct;1–9.
24. 24. Tielbeek JAW, Ziech MLW, Li Z, Lavini C, Bipat S, Bemelman WA, et al. Evaluation of
conventional, dynamic contrast enhanced and diffusion weighted MRI for quantitative
Crohn’s disease assessment with histopathology of surgical specimens. Eur Radiol. 2014
Mar;24(3):619–29.
25. Zappa M, Stefanescu C, Cazals-Hatem D, Bretagnol F, Deschamps L, Attar A, et al. Which
magnetic resonance imaging findings accurately evaluate inflammation in small bowel
Crohn’s disease? A retrospective comparison with surgical pathologic analysis. Inflamm
Bowel Dis. 2011;17(4):984–93.
26. Scardapane A, Ambrosi A, Salinaro E, Mancini ME, Principi M, Di Leo A, et al. Assessment
of Disease Activity in Small Bowel Crohn’s Disease: Comparison between Endoscopy and
Magnetic Resonance Enterography Using MRIA and Modified MRIA Score. Gastroenterol
Res Pract. 2015;2015.
27. Kim JS, Jang HY, Park SH, Kim KJ, Han K, Yang SK, et al. MR enterography assessment
of bowel inflammation severity in Crohn disease using the MR index of activity score:
Modifying roles of DWI and effects of contrast phases. Am J Roentgenol. 2017;208(5):1022–
9.
28. Kim K-J, Lee Y, Park SH, Kang B-K, Seo N, Yang S-K, et al. Diffusion-weighted MR
enterography for evaluating Crohn’s disease: how does it add diagnostically to
conventional MR enterography? Inflamm Bowel Dis. 2015 Jan;21(1):101–9.
29. Park SH. DWI at MR Enterography for Evaluating Bowel Inflammation in Crohn Disease.
AJR Am J Roentgenol. 2016 Mar;1–9.
30. Pendsé DA, Makanyanga JC, Plumb AA, Bhatnagar G, Atkinson D, Rodriguez-Justo M, et
Proefschrift2018-new2.indb 72 13/9/18 10:06
72
Chapter 3
Journal of Crohn’s and Colitis. 2011. p. 484–98.
16. Klang E, Amitai MM, Lahat A, Yablecovitch D, Avidan B, Neuman S, et al. Capsule endoscopy
validation of the magnetic enterography global score in patients with established Crohn’s
disease. J Crohns Colitis. 2017.
17. Hordonneau C, Buisson A, Scanzi J, Goutorbe F, Pereira B, Borderon C, et al. Diffusion-
weighted magnetic resonance imaging in ileocolonic Crohn’s disease: validation of
quantitative index of activity. Am J Gastroenterol. 2014 Jan;109(1):89–98.
18. Kopylov U, Klang E, Yablecovitch D, Lahat A, Avidan B, Neuman S, et al. Magnetic resonance
enterography versus capsule endoscopy activity indices for quantification of small bowel
inflammation in Crohn’s disease. Therap Adv Gastroenterol. 2016;9(5):655–63.
19. Buisson A, Hordonneau C, Goutte M, Scanzi J, Goutorbe F, Klotz T, et al. Diffusion-
weighted magnetic resonance enterocolonography in predicting remission after anti-
TNF induction therapy in Crohn’s disease. Dig Liver Dis. 2016 Mar;48(3):260–6.
20. Horsthuis K, Bipat S, Stokkers PCF, Stoker J. Magnetic resonance imaging for evaluation of
disease activity in Crohn’s disease: a systematic review. Eur Radiol. 2009;19(6):1450–60.
21. Coimbra AJF, Rimola J, O’Byrne S, Lu TT, Bengtsson T, De Crespigny A, et al. Magnetic
resonance enterography is feasible and reliable in multicenter clinical trials in patients
with Crohn’s disease, and may help select subjects with active inflammation. Aliment
Pharmacol Ther. 2016;43(1):61–72.
22. Buisson A, Hordonneau C, Goutte M, Boyer L, Pereira B, Bommelaer G. Diffusion-weighted
magnetic resonance imaging is effective to detect ileocolonic ulcerations in Crohn’s
disease. Aliment Pharmacol Ther. 2015 Aug;42(4):452–60.
23. Rimola J, Planell N, Rodríguez S, Delgado S, Ordás I, Ramírez-Morros A, et al.
Characterization of Inflammation and Fibrosis in Crohn’s Disease Lesions by Magnetic
Resonance Imaging. Am J Gastroenterol. 2015 Oct;1–9.
24. 24. Tielbeek JAW, Ziech MLW, Li Z, Lavini C, Bipat S, Bemelman WA, et al. Evaluation of
conventional, dynamic contrast enhanced and diffusion weighted MRI for quantitative
Crohn’s disease assessment with histopathology of surgical specimens. Eur Radiol. 2014
Mar;24(3):619–29.
25. Zappa M, Stefanescu C, Cazals-Hatem D, Bretagnol F, Deschamps L, Attar A, et al. Which
magnetic resonance imaging findings accurately evaluate inflammation in small bowel
Crohn’s disease? A retrospective comparison with surgical pathologic analysis. Inflamm
Bowel Dis. 2011;17(4):984–93.
26. Scardapane A, Ambrosi A, Salinaro E, Mancini ME, Principi M, Di Leo A, et al. Assessment
of Disease Activity in Small Bowel Crohn’s Disease: Comparison between Endoscopy and
Magnetic Resonance Enterography Using MRIA and Modified MRIA Score. Gastroenterol
Res Pract. 2015;2015.
27. Kim JS, Jang HY, Park SH, Kim KJ, Han K, Yang SK, et al. MR enterography assessment
of bowel inflammation severity in Crohn disease using the MR index of activity score:
Modifying roles of DWI and effects of contrast phases. Am J Roentgenol. 2017;208(5):1022–
9.
28. Kim K-J, Lee Y, Park SH, Kang B-K, Seo N, Yang S-K, et al. Diffusion-weighted MR
enterography for evaluating Crohn’s disease: how does it add diagnostically to
conventional MR enterography? Inflamm Bowel Dis. 2015 Jan;21(1):101–9.
29. Park SH. DWI at MR Enterography for Evaluating Bowel Inflammation in Crohn Disease.
AJR Am J Roentgenol. 2016 Mar;1–9.
30. Pendsé DA, Makanyanga JC, Plumb AA, Bhatnagar G, Atkinson D, Rodriguez-Justo M, et
Proefschrift2018-new2.indb 72 13/9/18 10:06
73
Comparison of MRI scoring systems
al. Diffusion-weighted imaging for evaluating inflammatory activity in Crohn’s disease:
comparison with histopathology, conventional MRI activity scores, and faecal calprotectin.
Abdom Radiol. 2017;42(1):115–23.
31. Giusti S, Faggioni L, Neri E, Fruzzetti E, Nardini L, Marchi S, et al. Dynamic MRI of the small
bowel: Usefulness of quantitative contrast-enhancement parameters and time-signal
intensity curves for differentiating between active and inactive Crohn’s disease. Abdom
Imaging. 2010;35(6):646–53.
32. Tielbeek JAW, Ziech MLW, Li Z, Lavini C, Bipat S, Bemelman WA, et al. Evaluation of
conventional, dynamic contrast enhanced and diffusion weighted MRI for quantitative
Crohn’s disease assessment with histopathology of surgical specimens. Eur Radiol.
2014;24(3):619–29.
Proefschrift2018-new2.indb 73 13/9/18 10:06
73
Comparison of MRI scoring systems
al. Diffusion-weighted imaging for evaluating inflammatory activity in Crohn’s disease:
comparison with histopathology, conventional MRI activity scores, and faecal calprotectin.
Abdom Radiol. 2017;42(1):115–23.
31. Giusti S, Faggioni L, Neri E, Fruzzetti E, Nardini L, Marchi S, et al. Dynamic MRI of the small
bowel: Usefulness of quantitative contrast-enhancement parameters and time-signal
intensity curves for differentiating between active and inactive Crohn’s disease. Abdom
Imaging. 2010;35(6):646–53.
32. Tielbeek JAW, Ziech MLW, Li Z, Lavini C, Bipat S, Bemelman WA, et al. Evaluation of
conventional, dynamic contrast enhanced and diffusion weighted MRI for quantitative
Crohn’s disease assessment with histopathology of surgical specimens. Eur Radiol.
2014;24(3):619–29.
Proefschrift2018-new2.indb 73 13/9/18 10:06
74
Chapter 3
Sup
ple
men
tary
dat
a 1.
Imag
ing
seq
uen
ces
an
d p
ara
mete
rs.
AM
CB
alan
ced
GE
BT
FE
dy-
nam
icT
2-w
SSF
SET
2-w
SSF
SET
2-w
SSF
SED
WI s
e-q
uenc
e3D
T1-
w S
PG
ED
CE
seq
uen-
cea
3D T
1-w
SP
GE
3D T
1-w
SP
GE
Pla
neC
oro
nal
Co
rona
lC
oro
nal
Axi
alA
xial
Axi
alC
oro
nal
Co
rona
lC
oro
nal
Axi
al
Fat
satu
rati
on
No
Yes
No
No
Yes
Yes
Yes
No
Yes
Yes
TR
/TE
(m
s)2.5
/1.2
52.1/1
.03
628
/60
75
9/1
199
67/
50
29
72/6
02.4
/1.0
62.9
/1.8
2.4
/1.0
62.3
/1.1
Flip
an
gle
(°)
60
45
90
90
90
90
1015
1010
Slic
e t
hic
k-
ness
/gap
(m
m)
5/1
10/0
4/1
4/1
7/1
7/1
2/0
2.5
/02/0
2/0
Slic
es
15V
ari
ab
le4
06
03
x15
20
90
30
90
90
Fie
ld o
f vie
w
(mm
)3
80
x3
80
38
0x3
80
38
0x3
80
40
0x4
00
38
0x3
02
38
0x26
03
80
x3
80
38
0x3
80
38
0x3
80
38
0x3
80
Acq
uis
itio
n
matr
ix25
2x20
919
2x15
824
0x211
25
6x19
829
2x18
915
2x10
119
2x18
812
8x12
8x15
192x18
820
4x20
5x4
5
Reco
nst
ruc-
tio
n m
atr
ix4
00
x4
00
192x19
23
84
x3
84
528
x5
28
320
x25
628
8x19
619
2x19
2224
x224
x3
019
2x19
23
84
x3
84
x9
0
UC
LHB
alan
ced
GE
BT
FE
dy-
nam
icT
2-w
SSF
SET
2-w
SSF
SET
2-w
SSF
SED
WI s
e-q
uenc
e3D
T1-
w S
PG
ED
CE
seq
uen-
cea
3D T
1-w
SP
GE
3D T
1-w
SP
GE
Pla
neC
oro
nal
Co
rona
lC
oro
nal
Axi
alA
xial
Axi
alC
oro
nal
Co
rona
lC
oro
nal
Axi
al
Fat
satu
rati
on
No
Yes
No
No
Yes
Yes
Yes
No
Yes
Yes
TR
/TE
(m
s)2.5
/1.2
51.
97/
0.9
86
61/
60
75
9/1
1913
14/5
04
975/4
82.2
/1.0
22.9
/1.8
2.2
/1.0
22.1/1
.03
Flip
an
gle
(°)
60
45
90
90
90
90
1015
1010
Slic
e t
hic
k-
ness
/gap
(m
m)
5/1
10/0
4/1
4/1
7/1
7/1
2/0
2/0
2/0
Slic
es
15V
ari
ab
le4
06
03
x15
20
90
30
90
90
Fie
ld o
f vie
w
(mm
)3
80
x3
80
38
0x3
80
38
0x3
80
40
0x4
00
38
0x3
02
38
0x26
03
80
x4
59
38
0x4
39
38
0x4
59
38
0x3
80
Acq
uis
itio
n
matr
ix25
2x211
192 x
15
823
6x210
25
2x19
829
2x18
915
2x10
119
2x23
012
8x14
5x15
192x23
019
6x19
7x4
5
Reco
nst
ruc-
tio
n m
atr
ix4
00
x4
00
192x19
24
00
x4
00
528
x5
28
320
x3
20
28
8x19
620
0x24
019
2x224
x3
020
0x24
03
84
x3
84
x9
0
Proefschrift2018-new2.indb 74 13/9/18 10:06
74
Chapter 3
Sup
ple
men
tary
dat
a 1.
Imag
ing
seq
uen
ces
an
d p
ara
mete
rs.
AM
CB
alan
ced
GE
BT
FE
dy-
nam
icT
2-w
SSF
SET
2-w
SSF
SET
2-w
SSF
SED
WI s
e-q
uenc
e3D
T1-
w S
PG
ED
CE
seq
uen-
cea
3D T
1-w
SP
GE
3D T
1-w
SP
GE
Pla
neC
oro
nal
Co
rona
lC
oro
nal
Axi
alA
xial
Axi
alC
oro
nal
Co
rona
lC
oro
nal
Axi
al
Fat
satu
rati
on
No
Yes
No
No
Yes
Yes
Yes
No
Yes
Yes
TR
/TE
(m
s)2.5
/1.2
52.1/1
.03
628
/60
75
9/1
199
67/
50
29
72/6
02.4
/1.0
62.9
/1.8
2.4
/1.0
62.3
/1.1
Flip
an
gle
(°)
60
45
90
90
90
90
1015
1010
Slic
e t
hic
k-
ness
/gap
(m
m)
5/1
10/0
4/1
4/1
7/1
7/1
2/0
2.5
/02/0
2/0
Slic
es
15V
ari
ab
le4
06
03
x15
20
90
30
90
90
Fie
ld o
f vie
w
(mm
)3
80
x3
80
38
0x3
80
38
0x3
80
40
0x4
00
38
0x3
02
38
0x26
03
80
x3
80
38
0x3
80
38
0x3
80
38
0x3
80
Acq
uis
itio
n
matr
ix25
2x20
919
2x15
824
0x211
25
6x19
829
2x18
915
2x10
119
2x18
812
8x12
8x15
192x18
820
4x20
5x4
5
Reco
nst
ruc-
tio
n m
atr
ix4
00
x4
00
192x19
23
84
x3
84
528
x5
28
320
x25
628
8x19
619
2x19
2224
x224
x3
019
2x19
23
84
x3
84
x9
0
UC
LHB
alan
ced
GE
BT
FE
dy-
nam
icT
2-w
SSF
SET
2-w
SSF
SET
2-w
SSF
SED
WI s
e-q
uenc
e3D
T1-
w S
PG
ED
CE
seq
uen-
cea
3D T
1-w
SP
GE
3D T
1-w
SP
GE
Pla
neC
oro
nal
Co
rona
lC
oro
nal
Axi
alA
xial
Axi
alC
oro
nal
Co
rona
lC
oro
nal
Axi
al
Fat
satu
rati
on
No
Yes
No
No
Yes
Yes
Yes
No
Yes
Yes
TR
/TE
(m
s)2.5
/1.2
51.
97/
0.9
86
61/
60
75
9/1
1913
14/5
04
975/4
82.2
/1.0
22.9
/1.8
2.2
/1.0
22.1/1
.03
Flip
an
gle
(°)
60
45
90
90
90
90
1015
1010
Slic
e t
hic
k-
ness
/gap
(m
m)
5/1
10/0
4/1
4/1
7/1
7/1
2/0
2/0
2/0
Slic
es
15V
ari
ab
le4
06
03
x15
20
90
30
90
90
Fie
ld o
f vie
w
(mm
)3
80
x3
80
38
0x3
80
38
0x3
80
40
0x4
00
38
0x3
02
38
0x26
03
80
x4
59
38
0x4
39
38
0x4
59
38
0x3
80
Acq
uis
itio
n
matr
ix25
2x211
192 x
15
823
6x210
25
2x19
829
2x18
915
2x10
119
2x23
012
8x14
5x15
192x23
019
6x19
7x4
5
Reco
nst
ruc-
tio
n m
atr
ix4
00
x4
00
192x19
24
00
x4
00
528
x5
28
320
x3
20
28
8x19
620
0x24
019
2x224
x3
020
0x24
03
84
x3
84
x9
0
Proefschrift2018-new2.indb 74 13/9/18 10:06
75
Comparison of MRI scoring systems
Supplementary data 2. Calculation of different MRI activity scoring systems.
2.1. CDMI score
Score 0 1 2 3
Mural thickness 1-3 mm >3-5 mm >5-7 mm >7 mm
Mural T2 signal Equivalent to normal bowel wall
Minor increase in signal - bowel wall appears dark grey on fat saturated images
Moderate in-crease in signal - bowel wall ap-pears light grey on fat saturated images
Marked incease in signal - bowel wall contains areas of white high signal approaching that of luminal content
Perimural T2 signal
Equivalent to normal mes-entery
Increase in mes-enteric signal but no fluid
Small fluid rim (≤2 mm)
Larger fluid rim (>2 mm)
Enhancement Equivalent to normal bowel wall
Minor enhance-ment - bowel wall signal greater than normal small bowel but significantly less than nearby va-sular structures
Moderate enhancement - bowel wall signal increased but somewhat less than nearby vascular struc-tures
Marked en-hancement - bowel wall signal ap-proaches that of nearby vascular structures
2.2. London score (features are scored using CDMI table)
1.79 + 1.34*mural thickness + 0.94*mural T2 score
2.3. MaRIA score
1.5*mural thickness (mm) + 0.02*RCE + 5*edema + 10*ulceration
The following formula was used to calculate RCE values:
RCE = (SI post-contrast - SI pre-contrast) / SI pre-contrast
2.4. Clermont score
1.646*mural thickness (mm) – 1.321*ADC + 5.613*edema + 8.306*ulceration
Proefschrift2018-new2.indb 75 13/9/18 10:06
75
Comparison of MRI scoring systems
Supplementary data 2. Calculation of different MRI activity scoring systems.
2.1. CDMI score
Score 0 1 2 3
Mural thickness 1-3 mm >3-5 mm >5-7 mm >7 mm
Mural T2 signal Equivalent to normal bowel wall
Minor increase in signal - bowel wall appears dark grey on fat saturated images
Moderate in-crease in signal - bowel wall ap-pears light grey on fat saturated images
Marked incease in signal - bowel wall contains areas of white high signal approaching that of luminal content
Perimural T2 signal
Equivalent to normal mes-entery
Increase in mes-enteric signal but no fluid
Small fluid rim (≤2 mm)
Larger fluid rim (>2 mm)
Enhancement Equivalent to normal bowel wall
Minor enhance-ment - bowel wall signal greater than normal small bowel but significantly less than nearby va-sular structures
Moderate enhancement - bowel wall signal increased but somewhat less than nearby vascular struc-tures
Marked en-hancement - bowel wall signal ap-proaches that of nearby vascular structures
2.2. London score (features are scored using CDMI table)
1.79 + 1.34*mural thickness + 0.94*mural T2 score
2.3. MaRIA score
1.5*mural thickness (mm) + 0.02*RCE + 5*edema + 10*ulceration
The following formula was used to calculate RCE values:
RCE = (SI post-contrast - SI pre-contrast) / SI pre-contrast
2.4. Clermont score
1.646*mural thickness (mm) – 1.321*ADC + 5.613*edema + 8.306*ulceration
Proefschrift2018-new2.indb 75 13/9/18 10:06
76
Chapter 3
Sup
ple
men
tary
dat
a 3.
Ind
ivid
ual an
aly
sis
for
each
of
the in
clu
din
g c
en
ters
(A
MC
an
d U
CL
H)
an
d in
div
idu
al o
bse
rvers
(A
,B,C
,D)
Rep
rod
ucib
ility
ICC
(9
5% C
I)
AM
CU
CLH
CD
MI sc
ore
0.8
0 (
0.6
9–0
.88
) 0
.74
(0
.49
–0.8
7)
Lo
nd
on
sco
re0
.72 (
0.5
5–0
.83
) 0
.73
(0
.51–
0.8
5)
MaR
IA s
co
re0
.69
(0
.53
–0.8
0)
0.7
2 (
0.5
2–0
.85
)
Cle
rmo
nt
sco
re0
.71
(0.5
5–0
.82)
0.7
1 (0
.49
–0.8
4)
Co
rrel
atio
n to
eA
ISA
MC
UC
LH
O
bse
rver
AO
bse
rver
BO
bse
rver
CO
bse
rver
D
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
CD
MI sc
ore
0.5
2<
0.0
01
0.5
0<
0.0
01
0.2
40
.15
0.2
20
.20
Lo
nd
on
sco
re0
.53
<0
.00
10
.50
<0
.00
10
.25
0.14
0.18
0.3
1
MaR
IA s
co
re0
.48
<0
.00
10
.45
<0
.00
10
.17
0.3
20
.27
0.11
Cle
rmo
nt
sco
re0
.51
<0
.00
10
.48
<0
.00
10
.16
0.3
90
.30
0.0
9
Co
rrel
atio
n to
CD
EIS
AM
CU
CLH
O
bse
rver
AO
bse
rver
BO
bse
rver
CO
bse
rver
D
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
CD
MI sc
ore
0.5
6<
0.0
01
0.7
0<
0.0
01
0.4
5<
0.0
10
.64
<0
.00
1
Lo
nd
on
sco
re0
.56
<0
.00
10
.70
<0
.00
10
.48
<0
.01
0.5
7<
0.0
01
MaR
IA s
co
re0
.65
<0
.00
10
.71
<0
.00
10
.36
0.0
40
.59
<0
.00
1
Cle
rmo
nt
sco
re0
.63
<0
.00
10
.73
<0
.00
10
.43
0.0
20
.58
0.0
01
CD
EIS
, C
roh
n’s
dis
ease
en
do
sco
pic
in
dex o
f se
veri
ty; C
I, c
on
fid
en
ce in
terv
al;
ICC
, in
tracla
ss c
orr
ela
tio
n c
oeff
icie
nt;
eA
IS, en
do
sco
pic
bio
psy
acu
te in
flam
mato
ry s
co
re; C
DM
I, C
roh
n’s
dis
ease
MR
I in
dex; M
aR
IA, m
ag
neti
c r
eso
nan
ce in
dex o
f acti
vit
y.
Proefschrift2018-new2.indb 76 13/9/18 10:06
76
Chapter 3
Sup
ple
men
tary
dat
a 3.
Ind
ivid
ual an
aly
sis
for
each
of
the in
clu
din
g c
en
ters
(A
MC
an
d U
CL
H)
an
d in
div
idu
al o
bse
rvers
(A
,B,C
,D)
Rep
rod
ucib
ility
ICC
(9
5% C
I)
AM
CU
CLH
CD
MI sc
ore
0.8
0 (
0.6
9–0
.88
) 0
.74
(0
.49
–0.8
7)
Lo
nd
on
sco
re0
.72 (
0.5
5–0
.83
) 0
.73
(0
.51–
0.8
5)
MaR
IA s
co
re0
.69
(0
.53
–0.8
0)
0.7
2 (
0.5
2–0
.85
)
Cle
rmo
nt
sco
re0
.71
(0.5
5–0
.82)
0.7
1 (0
.49
–0.8
4)
Co
rrel
atio
n to
eA
ISA
MC
UC
LH
O
bse
rver
AO
bse
rver
BO
bse
rver
CO
bse
rver
D
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
CD
MI sc
ore
0.5
2<
0.0
01
0.5
0<
0.0
01
0.2
40
.15
0.2
20
.20
Lo
nd
on
sco
re0
.53
<0
.00
10
.50
<0
.00
10
.25
0.14
0.18
0.3
1
MaR
IA s
co
re0
.48
<0
.00
10
.45
<0
.00
10
.17
0.3
20
.27
0.11
Cle
rmo
nt
sco
re0
.51
<0
.00
10
.48
<0
.00
10
.16
0.3
90
.30
0.0
9
Co
rrel
atio
n to
CD
EIS
AM
CU
CLH
O
bse
rver
AO
bse
rver
BO
bse
rver
CO
bse
rver
D
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
Co
effici
ent
p-V
alue
CD
MI sc
ore
0.5
6<
0.0
01
0.7
0<
0.0
01
0.4
5<
0.0
10
.64
<0
.00
1
Lo
nd
on
sco
re0
.56
<0
.00
10
.70
<0
.00
10
.48
<0
.01
0.5
7<
0.0
01
MaR
IA s
co
re0
.65
<0
.00
10
.71
<0
.00
10
.36
0.0
40
.59
<0
.00
1
Cle
rmo
nt
sco
re0
.63
<0
.00
10
.73
<0
.00
10
.43
0.0
20
.58
0.0
01
CD
EIS
, C
roh
n’s
dis
ease
en
do
sco
pic
in
dex o
f se
veri
ty; C
I, c
on
fid
en
ce in
terv
al;
ICC
, in
tracla
ss c
orr
ela
tio
n c
oeff
icie
nt;
eA
IS, en
do
sco
pic
bio
psy
acu
te in
flam
mato
ry s
co
re; C
DM
I, C
roh
n’s
dis
ease
MR
I in
dex; M
aR
IA, m
ag
neti
c r
eso
nan
ce in
dex o
f acti
vit
y.
Proefschrift2018-new2.indb 76 13/9/18 10:06
77
Comparison of MRI scoring systems
Supplementary data 4. Interobserver agreement between the two observer pairs for all MRI
features
MRI features Weighted kappa (95% CI)
Wall thickness (0-3) 0.52 (0.41–0.63)
Mural T2 signal 0.45 (0.33–0.57)
Perimural T2 signal 0.52 (0.33–0.70)
Enhancement 0.66 (0.55–0.77)
Multirater kappa
Mural edema 0.59 (0.43–0.75)
Ulcers 0.26 (0.06–0.46)
ICC (95% CI)
Wall thickness (in mm) 0.54 (0.22–0.72)
Apparent diffusion coefficient (ADC) 0.51 (0.32–0.65)
Relative contrast enhancement (RCE) 0.67 (0.55–0.77)
CI, confidence interval; ICC, intraclass correlation coefficient.
Supplementary data 5. Correlation between MRI features, eAIS and CDEIS
eAIS CDEIS
Observer 1 Observer 2 Observer 1 Observer 2
MRI features r p-Value r p-Value r p-Value r p-Value
Wall thickness (0-3) 0.51 <0.001 0.38 <0.001 0.63 <0.001 0.60 <0.001
Mural T2 signal 0.44 <0.001 0.37 <0.001 0.50 <0.001 0.62 <0.001
Perimural T2 signal 0.31 0.002 0.36 <0.001 0.26 <0.02 0.49 <0.001
Enhancement 0.43 <0.001 0.38 <0.001 0.57 <0.001 0.63 <0.001
Mural edema 0.35 <0.001 0.31 0.002 0.44 <0.001 0.56 <0.001
Ulcers 0.10 0.34 0.23 0.022 0.27 0.01 0.48 <0.001
Wall thickness (mm) 0.48 <0.001 0.41 <0.001 0.62 <0.001 0.59 <0.001
Apparent diffusion coefficient (ADC)
-0.41 <0.001 -0.26 0.013 -0.44 <0.001 -0.54 <0.001
Relative contrast enhancement (RCE)
0.11 0.30 0.16 0.11 0.39 <0.001 0.40 <0.001
CDEIS, Crohn’s disease endoscopic index of severity; eAIS, endoscopic biopsy acute in-flammatory score.
Proefschrift2018-new2.indb 77 13/9/18 10:06
77
Comparison of MRI scoring systems
Supplementary data 4. Interobserver agreement between the two observer pairs for all MRI
features
MRI features Weighted kappa (95% CI)
Wall thickness (0-3) 0.52 (0.41–0.63)
Mural T2 signal 0.45 (0.33–0.57)
Perimural T2 signal 0.52 (0.33–0.70)
Enhancement 0.66 (0.55–0.77)
Multirater kappa
Mural edema 0.59 (0.43–0.75)
Ulcers 0.26 (0.06–0.46)
ICC (95% CI)
Wall thickness (in mm) 0.54 (0.22–0.72)
Apparent diffusion coefficient (ADC) 0.51 (0.32–0.65)
Relative contrast enhancement (RCE) 0.67 (0.55–0.77)
CI, confidence interval; ICC, intraclass correlation coefficient.
Supplementary data 5. Correlation between MRI features, eAIS and CDEIS
eAIS CDEIS
Observer 1 Observer 2 Observer 1 Observer 2
MRI features r p-Value r p-Value r p-Value r p-Value
Wall thickness (0-3) 0.51 <0.001 0.38 <0.001 0.63 <0.001 0.60 <0.001
Mural T2 signal 0.44 <0.001 0.37 <0.001 0.50 <0.001 0.62 <0.001
Perimural T2 signal 0.31 0.002 0.36 <0.001 0.26 <0.02 0.49 <0.001
Enhancement 0.43 <0.001 0.38 <0.001 0.57 <0.001 0.63 <0.001
Mural edema 0.35 <0.001 0.31 0.002 0.44 <0.001 0.56 <0.001
Ulcers 0.10 0.34 0.23 0.022 0.27 0.01 0.48 <0.001
Wall thickness (mm) 0.48 <0.001 0.41 <0.001 0.62 <0.001 0.59 <0.001
Apparent diffusion coefficient (ADC)
-0.41 <0.001 -0.26 0.013 -0.44 <0.001 -0.54 <0.001
Relative contrast enhancement (RCE)
0.11 0.30 0.16 0.11 0.39 <0.001 0.40 <0.001
CDEIS, Crohn’s disease endoscopic index of severity; eAIS, endoscopic biopsy acute in-flammatory score.
Proefschrift2018-new2.indb 77 13/9/18 10:06
Proefschrift2018-new2.indb 78 13/9/18 10:06 Proefschrift2018-new2.indb 78 13/9/18 10:06
CHAPTER 4
Semiautomatic assessment of the terminal ileum and colon in patients with Crohn disease using MRI (the VIGOR++ project)
Carl. A.J. Puylaert*, Peter J. Schüffler*, Robiel E. Naziroglu, Jeroen A.W. Tielbeek,
Zhang Li, Jesica C. Makanyanga, Charlotte J. Tutein Nolthenius, C. Yung Nio, Doug
A. Pendse, Alex Menys, Cyriel Y. Ponsioen, David Atkinson, Alastair Forbes, Joachim
M. Buhmann, Thomas J. Fuchs, Haralambos Hatzakis, Lucas J. van Vliet, Jaap Stoker,
Stuart A. Taylor, Frans M. Vos
*Both authors contributed equally
Proefschrift2018-new2.indb 79 13/9/18 10:06
CHAPTER 4
Semiautomatic assessment of the terminal ileum and colon in patients with Crohn disease using MRI (the VIGOR++ project)
Carl. A.J. Puylaert*, Peter J. Schüffler*, Robiel E. Naziroglu, Jeroen A.W. Tielbeek,
Zhang Li, Jesica C. Makanyanga, Charlotte J. Tutein Nolthenius, C. Yung Nio, Doug
A. Pendse, Alex Menys, Cyriel Y. Ponsioen, David Atkinson, Alastair Forbes, Joachim
M. Buhmann, Thomas J. Fuchs, Haralambos Hatzakis, Lucas J. van Vliet, Jaap Stoker,
Stuart A. Taylor, Frans M. Vos
*Both authors contributed equally
Proefschrift2018-new2.indb 79 13/9/18 10:06
80
Chapter 4
ABSTRACT
Objectives
The objective of this study was to develop and validate a predictive magnetic
resonance imaging (MRI) activity score for ileocolonic Crohn disease activity based
on both subjective and semiautomatic MRI features.
Materials and Methods
An MRI activity score (the “virtual gastrointestinal tract [VIGOR]” score) was
developed from 27 validated magnetic resonance enterography datasets,
including subjective radiologist observation of mural T2 signal and semiautomatic
measurements of bowel wall thickness, excess volume, and dynamic contrast
enhancement (initial slope of increase). A second subjective score was developed
based on only radiologist observations. For validation, two observers applied both
scores and three existing scores to a prospective dataset of 106 patients (59 women,
median age 33) with known Crohn disease, using the endoscopic Crohn’s Disease
Endoscopic Index of Severity (CDEIS) as a reference standard.
Results
The VIGOR score (17.1 x initial slope of increase + 0.2 x excess volume + 2.3 x
mural T2) and other activity scores all had comparable correlation to the CDEIS
scores (observer 1: r = 0.58 and 0.59, and observer 2: r = 0.34–0.40 and 0.43–0.51,
respectively). The VIGOR score, however, improved interobserver agreement
compared to the other activity scores (intraclass correlation coefficient = 0.81 vs
0.44–0.59). A diagnostic accuracy of 80%–81% was seen for the VIGOR score,
similar to the other scores.
Conclusions
The VIGOR score achieves comparable accuracy to conventional MRI activity
scores, but with significantly improved reproducibility, favoring its use for disease
monitoring and therapy evaluation.
Proefschrift2018-new2.indb 80 13/9/18 10:06
80
Chapter 4
ABSTRACT
Objectives
The objective of this study was to develop and validate a predictive magnetic
resonance imaging (MRI) activity score for ileocolonic Crohn disease activity based
on both subjective and semiautomatic MRI features.
Materials and Methods
An MRI activity score (the “virtual gastrointestinal tract [VIGOR]” score) was
developed from 27 validated magnetic resonance enterography datasets,
including subjective radiologist observation of mural T2 signal and semiautomatic
measurements of bowel wall thickness, excess volume, and dynamic contrast
enhancement (initial slope of increase). A second subjective score was developed
based on only radiologist observations. For validation, two observers applied both
scores and three existing scores to a prospective dataset of 106 patients (59 women,
median age 33) with known Crohn disease, using the endoscopic Crohn’s Disease
Endoscopic Index of Severity (CDEIS) as a reference standard.
Results
The VIGOR score (17.1 x initial slope of increase + 0.2 x excess volume + 2.3 x
mural T2) and other activity scores all had comparable correlation to the CDEIS
scores (observer 1: r = 0.58 and 0.59, and observer 2: r = 0.34–0.40 and 0.43–0.51,
respectively). The VIGOR score, however, improved interobserver agreement
compared to the other activity scores (intraclass correlation coefficient = 0.81 vs
0.44–0.59). A diagnostic accuracy of 80%–81% was seen for the VIGOR score,
similar to the other scores.
Conclusions
The VIGOR score achieves comparable accuracy to conventional MRI activity
scores, but with significantly improved reproducibility, favoring its use for disease
monitoring and therapy evaluation.
Proefschrift2018-new2.indb 80 13/9/18 10:06
81
Crohn disease: Semiautomatic MRI assessment
INTRODUCTION
Crohn disease (CD) is an inflammatory bowel disease, which can present
throughout the gastrointestinal tract, particularly affecting the small bowel and
the colon. Magnetic resonance imaging (MRI) is increasingly used for diagnosis
and phenotyping of CD because it is safe, noninvasive, and has high accuracy for
evaluating enteric disease and extramural complications (1). MRI features such as
wall thickness and T1 and T2 bowel wall signals have been validated as biomarkers of
CD activity, demonstrating good correlation with endoscopic and histopathologic
grading of inflammation (2, 3, 4).
Recent years have seen several MRI disease activity scores being developed and
externally validated, combining multiple MRI features to predict overall disease
activity (3, 4, 5, 6). These scores are gradually disseminating into clinical practice,
although at present, they are predominantly employed as research tools. The
magnetic resonance index of activity (MaRIA), for example, has been developed
using the Crohn’s Disease Endoscopic Index of Severity (CDEIS) as a reference
standard. The MaRIA is based on quantitative measurement of bowel wall relative
contrast enhancement, along with subjective evaluation of mural ulceration and
abnormal T2 signal (3). Other indices, such as the London score and the Crohn
disease MRI index (CDMI), rely on qualitative grading of various features by
reporting radiologists (4, 6). Such activity scores can be applied to individual bowel
segments, as well as to the patient as a whole, as both are important to clinical
management. Before MRI scores can be widely adopted for evaluating disease
activity and therapeutic monitoring, high accuracy across the spectrum of disease
severity and good reproducibility among radiologists must be proven. The current
literature, however, reports variable reproducibility for many features used in MRI
activity scores (6, 7).
One potential solution to the current limitations of MRI activity scoring is to
incorporate novel software solutions, which can automatically extract relevant
features from MRI data. Such software could reduce both interobserver variability
and the risk of observer bias inherent to subjective evaluation (8). New MRI image
processing methods are available, which give semiautomatic measurements of
bowel wall thickness, providing superior reproducibility over manual measurement
(9). Further techniques have been developed that automatically extract perfusion
Proefschrift2018-new2.indb 81 13/9/18 10:06
81
Crohn disease: Semiautomatic MRI assessment
INTRODUCTION
Crohn disease (CD) is an inflammatory bowel disease, which can present
throughout the gastrointestinal tract, particularly affecting the small bowel and
the colon. Magnetic resonance imaging (MRI) is increasingly used for diagnosis
and phenotyping of CD because it is safe, noninvasive, and has high accuracy for
evaluating enteric disease and extramural complications (1). MRI features such as
wall thickness and T1 and T2 bowel wall signals have been validated as biomarkers of
CD activity, demonstrating good correlation with endoscopic and histopathologic
grading of inflammation (2, 3, 4).
Recent years have seen several MRI disease activity scores being developed and
externally validated, combining multiple MRI features to predict overall disease
activity (3, 4, 5, 6). These scores are gradually disseminating into clinical practice,
although at present, they are predominantly employed as research tools. The
magnetic resonance index of activity (MaRIA), for example, has been developed
using the Crohn’s Disease Endoscopic Index of Severity (CDEIS) as a reference
standard. The MaRIA is based on quantitative measurement of bowel wall relative
contrast enhancement, along with subjective evaluation of mural ulceration and
abnormal T2 signal (3). Other indices, such as the London score and the Crohn
disease MRI index (CDMI), rely on qualitative grading of various features by
reporting radiologists (4, 6). Such activity scores can be applied to individual bowel
segments, as well as to the patient as a whole, as both are important to clinical
management. Before MRI scores can be widely adopted for evaluating disease
activity and therapeutic monitoring, high accuracy across the spectrum of disease
severity and good reproducibility among radiologists must be proven. The current
literature, however, reports variable reproducibility for many features used in MRI
activity scores (6, 7).
One potential solution to the current limitations of MRI activity scoring is to
incorporate novel software solutions, which can automatically extract relevant
features from MRI data. Such software could reduce both interobserver variability
and the risk of observer bias inherent to subjective evaluation (8). New MRI image
processing methods are available, which give semiautomatic measurements of
bowel wall thickness, providing superior reproducibility over manual measurement
(9). Further techniques have been developed that automatically extract perfusion
Proefschrift2018-new2.indb 81 13/9/18 10:06
82
Chapter 4
parameters from motion corrected free-breathing dynamic contrast-enhanced
(DCE) MRI (10. Although several studies have shown the potential of semiautomatic
MRI assessment of CD (9, 10, 11), none of those have examined clinical practicability
or validated their results using a large, independent cohort.
We hypothesize that a scoring system combining semiautomatic software
measurements with conventional subjective radiologist scoring of MRI features
can improve accuracy and reproducibility in comparison to existing MRI scores.
Accordingly, our aim was to develop and validate a predictive MRI score for
ileocolonic CD activity incorporating novel software-assisted semiautomatic
measurement of MRI features using an ileocolonoscopic standard of reference, and
to compare its performance with existing MRI activity scores.
MATERIALS AND METHODS
The study was divided in two phases. Firstly, a detailed modeling process was
undertaken to derive two new MRI activity scores. Secondly, these new scores were
validated and compared to existing scores regarding accuracy for diagnosis and
grading of disease, as well as score reproducibility. Ethical permission was obtained
from both institutions’ medical ethics committee, and written informed consent was
obtained from all patients.
Phase 1 — Model Development
The modeling process employed a previously described cohort of 27 patients with
known CD (6. The first developed score specifically incorporated semiautomatic
measurements of bowel wall thickness and enhancement (described in more detail
further in phase 2) and was termed the “virtual gastrointestinal tract (VIGOR) score.”
The second score incorporated only the best performing combination of a number
of subjective evaluations made by radiologists (termed the “subjective score”). A
full description of the model development is given in Appendix A.
Phase 2 — Prospective Activity Score Testing and Model Comparison
The validation and comparison of the newly developed and existing activity scores
Proefschrift2018-new2.indb 82 13/9/18 10:06
82
Chapter 4
parameters from motion corrected free-breathing dynamic contrast-enhanced
(DCE) MRI (10. Although several studies have shown the potential of semiautomatic
MRI assessment of CD (9, 10, 11), none of those have examined clinical practicability
or validated their results using a large, independent cohort.
We hypothesize that a scoring system combining semiautomatic software
measurements with conventional subjective radiologist scoring of MRI features
can improve accuracy and reproducibility in comparison to existing MRI scores.
Accordingly, our aim was to develop and validate a predictive MRI score for
ileocolonic CD activity incorporating novel software-assisted semiautomatic
measurement of MRI features using an ileocolonoscopic standard of reference, and
to compare its performance with existing MRI activity scores.
MATERIALS AND METHODS
The study was divided in two phases. Firstly, a detailed modeling process was
undertaken to derive two new MRI activity scores. Secondly, these new scores were
validated and compared to existing scores regarding accuracy for diagnosis and
grading of disease, as well as score reproducibility. Ethical permission was obtained
from both institutions’ medical ethics committee, and written informed consent was
obtained from all patients.
Phase 1 — Model Development
The modeling process employed a previously described cohort of 27 patients with
known CD (6. The first developed score specifically incorporated semiautomatic
measurements of bowel wall thickness and enhancement (described in more detail
further in phase 2) and was termed the “virtual gastrointestinal tract (VIGOR) score.”
The second score incorporated only the best performing combination of a number
of subjective evaluations made by radiologists (termed the “subjective score”). A
full description of the model development is given in Appendix A.
Phase 2 — Prospective Activity Score Testing and Model Comparison
The validation and comparison of the newly developed and existing activity scores
Proefschrift2018-new2.indb 82 13/9/18 10:06
83
Crohn disease: Semiautomatic MRI assessment
were performed using an independent prospective cohort. Between October 2011
and September 2014, consecutive patients aged ≥18 years with suspected or known
CD and scheduled for ileocolonoscopy were recruited from two European tertiary
referral centres for inflammatory bowel disease (1. Academic Medical Center (AMC),
Amsterdam, the Netherlands, and 2. University College London Hospital (UCLH),
London, United Kingdom). All included patients underwent MRI and ileocolonoscopy
within 2 weeks. The Harvey-Bradshaw Index (HBI) was collected at the time of MRI
(12). Patient exclusion criteria were contraindications to MRI (eg, pacemakers and
claustrophobia), a final diagnosis other than CD, failure to comply with the oral
contrast protocol, >2 weeks between MRI and ileocolonoscopy, and incomplete
MRI protocol (eg, missing sequences or incomplete imaging), or insufficient
bowel cleansing precluding accurate mucosal assessment, as determined by the
endoscopist.
Reference Standard
Ileocolonoscopy was performed within 2 weeks of MRI using a standard endoscope
(model CF-160L, Olympus) by either a gastroenterologist or a senior resident in
gastroenterology under direct supervision of a gastroenterologist. The endoscopist
applied the CDEIS to evaluate endoscopic disease (13). The endoscopist was
blinded to findings on MRI, except for cases where a balloon-dilatation procedure
was indicated. In these cases, the length of stenosis on MRI was used to determine
the feasibility of balloon dilatation.
MRI Protocol
Patients fasted for at least 4 hours before the examination and were instructed to
drink a total of 2400 mL 2.5% mannitol solution (Baxter, Utrecht, The Netherlands)
split in two doses, 800 mL (3 hours before MRI) and 1600 mL (1 hour before MRI), to
achieve distension of both colonic and small bowel segments. MRI examinations were
performed on a 3-T MRI unit (Ingenia and Achieva; Philips, Best, The Netherlands) in
the supine position using a phased-array body coil. The MRI protocol used in both
centers is outlined in Appendix A. DCE images were mutually aligned using the
registration method described by Li et al. (10, 14).
Proefschrift2018-new2.indb 83 13/9/18 10:06
83
Crohn disease: Semiautomatic MRI assessment
were performed using an independent prospective cohort. Between October 2011
and September 2014, consecutive patients aged ≥18 years with suspected or known
CD and scheduled for ileocolonoscopy were recruited from two European tertiary
referral centres for inflammatory bowel disease (1. Academic Medical Center (AMC),
Amsterdam, the Netherlands, and 2. University College London Hospital (UCLH),
London, United Kingdom). All included patients underwent MRI and ileocolonoscopy
within 2 weeks. The Harvey-Bradshaw Index (HBI) was collected at the time of MRI
(12). Patient exclusion criteria were contraindications to MRI (eg, pacemakers and
claustrophobia), a final diagnosis other than CD, failure to comply with the oral
contrast protocol, >2 weeks between MRI and ileocolonoscopy, and incomplete
MRI protocol (eg, missing sequences or incomplete imaging), or insufficient
bowel cleansing precluding accurate mucosal assessment, as determined by the
endoscopist.
Reference Standard
Ileocolonoscopy was performed within 2 weeks of MRI using a standard endoscope
(model CF-160L, Olympus) by either a gastroenterologist or a senior resident in
gastroenterology under direct supervision of a gastroenterologist. The endoscopist
applied the CDEIS to evaluate endoscopic disease (13). The endoscopist was
blinded to findings on MRI, except for cases where a balloon-dilatation procedure
was indicated. In these cases, the length of stenosis on MRI was used to determine
the feasibility of balloon dilatation.
MRI Protocol
Patients fasted for at least 4 hours before the examination and were instructed to
drink a total of 2400 mL 2.5% mannitol solution (Baxter, Utrecht, The Netherlands)
split in two doses, 800 mL (3 hours before MRI) and 1600 mL (1 hour before MRI), to
achieve distension of both colonic and small bowel segments. MRI examinations were
performed on a 3-T MRI unit (Ingenia and Achieva; Philips, Best, The Netherlands) in
the supine position using a phased-array body coil. The MRI protocol used in both
centers is outlined in Appendix A. DCE images were mutually aligned using the
registration method described by Li et al. (10, 14).
Proefschrift2018-new2.indb 83 13/9/18 10:06
84
Chapter 4
Image Analysis
MRI examinations were evaluated using online viewer software (3Dnet Suite,
Biotronics3D, London, UK) by two pairs of observers (Ob1: C.Y.N, J.S.; Ob2. D.P,
S.T.) with extensive experience in MR enterography (>1100, >800, >500 and >1500
examinations, respectively). The first pair of observers was from AMC, the second
pair from UCLH. Each MRI dataset was independently evaluated by one observer
from both pairs, resulting in two evaluations per dataset. Observers were blinded
to each other’s findings and clinical data. Scan quality, luminal distension, and MRI
features from three existing validated MRI disease activity scores (MaRIA, London
score, and CDMI) were evaluated (3, 4). Details of the image analysis and the score
calculation are found in Appendix A.
Figure 1. (a) Placement of centerline points in the lumen of an affected transverse colon
segment on a coronal contrast-enhanced 3D T1-weighted SPGE image with fat saturation.
A few centerline points are placed in the middle of the lumen in one or more slices (yellow
dots). (b) The delineation of the inner and outer bowel wall surfaces is visualized by a red lines.
Presently, this is shown on a coronal slice but can be visualized in a similar way in reconstructed
sagittal or transversal planes.
Semiautomatic Measurements
Using our online viewer software, the bowel’s centerline was indicated on MRI
individually by each observer by manually placing a number of widely spaced points
Proefschrift2018-new2.indb 84 13/9/18 10:06
84
Chapter 4
Image Analysis
MRI examinations were evaluated using online viewer software (3Dnet Suite,
Biotronics3D, London, UK) by two pairs of observers (Ob1: C.Y.N, J.S.; Ob2. D.P,
S.T.) with extensive experience in MR enterography (>1100, >800, >500 and >1500
examinations, respectively). The first pair of observers was from AMC, the second
pair from UCLH. Each MRI dataset was independently evaluated by one observer
from both pairs, resulting in two evaluations per dataset. Observers were blinded
to each other’s findings and clinical data. Scan quality, luminal distension, and MRI
features from three existing validated MRI disease activity scores (MaRIA, London
score, and CDMI) were evaluated (3, 4). Details of the image analysis and the score
calculation are found in Appendix A.
Figure 1. (a) Placement of centerline points in the lumen of an affected transverse colon
segment on a coronal contrast-enhanced 3D T1-weighted SPGE image with fat saturation.
A few centerline points are placed in the middle of the lumen in one or more slices (yellow
dots). (b) The delineation of the inner and outer bowel wall surfaces is visualized by a red lines.
Presently, this is shown on a coronal slice but can be visualized in a similar way in reconstructed
sagittal or transversal planes.
Semiautomatic Measurements
Using our online viewer software, the bowel’s centerline was indicated on MRI
individually by each observer by manually placing a number of widely spaced points
Proefschrift2018-new2.indb 84 13/9/18 10:06
85
Crohn disease: Semiautomatic MRI assessment
within the lumen of the bowel on the postcontrast coronal T1-weighted sequence
(Fig 1). If a bowel segment harbored active disease (defined as a >0 score on at least
one subjective MRI feature), the centerline was placed across the affected part. In
the absence of disease activity, the centerline was placed in a representative part of
the bowel segment. Subsequently, the volume of the bowel wall was automatically
delineated using the segmentation method available in our online imaging viewers’
postprocessing environment (9). From this delineation, the following features were
automatically obtained: maximum bowel wall thickness (mm), mean bowel wall
thickness (mm), and excess bowel wall volume (mm3) (Appendix A). Additionally,
each delineation was used as a three-dimensional region of interest on DCE images
to extract the initial slope of increase (ISI) of the enhancement curve (the ISI
corresponds to the mathematically defined A1 feature in the reference paper) (10).
Validation of MRI Activity Scores and Statistical Analysis
Assessment of the validity of segmental scores in patients with CD can be challenging
because of the high numbers of healthy segments relative to the small number of
actively diseased segments (which may skew and inflate agreement statistics). For
this reason, we validated the newly developed scores in two ways.
The primary validation was restricted to segments with active disease on MRI from
the full prospective cohort. The applied definition of active disease (>0 score on at
least one subjective MRI feature) was chosen as a low threshold to obtain the highest
yield of segments in this primary analysis without creating a selection bias to one of
the activity scores. The selection was not based on endoscopic disease activity, as
this would require unblinding of endoscopic information to the radiologist. Grading
accuracy was evaluated by correlating segmental activity scores for each observer
individually against the segmental CDEIS score. Segments with missing model
features (ie, nonevaluable subjective features or failure to generate semiautomatic
features) were excluded, so that all activity scores were available in each segment.
Additionally, interobserver agreement was calculated for all overlapping active
segments (ie, deemed active by both observers) using the intraclass correlation
coefficient (ICC) for absolute agreement.
The secondary validation concerned the same evaluation of grading accuracy and
interobserver agreement on all segments (ie, active and healthy or in remission)
Proefschrift2018-new2.indb 85 13/9/18 10:06
85
Crohn disease: Semiautomatic MRI assessment
within the lumen of the bowel on the postcontrast coronal T1-weighted sequence
(Fig 1). If a bowel segment harbored active disease (defined as a >0 score on at least
one subjective MRI feature), the centerline was placed across the affected part. In
the absence of disease activity, the centerline was placed in a representative part of
the bowel segment. Subsequently, the volume of the bowel wall was automatically
delineated using the segmentation method available in our online imaging viewers’
postprocessing environment (9). From this delineation, the following features were
automatically obtained: maximum bowel wall thickness (mm), mean bowel wall
thickness (mm), and excess bowel wall volume (mm3) (Appendix A). Additionally,
each delineation was used as a three-dimensional region of interest on DCE images
to extract the initial slope of increase (ISI) of the enhancement curve (the ISI
corresponds to the mathematically defined A1 feature in the reference paper) (10).
Validation of MRI Activity Scores and Statistical Analysis
Assessment of the validity of segmental scores in patients with CD can be challenging
because of the high numbers of healthy segments relative to the small number of
actively diseased segments (which may skew and inflate agreement statistics). For
this reason, we validated the newly developed scores in two ways.
The primary validation was restricted to segments with active disease on MRI from
the full prospective cohort. The applied definition of active disease (>0 score on at
least one subjective MRI feature) was chosen as a low threshold to obtain the highest
yield of segments in this primary analysis without creating a selection bias to one of
the activity scores. The selection was not based on endoscopic disease activity, as
this would require unblinding of endoscopic information to the radiologist. Grading
accuracy was evaluated by correlating segmental activity scores for each observer
individually against the segmental CDEIS score. Segments with missing model
features (ie, nonevaluable subjective features or failure to generate semiautomatic
features) were excluded, so that all activity scores were available in each segment.
Additionally, interobserver agreement was calculated for all overlapping active
segments (ie, deemed active by both observers) using the intraclass correlation
coefficient (ICC) for absolute agreement.
The secondary validation concerned the same evaluation of grading accuracy and
interobserver agreement on all segments (ie, active and healthy or in remission)
Proefschrift2018-new2.indb 85 13/9/18 10:06
86
Chapter 4
from the subset of 50 patients. In these data, the distribution of disease forms a
skewed distribution of segmental score values, violating the assumption of normality
for the ICC, the standard measure for interobserver agreement in continuous data.
Accordingly, we applied both the conventional ICC and a modified, nonparametric
ICC by Rothery for a comprehensive evaluation of interobserver agreement (15).
This measure has been used in several studies (16, 17). The subset was determined by
random number generation from within the set of complete studies to minimize risk
of selection bias, whereas a sample size calculation was performed using previous
MRI performance data (Appendix A) (6). In both analyses, the developed scores
from phase 1 were compared to three existing MRI activity scores (MaRIA, London
score, and CDMI). Diagnostic accuracy and per-patient analysis were performed
using the subset of 50 patients, as detailed in Appendix A.
Spearman rank correlations were interpreted as follows: 0–0.20, very weak;
≥0–0.40, weak; ≥0.40–0.60, moderate; ≥0.60–0.80, strong; and ≥0–1.00, very
strong. Correlation coefficients were then compared using the Steiger Z test for
(non)overlapping, dependent correlations (18. Interobserver agreement (ICC or
nonparametric ICC) was evaluated using the following criteria for interpretation:
0–0.20, poor; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, good; and 0.81–1.00,
very good (19. Diagnostic accuracy values were compared using the McNemar test.
We considered a P value of <.05 to indicate a statistically significant difference.
Model development and validation were implemented with R Statistical language
(v3.1.2, Vienna, Austria) (20. Descriptive statistics were analyzed using SPSS 22 for
Mac (SPSS, Chicago, IL).
RESULTS
Phase 1 — Model Development
The developed VIGOR and subjective models were:
VIGOR score = 17.1 x ISI + 0.2 x excess volume + 2.3 x mural T2
Subjective score = 0.03 x RCE + 0.9 x mural thickness (mm) + 3 x mural T2
Proefschrift2018-new2.indb 86 13/9/18 10:06
86
Chapter 4
from the subset of 50 patients. In these data, the distribution of disease forms a
skewed distribution of segmental score values, violating the assumption of normality
for the ICC, the standard measure for interobserver agreement in continuous data.
Accordingly, we applied both the conventional ICC and a modified, nonparametric
ICC by Rothery for a comprehensive evaluation of interobserver agreement (15).
This measure has been used in several studies (16, 17). The subset was determined by
random number generation from within the set of complete studies to minimize risk
of selection bias, whereas a sample size calculation was performed using previous
MRI performance data (Appendix A) (6). In both analyses, the developed scores
from phase 1 were compared to three existing MRI activity scores (MaRIA, London
score, and CDMI). Diagnostic accuracy and per-patient analysis were performed
using the subset of 50 patients, as detailed in Appendix A.
Spearman rank correlations were interpreted as follows: 0–0.20, very weak;
≥0–0.40, weak; ≥0.40–0.60, moderate; ≥0.60–0.80, strong; and ≥0–1.00, very
strong. Correlation coefficients were then compared using the Steiger Z test for
(non)overlapping, dependent correlations (18. Interobserver agreement (ICC or
nonparametric ICC) was evaluated using the following criteria for interpretation:
0–0.20, poor; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, good; and 0.81–1.00,
very good (19. Diagnostic accuracy values were compared using the McNemar test.
We considered a P value of <.05 to indicate a statistically significant difference.
Model development and validation were implemented with R Statistical language
(v3.1.2, Vienna, Austria) (20. Descriptive statistics were analyzed using SPSS 22 for
Mac (SPSS, Chicago, IL).
RESULTS
Phase 1 — Model Development
The developed VIGOR and subjective models were:
VIGOR score = 17.1 x ISI + 0.2 x excess volume + 2.3 x mural T2
Subjective score = 0.03 x RCE + 0.9 x mural thickness (mm) + 3 x mural T2
Proefschrift2018-new2.indb 86 13/9/18 10:06
87
Crohn disease: Semiautomatic MRI assessment
A VIGOR score of ≥5.6 was determined via receiver operating characteristic
analysis as the optimal cutoff value for active disease (CDEIS score ≥3). For the
subjective score, the optimal cutoff value for active disease was ≥4.8. Details of the
development cohorts’ segmental exclusions are shown in Appendix B.
Phase 2 — Prospective Activity Score Testing and Comparison
After exclusions (Fig 2), the final prospective study cohort consisted of 106 patients
with known CD, for which demographics and clinical characteristics are provided
in Table 1. Characteristics of the 50 patients’ randomly determined subset used for
evaluation of diagnostic accuracy and per-patient scores can be found in Appendix
B. One patient experienced abdominal pain and cramping after the MRI examination,
which were successfully treated with simple analgesia.
The mean scan image quality (0–3) was 2.2 (standard deviation: 0.6). The mean
distension value (0–4) for both terminal ileum and colon was 3.4 (standard
deviation: 0.7). Within evaluable segments (evaluable on MRI by the radiologist and
at endoscopic intubation), Ob1 and Ob2 identified 88 and 95 segments with active
disease on MRI, respectively. In the subset of 50 patients, a total of 230 and 229
segments (both active and healthy and in remission) were evaluable for Ob1 and
Ob2, respectively.
Figure 2. Flow diagram detailing patient inclusions and exclusions
Proefschrift2018-new2.indb 87 13/9/18 10:06
87
Crohn disease: Semiautomatic MRI assessment
A VIGOR score of ≥5.6 was determined via receiver operating characteristic
analysis as the optimal cutoff value for active disease (CDEIS score ≥3). For the
subjective score, the optimal cutoff value for active disease was ≥4.8. Details of the
development cohorts’ segmental exclusions are shown in Appendix B.
Phase 2 — Prospective Activity Score Testing and Comparison
After exclusions (Fig 2), the final prospective study cohort consisted of 106 patients
with known CD, for which demographics and clinical characteristics are provided
in Table 1. Characteristics of the 50 patients’ randomly determined subset used for
evaluation of diagnostic accuracy and per-patient scores can be found in Appendix
B. One patient experienced abdominal pain and cramping after the MRI examination,
which were successfully treated with simple analgesia.
The mean scan image quality (0–3) was 2.2 (standard deviation: 0.6). The mean
distension value (0–4) for both terminal ileum and colon was 3.4 (standard
deviation: 0.7). Within evaluable segments (evaluable on MRI by the radiologist and
at endoscopic intubation), Ob1 and Ob2 identified 88 and 95 segments with active
disease on MRI, respectively. In the subset of 50 patients, a total of 230 and 229
segments (both active and healthy and in remission) were evaluable for Ob1 and
Ob2, respectively.
Figure 2. Flow diagram detailing patient inclusions and exclusions
Proefschrift2018-new2.indb 87 13/9/18 10:06
88
Chapter 4
Table 1. Clinical Characteristics of the Prospective Cohort
Total no. of patients 106
Female, n (%) 59 (56)
Age at MRI (y), median (IQR) 33 (26–44)
Previous surgery, n (%) 42 (40)
Concomitant treatments
Anti-TNF antibodies, n (%) 30 (28)
Steroids, n (%) 18 (17)
Thiopurines, n (%) 14 (13)
5-ASA, n (%) 19 (18)
Methotrexate, n (%) 8 (8)
CRP (mg/L), median (IQR) 5 (1–13)
HBI value, median (IQR) 5 (2–8)
CDEIS score, median (IQR) 3.2 (0.5–6.4)
Montreal classification
Age at diagnosis (y), median (IQR) 22 (17–28)
Disease location
L1 ileal, n (%) 43 (41)
L2 colonic, n (%) 15 (14)
L3 ileocolonic, n (%) 48 (45)
L4 upper GI tract involvement, n (%) 4 (4)
Disease behavior
B1 inflammatory 54 (51)
B2 stricturing 36 (34)
B3 penetrating 16 (15)
Perianal involvement, n (%) 23 (22)
5-ASA, 5-acetylsalicylic acid; CDEIS, Crohn’s Disease Endoscopic Index of Severity; CRP, C-reactive protein; GI, gastrointestinal; HBI, Harvey-Bradshaw Index; IQR, interquartile range; MRI, magnetic resonance imaging; TNF, tumor necrosis factor.
In active segments (>0 score on at least one subjective feature), the VIGOR score
could be calculated in 83% (73/88) of the segments for Ob1 and in 73% (69/95) of
the segments for Ob2. In the subset with 50 patients, the VIGOR score could be
applied to 73% (167/230) of the segments for Ob1. Exclusion of rectum segments
from the analysis increased this rate to 87% (161/186). For Ob2, the VIGOR score
was applied to 70% (161/229) of the segments, which increased to 82% (153/187)
after the exclusion of rectum segments. Details on the inclusion of bowel segments
can be found in Table 2.
Proefschrift2018-new2.indb 88 13/9/18 10:06
88
Chapter 4
Table 1. Clinical Characteristics of the Prospective Cohort
Total no. of patients 106
Female, n (%) 59 (56)
Age at MRI (y), median (IQR) 33 (26–44)
Previous surgery, n (%) 42 (40)
Concomitant treatments
Anti-TNF antibodies, n (%) 30 (28)
Steroids, n (%) 18 (17)
Thiopurines, n (%) 14 (13)
5-ASA, n (%) 19 (18)
Methotrexate, n (%) 8 (8)
CRP (mg/L), median (IQR) 5 (1–13)
HBI value, median (IQR) 5 (2–8)
CDEIS score, median (IQR) 3.2 (0.5–6.4)
Montreal classification
Age at diagnosis (y), median (IQR) 22 (17–28)
Disease location
L1 ileal, n (%) 43 (41)
L2 colonic, n (%) 15 (14)
L3 ileocolonic, n (%) 48 (45)
L4 upper GI tract involvement, n (%) 4 (4)
Disease behavior
B1 inflammatory 54 (51)
B2 stricturing 36 (34)
B3 penetrating 16 (15)
Perianal involvement, n (%) 23 (22)
5-ASA, 5-acetylsalicylic acid; CDEIS, Crohn’s Disease Endoscopic Index of Severity; CRP, C-reactive protein; GI, gastrointestinal; HBI, Harvey-Bradshaw Index; IQR, interquartile range; MRI, magnetic resonance imaging; TNF, tumor necrosis factor.
In active segments (>0 score on at least one subjective feature), the VIGOR score
could be calculated in 83% (73/88) of the segments for Ob1 and in 73% (69/95) of
the segments for Ob2. In the subset with 50 patients, the VIGOR score could be
applied to 73% (167/230) of the segments for Ob1. Exclusion of rectum segments
from the analysis increased this rate to 87% (161/186). For Ob2, the VIGOR score
was applied to 70% (161/229) of the segments, which increased to 82% (153/187)
after the exclusion of rectum segments. Details on the inclusion of bowel segments
can be found in Table 2.
Proefschrift2018-new2.indb 88 13/9/18 10:06
89
Crohn disease: Semiautomatic MRI assessment
Model Validation and Comparison
Correlations to CDEIS scores for each observer pair and interobserver agreement
are presented in Table 3. In active segments, the VIGOR score showed moderate
correlations to CDEIS scores (Ob1: r = 0.58 and Ob2: r = 0.59). Weak-to-moderate
correlations to CDEIS scores were seen for the subjective score (r = 0.39 and 0.51),
the MaRIA (r = 0.40 and 0.43), the London score (r = 0.38 and 0.45), and the CDMI
(r = 0.34 and 0.48). Significant differences were seen for Ob1 between the VIGOR
score and the subjective score (P = .04), the London score (P = .03), and the CDMI
(P = .01), but not for the MaRIA (P = .05). For Ob2, no significant differences were
seen (P = .10–.35). The VIGOR score showed very good interobserver agreement in
active segments (ICC = 0.81) compared to fair agreement for other activity scores
(ICC = 0.44–0.59). Interobserver scatter plots for all scores can be found in Appendix
B, which shows visually similar agreement for the analyses on the active segments
of the full dataset and all segments of the subset, whereas in the latter, all scores
show narrow clustering (ie, high reproducibility) of healthy segments.
In the subset of 50 patients including all segments (active and healthy and
remission), the VIGOR score showed moderate correlation to CDEIS scores (Ob1:
r = 0.57 and Ob2: r = 0.53) for segmental disease activity, whereas the correlations
for the other activity scores ranged between 0.50 and 0.61 for Ob1 and between
0.53 and 0.64 for Ob2. No significant differences were seen between the VIGOR
score and other activity scores for Ob1 (P = .2–.6). For Ob2, the CDMI and the London
scores showed significantly higher correlation to CDEIS scores compared to the
other activity scores (P = .02–.03). Conventional ICC values for active segments and
all segments and nonparametric ICC values for all segments from the subset of 50
patients are shown in Table 4. It can be observed that the conventional ICC values
for all segments were evidently higher compared to ICC values in active segments
and the nonparametric ICC values, especially for the subjective and existing activity
scores. Using the nonparametric ICC values, the VIGOR score showed very good
agreement of (ICC = 0.89) compared to poor-to-fair agreement for other activity
scores (ICC = 0.33–0.56), which was a significant difference (P < .001).
Proefschrift2018-new2.indb 89 13/9/18 10:06
89
Crohn disease: Semiautomatic MRI assessment
Model Validation and Comparison
Correlations to CDEIS scores for each observer pair and interobserver agreement
are presented in Table 3. In active segments, the VIGOR score showed moderate
correlations to CDEIS scores (Ob1: r = 0.58 and Ob2: r = 0.59). Weak-to-moderate
correlations to CDEIS scores were seen for the subjective score (r = 0.39 and 0.51),
the MaRIA (r = 0.40 and 0.43), the London score (r = 0.38 and 0.45), and the CDMI
(r = 0.34 and 0.48). Significant differences were seen for Ob1 between the VIGOR
score and the subjective score (P = .04), the London score (P = .03), and the CDMI
(P = .01), but not for the MaRIA (P = .05). For Ob2, no significant differences were
seen (P = .10–.35). The VIGOR score showed very good interobserver agreement in
active segments (ICC = 0.81) compared to fair agreement for other activity scores
(ICC = 0.44–0.59). Interobserver scatter plots for all scores can be found in Appendix
B, which shows visually similar agreement for the analyses on the active segments
of the full dataset and all segments of the subset, whereas in the latter, all scores
show narrow clustering (ie, high reproducibility) of healthy segments.
In the subset of 50 patients including all segments (active and healthy and
remission), the VIGOR score showed moderate correlation to CDEIS scores (Ob1:
r = 0.57 and Ob2: r = 0.53) for segmental disease activity, whereas the correlations
for the other activity scores ranged between 0.50 and 0.61 for Ob1 and between
0.53 and 0.64 for Ob2. No significant differences were seen between the VIGOR
score and other activity scores for Ob1 (P = .2–.6). For Ob2, the CDMI and the London
scores showed significantly higher correlation to CDEIS scores compared to the
other activity scores (P = .02–.03). Conventional ICC values for active segments and
all segments and nonparametric ICC values for all segments from the subset of 50
patients are shown in Table 4. It can be observed that the conventional ICC values
for all segments were evidently higher compared to ICC values in active segments
and the nonparametric ICC values, especially for the subjective and existing activity
scores. Using the nonparametric ICC values, the VIGOR score showed very good
agreement of (ICC = 0.89) compared to poor-to-fair agreement for other activity
scores (ICC = 0.33–0.56), which was a significant difference (P < .001).
Proefschrift2018-new2.indb 89 13/9/18 10:06
90
Chapter 4
Diagnostic Accuracy
The diagnostic accuracy for all MRI scores are presented in Table 5. No significant
differences in diagnostic accuracy were seen (P > .05), except for the subjective
scores’ significantly lower accuracy for Ob1 compared to other activity scores
(P < .01).
Per-patient activity scores in the subset showed moderate correlations to CDEIS
scores for the VIGOR score (Ob1: r = 0.53 and Ob2: r = 0.54), the subjective score
(r = 0.60 and 0.57), the MaRIA (r = 0.58 and 0.51), the London score (r = 0.58 and
0.56), and the CDMI (r = 0.53 and 0.59). There were no significant differences
between any pair of activity scores (P > .05). Per-patient scores showed similar
(conventional) ICC values for the VIGOR score (0.77, 95% confidence interval [CI]:
0.62–0.86), the subjective score (0.71, 95% CI: 0.51–0.83), the MaRIA (0.75, 95% CI:
0.54–0.87), the London score (0.74, 95% CI: 0.57–0.84), and the CDMI (0.79, 95%
CI: 0.65–0.88).
Table 2. Segment inclusions and exclusions
Active segments Subset (n=50), all Subset (n=50), Rectum excluded
Ob1 Ob2 Ob1 Ob2 Ob1 Ob2
Total no. of segment* 88 95 230 229 186 187
Inclusions (%) 73 (83) 69 (73) 167 (73) 161 (70) 161 (87) 153 (82)
Terminal ileum 54 49 39 41 39 41
Ascending colon 9 9 44 41 44 41
Transverse colon 4 2 39 38 39 38
Desc/sigmoid colon 6 9 39 33 39 33
Rectum 0 0 6 8 – –
Exclusions (%) 15 (17) 26 (27) 63 (27) 68 (30) 25 (13) 34 (18)
Outside DCE 3 7 42 40 12 13
Failed DCE registra-tion
7 7 1 1 1 1
Fecal residue 3 1 6 6 2 2
Poor distension 0 2 6 6 3 3
Artefacts 0 2 0 1 0 1
Failed segmentation 2 7 8 14 7 14
DCE, dynamic contrast enhanced; Ob1, observer 1; Ob2, observer 2.* All segments which could be evaluated by the radiologist and endoscopist.
Proefschrift2018-new2.indb 90 13/9/18 10:06
90
Chapter 4
Diagnostic Accuracy
The diagnostic accuracy for all MRI scores are presented in Table 5. No significant
differences in diagnostic accuracy were seen (P > .05), except for the subjective
scores’ significantly lower accuracy for Ob1 compared to other activity scores
(P < .01).
Per-patient activity scores in the subset showed moderate correlations to CDEIS
scores for the VIGOR score (Ob1: r = 0.53 and Ob2: r = 0.54), the subjective score
(r = 0.60 and 0.57), the MaRIA (r = 0.58 and 0.51), the London score (r = 0.58 and
0.56), and the CDMI (r = 0.53 and 0.59). There were no significant differences
between any pair of activity scores (P > .05). Per-patient scores showed similar
(conventional) ICC values for the VIGOR score (0.77, 95% confidence interval [CI]:
0.62–0.86), the subjective score (0.71, 95% CI: 0.51–0.83), the MaRIA (0.75, 95% CI:
0.54–0.87), the London score (0.74, 95% CI: 0.57–0.84), and the CDMI (0.79, 95%
CI: 0.65–0.88).
Table 2. Segment inclusions and exclusions
Active segments Subset (n=50), all Subset (n=50), Rectum excluded
Ob1 Ob2 Ob1 Ob2 Ob1 Ob2
Total no. of segment* 88 95 230 229 186 187
Inclusions (%) 73 (83) 69 (73) 167 (73) 161 (70) 161 (87) 153 (82)
Terminal ileum 54 49 39 41 39 41
Ascending colon 9 9 44 41 44 41
Transverse colon 4 2 39 38 39 38
Desc/sigmoid colon 6 9 39 33 39 33
Rectum 0 0 6 8 – –
Exclusions (%) 15 (17) 26 (27) 63 (27) 68 (30) 25 (13) 34 (18)
Outside DCE 3 7 42 40 12 13
Failed DCE registra-tion
7 7 1 1 1 1
Fecal residue 3 1 6 6 2 2
Poor distension 0 2 6 6 3 3
Artefacts 0 2 0 1 0 1
Failed segmentation 2 7 8 14 7 14
DCE, dynamic contrast enhanced; Ob1, observer 1; Ob2, observer 2.* All segments which could be evaluated by the radiologist and endoscopist.
Proefschrift2018-new2.indb 90 13/9/18 10:06
91
Crohn disease: Semiautomatic MRI assessment
Table 3. Correlations between MRI activity scores and Crohn’s disease endoscopic index of
severity (CDEIS) and interobserver agreement in the active segments of the full prospective
cohort
Observer 1 (n=73)
Observer 2 (n=69)
Interobserver agreement (n=56)
MRI features r P Value r P Value ICC (95% CI)
VIGOR score 0.58 <0.001 0.59 <0.001 0.81 (0.56–0.91)
Subjective score 0.39 0.001 0.51 <0.001 0.44 (0.21–0.63)
MaRIA 0.40 0.001 0.43 <0.001 0.44 (0.21–0.63)
London score 0.38 0.001 0.45 <0.001 0.47 (0.24–0.65)
CDMI 0.34 0.003 0.48 <0.001 0.59 (0.40–0.74)
CDMI, Crohn disease MRI index; CI, confidence interval; ICC, intraclass correlation coef-ficient; MaRIA, magnetic resonance index of activity; MRI, magnetic resonance imaging; VIGOR, virtual gastrointestinal tract.
Table 4. Interobserver agreement for segmental scores of the 50-patient subset in active
segments and in all segments
Active (n=43) All (n=146)
MRI features ICC (95% CI) ICC (95% CI) Nonparametric ICC (Rothery)
VIGOR score 0.70 (0.51-0.82) 0.87 (0.83-0.91) 0.89
Subjective score 0.44 (0.16-0.65) 0.77 (0.69-0.83) 0.53
MaRIA 0.45 (0.18-0.66) 0.77 (0.69-0.83) 0.33
London score 0.44 (0.16-0.65) 0.81 (0.75-0.86) 0.53
CDMI 0.55 (0.30-0.73) 0.86 (0.81-0.90) 0.56
CDMI, Crohn disease MRI index; CI, confidence interval; ICC, intraclass correlation coef-ficient; MaRIA, magnetic resonance index of activity; MRI, magnetic resonance imaging; VIGOR, virtual gastrointestinal tract.Original ICC values are shown for both groups, whereas the nonparametric ICC is shown for all segments to account for the skewed distribution in this dataset.
Proefschrift2018-new2.indb 91 13/9/18 10:06
91
Crohn disease: Semiautomatic MRI assessment
Table 3. Correlations between MRI activity scores and Crohn’s disease endoscopic index of
severity (CDEIS) and interobserver agreement in the active segments of the full prospective
cohort
Observer 1 (n=73)
Observer 2 (n=69)
Interobserver agreement (n=56)
MRI features r P Value r P Value ICC (95% CI)
VIGOR score 0.58 <0.001 0.59 <0.001 0.81 (0.56–0.91)
Subjective score 0.39 0.001 0.51 <0.001 0.44 (0.21–0.63)
MaRIA 0.40 0.001 0.43 <0.001 0.44 (0.21–0.63)
London score 0.38 0.001 0.45 <0.001 0.47 (0.24–0.65)
CDMI 0.34 0.003 0.48 <0.001 0.59 (0.40–0.74)
CDMI, Crohn disease MRI index; CI, confidence interval; ICC, intraclass correlation coef-ficient; MaRIA, magnetic resonance index of activity; MRI, magnetic resonance imaging; VIGOR, virtual gastrointestinal tract.
Table 4. Interobserver agreement for segmental scores of the 50-patient subset in active
segments and in all segments
Active (n=43) All (n=146)
MRI features ICC (95% CI) ICC (95% CI) Nonparametric ICC (Rothery)
VIGOR score 0.70 (0.51-0.82) 0.87 (0.83-0.91) 0.89
Subjective score 0.44 (0.16-0.65) 0.77 (0.69-0.83) 0.53
MaRIA 0.45 (0.18-0.66) 0.77 (0.69-0.83) 0.33
London score 0.44 (0.16-0.65) 0.81 (0.75-0.86) 0.53
CDMI 0.55 (0.30-0.73) 0.86 (0.81-0.90) 0.56
CDMI, Crohn disease MRI index; CI, confidence interval; ICC, intraclass correlation coef-ficient; MaRIA, magnetic resonance index of activity; MRI, magnetic resonance imaging; VIGOR, virtual gastrointestinal tract.Original ICC values are shown for both groups, whereas the nonparametric ICC is shown for all segments to account for the skewed distribution in this dataset.
Proefschrift2018-new2.indb 91 13/9/18 10:06
92
Chapter 4
Tab
le 5
. Dia
gn
ost
ic A
ccu
racy f
or
Seg
men
tal M
ag
neti
c R
eso
nan
ce Im
ag
ing
Acti
vit
y S
co
res
for
Dete
cti
on
of
Acti
ve D
isease
(C
roh
n’s
Dis
ease
En
do
sco
pic
In
dex [
CD
EIS
] ≥
3)
Ob
serv
er 1
Ob
serv
er 2
Sens
itiv
ity
(%)
Spec
ifici
ty
(%)
PP
V
(%)
NP
V
(%)
Acc
urac
y (%
)Se
nsit
ivit
y (%
)Sp
ecifi
city
(%
)P
PV
(%
)N
PV
(%
)A
ccur
acy
(%)
VIG
OR
sco
re76
84
63
90
81
748
25
89
08
0
Su
bje
cti
ve s
co
re78
67
47
89
70
748
25
89
08
0
MaR
IA6
78
66
48
88
16
49
17
18
88
4
Lo
nd
on
sco
re6
09
68
48
7 8
657
94
77
86
84
CD
MI
60
92
73
86
83
62
91
72
87
83
CD
MI, C
roh
n d
isease
MR
I in
dex; M
aR
IA, m
ag
neti
c r
eso
nan
ce in
dex o
f acti
vit
y; N
PV
, n
eg
ati
ve p
red
icti
ve v
alu
e; P
PV
, p
osi
tive p
red
icti
ve v
alu
e; V
IGO
R, vir
tual
gast
roin
test
inal tr
act.
Proefschrift2018-new2.indb 92 13/9/18 10:06
92
Chapter 4
Tab
le 5
. Dia
gn
ost
ic A
ccu
racy f
or
Seg
men
tal M
ag
neti
c R
eso
nan
ce Im
ag
ing
Acti
vit
y S
co
res
for
Dete
cti
on
of
Acti
ve D
isease
(C
roh
n’s
Dis
ease
En
do
sco
pic
In
dex [
CD
EIS
] ≥
3)
Ob
serv
er 1
Ob
serv
er 2
Sens
itiv
ity
(%)
Spec
ifici
ty
(%)
PP
V
(%)
NP
V
(%)
Acc
urac
y (%
)Se
nsit
ivit
y (%
)Sp
ecifi
city
(%
)P
PV
(%
)N
PV
(%
)A
ccur
acy
(%)
VIG
OR
sco
re76
84
63
90
81
748
25
89
08
0
Su
bje
cti
ve s
co
re78
67
47
89
70
748
25
89
08
0
MaR
IA6
78
66
48
88
16
49
17
18
88
4
Lo
nd
on
sco
re6
09
68
48
7 8
657
94
77
86
84
CD
MI
60
92
73
86
83
62
91
72
87
83
CD
MI, C
roh
n d
isease
MR
I in
dex; M
aR
IA, m
ag
neti
c r
eso
nan
ce in
dex o
f acti
vit
y; N
PV
, n
eg
ati
ve p
red
icti
ve v
alu
e; P
PV
, p
osi
tive p
red
icti
ve v
alu
e; V
IGO
R, vir
tual
gast
roin
test
inal tr
act.
Proefschrift2018-new2.indb 92 13/9/18 10:06
93
Crohn disease: Semiautomatic MRI assessment
DISCUSSION
In this development and validation study, evidence is provided for a new MRI
CD activity scoring system, the “VIGOR score”, incorporating both subjective
observations and semiautomatic features. The VIGOR score achieved improved
segmental reproducibility compared to existing activity scores, such as the MaRIA,
the London score, and the CDMI. The VIGOR score showed similar correlation with
the endoscopic standard of reference and diagnostic accuracy compared to other
activity scores. The VIGOR score also showed superior performance in comparison
to a new subjective score, which was developed and validated using the same
cohorts. When considering the per-patient VIGOR score, correlation with CDEIS
scores remained moderate and interobserver agreement remained very good. In
contrast to the segmental analyses, per-patient scores showed high agreement for
all activity scores. This difference can be explained through the high reproducibility
of all activity scores in healthy segments (Appendix B), which considerably
influences the per-patient scores’ agreement due to their high prevalence.
MRI activity scores are currently being investigated for use as outcome measures in
clinical trials, with some success (21, 22). Clearly, for use in multicenter studies, a high
level of reproducibility between readers is imperative. Therapeutic management
requires high reproducibility in both segmental and patient scores, as these serve
different purposes in guidance and evaluation of surgical and medical therapies.
Many patients CD have limited segmental disease (usually ileocecal disease), such
that segmental reproducibility for disease activity is paramount. Conversely, a more
global overview is important in those with multifocal disease. Our study reports very
encouraging performance characteristics for the newly developed semiautomatic
score: correlation with CDEIS scores is at least as good as existing scores, yet only
the VIGOR score maintained high reproducibility in both per-segment and per-
patient analyses. The next stage of development should now investigate the ability
of the VIGOR score to monitor therapy via longitudinal studies, similar to the work
by Ordas et al. evaluating the MaRIA (22).
Compared to existing evaluations of MRI activity scores, we found relatively low
correlations with CDEIS scores (5, 6, 22). We hypothesize that this is caused by the
disease spectrum in our prospective cohort, with relatively high prevalence of mild
disease. This hypothesis is confirmed by the median CDEIS, C-reactive protein, and
Proefschrift2018-new2.indb 93 13/9/18 10:06
93
Crohn disease: Semiautomatic MRI assessment
DISCUSSION
In this development and validation study, evidence is provided for a new MRI
CD activity scoring system, the “VIGOR score”, incorporating both subjective
observations and semiautomatic features. The VIGOR score achieved improved
segmental reproducibility compared to existing activity scores, such as the MaRIA,
the London score, and the CDMI. The VIGOR score showed similar correlation with
the endoscopic standard of reference and diagnostic accuracy compared to other
activity scores. The VIGOR score also showed superior performance in comparison
to a new subjective score, which was developed and validated using the same
cohorts. When considering the per-patient VIGOR score, correlation with CDEIS
scores remained moderate and interobserver agreement remained very good. In
contrast to the segmental analyses, per-patient scores showed high agreement for
all activity scores. This difference can be explained through the high reproducibility
of all activity scores in healthy segments (Appendix B), which considerably
influences the per-patient scores’ agreement due to their high prevalence.
MRI activity scores are currently being investigated for use as outcome measures in
clinical trials, with some success (21, 22). Clearly, for use in multicenter studies, a high
level of reproducibility between readers is imperative. Therapeutic management
requires high reproducibility in both segmental and patient scores, as these serve
different purposes in guidance and evaluation of surgical and medical therapies.
Many patients CD have limited segmental disease (usually ileocecal disease), such
that segmental reproducibility for disease activity is paramount. Conversely, a more
global overview is important in those with multifocal disease. Our study reports very
encouraging performance characteristics for the newly developed semiautomatic
score: correlation with CDEIS scores is at least as good as existing scores, yet only
the VIGOR score maintained high reproducibility in both per-segment and per-
patient analyses. The next stage of development should now investigate the ability
of the VIGOR score to monitor therapy via longitudinal studies, similar to the work
by Ordas et al. evaluating the MaRIA (22).
Compared to existing evaluations of MRI activity scores, we found relatively low
correlations with CDEIS scores (5, 6, 22). We hypothesize that this is caused by the
disease spectrum in our prospective cohort, with relatively high prevalence of mild
disease. This hypothesis is confirmed by the median CDEIS, C-reactive protein, and
Proefschrift2018-new2.indb 93 13/9/18 10:06
94
Chapter 4
HBI values from our prospective cohort (Table 1 and Appendix B), which are much
lower than those in previous studies (3, 4). Furthermore, our results are accordant
with previous results from our two inclusion centers (4, 6).
The presence of mural ulceration has been reported as a useful sign of activity and
is incorporated in the MaRIA. However, we did not include evaluation of ulceration
in our model development as data suggest that it is highly reader dependent (6.
Furthermore, all five MRI scores (four of which did not include ulceration) achieved
similar correlation to CDEIS scores and diagnostic accuracy for active segments.
Our primary analysis was limited to active segments as large numbers of normal
segments can skew agreement statistics and result in overoptimistic estimates.
The skewing of data is confirmed by our results; increased ICC values are seen
for subjective activity scores in the inclusive analyses of all segments, whereas
no improved agreement is observed visually in the corresponding scatter plots or
when using the nonparametric ICC values.
Our study has several limitations. The DCE sequence employed in our development
cohort used a smaller field of view compared to the sequence used in the
prospective cohort, which limited the amount of ISI data for model development.
Because the field was positioned on the terminal ileum, the excluded segments
from the development cohort were mainly colonic and rectum segments (81% of
exclusions). Exclusions were improved considerably in the prospective cohort,
although a relatively large number of rectum segments were excluded for being out
of the field of view on DCE. Simultaneously, our results do reveal current limitations
of semiautomatic features, as measurements in segments with suboptimal
preparation were limited. Although subjective evaluation is also affected, human
interpretation remains superior in coping with the effects of suboptimal preparation
on mural thickness and contrast enhancement. However, semiautomatic software,
together with MRI sequences, continuously undergoes improvement, and as such,
an increase in success rate can be expected. These improvements might prove
especially beneficial for inexperienced MRI readers. Although all readers in our study
had extensive experience in magnetic resonance enterography, future research
should explore the semiautomatic scores’ application by readers of different levels
of experience.
Proefschrift2018-new2.indb 94 13/9/18 10:06
94
Chapter 4
HBI values from our prospective cohort (Table 1 and Appendix B), which are much
lower than those in previous studies (3, 4). Furthermore, our results are accordant
with previous results from our two inclusion centers (4, 6).
The presence of mural ulceration has been reported as a useful sign of activity and
is incorporated in the MaRIA. However, we did not include evaluation of ulceration
in our model development as data suggest that it is highly reader dependent (6.
Furthermore, all five MRI scores (four of which did not include ulceration) achieved
similar correlation to CDEIS scores and diagnostic accuracy for active segments.
Our primary analysis was limited to active segments as large numbers of normal
segments can skew agreement statistics and result in overoptimistic estimates.
The skewing of data is confirmed by our results; increased ICC values are seen
for subjective activity scores in the inclusive analyses of all segments, whereas
no improved agreement is observed visually in the corresponding scatter plots or
when using the nonparametric ICC values.
Our study has several limitations. The DCE sequence employed in our development
cohort used a smaller field of view compared to the sequence used in the
prospective cohort, which limited the amount of ISI data for model development.
Because the field was positioned on the terminal ileum, the excluded segments
from the development cohort were mainly colonic and rectum segments (81% of
exclusions). Exclusions were improved considerably in the prospective cohort,
although a relatively large number of rectum segments were excluded for being out
of the field of view on DCE. Simultaneously, our results do reveal current limitations
of semiautomatic features, as measurements in segments with suboptimal
preparation were limited. Although subjective evaluation is also affected, human
interpretation remains superior in coping with the effects of suboptimal preparation
on mural thickness and contrast enhancement. However, semiautomatic software,
together with MRI sequences, continuously undergoes improvement, and as such,
an increase in success rate can be expected. These improvements might prove
especially beneficial for inexperienced MRI readers. Although all readers in our study
had extensive experience in magnetic resonance enterography, future research
should explore the semiautomatic scores’ application by readers of different levels
of experience.
Proefschrift2018-new2.indb 94 13/9/18 10:06
95
Crohn disease: Semiautomatic MRI assessment
Currently, steps are being taken to further technically optimize the semiautomatic
MRI measurements and to provide full integration in viewer software. Clearly, these
aspects are essential for clinical applicability, which requires easy-to-use techniques.
In conclusion, the use of semiautomatic features for the assessment of patients with
CD maintains diagnostic and grading accuracy while improving reproducibility over
conventional activity scores. These characteristics make it potentially suitable for
therapy evaluation and monitoring of disease activity. Furthermore, accurate and
reproducible MRI scores could improve the physician’s trust in these scores to make
consistent and effective treatment decisions.
ACKNOWLEDGMENTS
The authors thank Ernst Harting for the management of the VIGOR++ project and
Costis Kompis for project exploitation, Rado Andriantsimiavona and Laurence
Bourne from Biotronics3D for technical support, Christopher Pawley and Asif Jaffar
for patient recruitment, and Isha Verkaik for database management.
Proefschrift2018-new2.indb 95 13/9/18 10:06
95
Crohn disease: Semiautomatic MRI assessment
Currently, steps are being taken to further technically optimize the semiautomatic
MRI measurements and to provide full integration in viewer software. Clearly, these
aspects are essential for clinical applicability, which requires easy-to-use techniques.
In conclusion, the use of semiautomatic features for the assessment of patients with
CD maintains diagnostic and grading accuracy while improving reproducibility over
conventional activity scores. These characteristics make it potentially suitable for
therapy evaluation and monitoring of disease activity. Furthermore, accurate and
reproducible MRI scores could improve the physician’s trust in these scores to make
consistent and effective treatment decisions.
ACKNOWLEDGMENTS
The authors thank Ernst Harting for the management of the VIGOR++ project and
Costis Kompis for project exploitation, Rado Andriantsimiavona and Laurence
Bourne from Biotronics3D for technical support, Christopher Pawley and Asif Jaffar
for patient recruitment, and Isha Verkaik for database management.
Proefschrift2018-new2.indb 95 13/9/18 10:06
96
Chapter 4
REFERENCES1. Panes J, Bouhnik Y, Reinisch W, et al. Imaging techniques for assessment of inflammatory
bowel disease: joint ECCO and ESGAR evidencebased consensus guidelines. J Crohns
Colitis 2013; 7:556–585. http://dx.doi.org/10.1016/j.crohns.2013.02.020.
2. Zappa M, Stefanescu C, Cazals-Hatem D, et al. Which magnetic resonance imaging
findings accurately evaluate inflammation in small bowel Crohn’s disease? A retrospective
comparison with surgical pathologic analysis. Inflamm Bowel Dis 2011; 17:984–993.
3. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment of
disease activity and severity in ileocolonic Crohn’s disease. Gut 2009; 58:1113–1120.
4. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: Derivation and histopathological validation of an MR-
based activity index. Eur J Radiol 2012;81:2080–2088.
5. Rimola J, Ordás I, Rodriguez S, et al. Magnetic resonance imaging for evaluation of Crohn’s
disease: validation of parameters of severity and quantitative index of activity. Inflamm
Bowel Dis 2011; 17:1759–1768.
6. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading crohn disease activity with MRI:
Interobserver variability of MRI features, MRI scoring of severity, and correlation with
crohn disease endoscopic index of severity. Am. J. Roentgenol. 2013; 201:1220–1228.
7. Ziech MLW, Bipat S, Roelofs JJTH, et al. Retrospective comparison of magnetic resonance
imaging features and histopathology in Crohn’s disease patients. Eur J Radiol 2011;
80:e299–e305. http://dx.doi.org/10.1016/j.ejrad.2010.12.075.
8. Tielbeek JAW, Vos FM, Stoker J. A computer-assisted model for detection of MRI signs of
Crohn’s disease activity: Future or fiction? Abdom Imaging 2012; 37:967–973.
9. Naziroglu RE, Puylaert CAJ, Tielbeek JAW, et al. Semi-automatic bowel wall thickness
measurements on MR enterography in patients with Crohn’s disease. Br J Radiol 2017;
20160654.
10. Li Z, Tielbeek JAW, Caan MWA, et al. Expiration-Phase Template-Based Motion Correction
of Free-Breathing Abdominal Dynamic Contrast Enhanced MRI. IEEE Trans Biomed Eng
2015; 62:1215–1225.
11. Schüffler PJ, Mahapatra D, Tielbeek JAW, et al. A Model Development Pipeline for Crohn’s
Disease Severity Assessment from Magnetic Resonance Images. Abdom Imaging Comput
Clin Appl 2013; 8198:1–10.
12. Harvey RF, Bradshaw JM. A simple index of Crohn’s-disease activity. Lancet 1980; 1:514.
13. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Thérapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut 1989; 30:983–989.
14. Li Z, Mahapatra D, Tielbeek J, et al. Image registration based on autocorrelation of local
structure. IEEE Trans Med Imaging 2015; 35:1.
15. Rothery P. A nonparametric measure of intraclass correlation. Biometrika 1979; 66:629–
639.
16. van Ierssel SH, Van Craenenbroeck EM, Conraads VM, et al. Flow cytometric detection
of endothelial microparticles (EMP): effects of centrifugation and storage alter with the
phenotype studied. Thromb Res 2010; 125:332–339.
17. Vuillemin A, Oppert JM, Guillemin F, et al. Self-administered questionnaire compared with
interview to assess past-year physical activity. Med Sci Sports Exerc 2000; 32(July):1119–
1124.
Proefschrift2018-new2.indb 96 13/9/18 10:06
96
Chapter 4
REFERENCES1. Panes J, Bouhnik Y, Reinisch W, et al. Imaging techniques for assessment of inflammatory
bowel disease: joint ECCO and ESGAR evidencebased consensus guidelines. J Crohns
Colitis 2013; 7:556–585. http://dx.doi.org/10.1016/j.crohns.2013.02.020.
2. Zappa M, Stefanescu C, Cazals-Hatem D, et al. Which magnetic resonance imaging
findings accurately evaluate inflammation in small bowel Crohn’s disease? A retrospective
comparison with surgical pathologic analysis. Inflamm Bowel Dis 2011; 17:984–993.
3. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment of
disease activity and severity in ileocolonic Crohn’s disease. Gut 2009; 58:1113–1120.
4. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: Derivation and histopathological validation of an MR-
based activity index. Eur J Radiol 2012;81:2080–2088.
5. Rimola J, Ordás I, Rodriguez S, et al. Magnetic resonance imaging for evaluation of Crohn’s
disease: validation of parameters of severity and quantitative index of activity. Inflamm
Bowel Dis 2011; 17:1759–1768.
6. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading crohn disease activity with MRI:
Interobserver variability of MRI features, MRI scoring of severity, and correlation with
crohn disease endoscopic index of severity. Am. J. Roentgenol. 2013; 201:1220–1228.
7. Ziech MLW, Bipat S, Roelofs JJTH, et al. Retrospective comparison of magnetic resonance
imaging features and histopathology in Crohn’s disease patients. Eur J Radiol 2011;
80:e299–e305. http://dx.doi.org/10.1016/j.ejrad.2010.12.075.
8. Tielbeek JAW, Vos FM, Stoker J. A computer-assisted model for detection of MRI signs of
Crohn’s disease activity: Future or fiction? Abdom Imaging 2012; 37:967–973.
9. Naziroglu RE, Puylaert CAJ, Tielbeek JAW, et al. Semi-automatic bowel wall thickness
measurements on MR enterography in patients with Crohn’s disease. Br J Radiol 2017;
20160654.
10. Li Z, Tielbeek JAW, Caan MWA, et al. Expiration-Phase Template-Based Motion Correction
of Free-Breathing Abdominal Dynamic Contrast Enhanced MRI. IEEE Trans Biomed Eng
2015; 62:1215–1225.
11. Schüffler PJ, Mahapatra D, Tielbeek JAW, et al. A Model Development Pipeline for Crohn’s
Disease Severity Assessment from Magnetic Resonance Images. Abdom Imaging Comput
Clin Appl 2013; 8198:1–10.
12. Harvey RF, Bradshaw JM. A simple index of Crohn’s-disease activity. Lancet 1980; 1:514.
13. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Thérapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut 1989; 30:983–989.
14. Li Z, Mahapatra D, Tielbeek J, et al. Image registration based on autocorrelation of local
structure. IEEE Trans Med Imaging 2015; 35:1.
15. Rothery P. A nonparametric measure of intraclass correlation. Biometrika 1979; 66:629–
639.
16. van Ierssel SH, Van Craenenbroeck EM, Conraads VM, et al. Flow cytometric detection
of endothelial microparticles (EMP): effects of centrifugation and storage alter with the
phenotype studied. Thromb Res 2010; 125:332–339.
17. Vuillemin A, Oppert JM, Guillemin F, et al. Self-administered questionnaire compared with
interview to assess past-year physical activity. Med Sci Sports Exerc 2000; 32(July):1119–
1124.
Proefschrift2018-new2.indb 96 13/9/18 10:06
97
Crohn disease: Semiautomatic MRI assessment
18. Steiger JH. Tests for comparing elements of a correlation matrix. Psychol Bull 1980;
87:245–251.
19. Landis JR, Koch GG. The measurement of observer agreement for categorical data.
Biometrics 1977; 33:159–174.
20. R Core Team. R: a language and environment for statistical computing. 2014. ISBN
3-900051-07-0. Available at: http://www.r-project.org/. R Found. Stat. Comput.
21. Coimbra AJF, Rimola J, O’Byrne S, et al. Magnetic resonance enterography is feasible and
reliable in multicenter clinical trials in patients with Crohn’s disease, and may help select
subjects with active inflammation. Aliment Pharmacol Ther 2016; 43:61–72.
22. Ordás I, Rimola J, Rodríguez S, et al. Accuracy of magnetic resonance enterography in
assessing response to therapy and mucosal healing in patients with Crohn’s disease.
Gastroenterology 2014; 146:374–382, e1.
Proefschrift2018-new2.indb 97 13/9/18 10:06
97
Crohn disease: Semiautomatic MRI assessment
18. Steiger JH. Tests for comparing elements of a correlation matrix. Psychol Bull 1980;
87:245–251.
19. Landis JR, Koch GG. The measurement of observer agreement for categorical data.
Biometrics 1977; 33:159–174.
20. R Core Team. R: a language and environment for statistical computing. 2014. ISBN
3-900051-07-0. Available at: http://www.r-project.org/. R Found. Stat. Comput.
21. Coimbra AJF, Rimola J, O’Byrne S, et al. Magnetic resonance enterography is feasible and
reliable in multicenter clinical trials in patients with Crohn’s disease, and may help select
subjects with active inflammation. Aliment Pharmacol Ther 2016; 43:61–72.
22. Ordás I, Rimola J, Rodríguez S, et al. Accuracy of magnetic resonance enterography in
assessing response to therapy and mucosal healing in patients with Crohn’s disease.
Gastroenterology 2014; 146:374–382, e1.
Proefschrift2018-new2.indb 97 13/9/18 10:06
98
Chapter 4
APPENDIX A: SUPPLEMENTAL METHODS
Model development
For development of the scoring systems, an independent cohort was used, consisting
of 27 patients with known Crohn’s disease undergoing MR enterography (MRE)
and ileocolonoscopy (with segmental CDEIS scoring) within four weeks. Prior to
MRE, a standardized small bowel preparation was used consisting of 4 hours fasting
and 1600 mL 2.5% Mannitol solution ingested over 1 hour before the scan. This
cohort was recruited for a previous study [1]. Three patients were excluded from the
original cohort, because no informed consent could be obtained for future research.
Scans from the development cohort were all individually evaluated by four observers
(C.Y.N., D.P., J.S., J.M.) resulting in four evaluations per dataset [1]. All readers were
unaware of the findings at the initial reading (> 1 year before study reading for all
cases) and the findings from ileocolonoscopy, but were aware of patients’ surgical
history. MRI examinations from were evaluated using online viewer software (3Dnet
Suite, Biotronics3D, London, UK). Features common to three previously validated
MRI activity scores (MaRIA, London score and CDMI) were evaluated in all segments
of the dataset and included in the selection process for model development. By
reducing the number of features to include only the most essential, the potential
validity of the developed model is increased. The included features comprised three
categories: 1. mural thickness, 2. contrast enhancement (either subjectively graded
or quantified using RCE) and 3. T2 mural signal intensity (classified in the MaRIA as
mural edema) (see Score calculation, Appendix A). Additional features, for example
perimural T2 signal and ulceration, were not included as they are not common to
all MRI activity scores. Semi-automatic measurements – maximum and mean bowel
wall thickness, excess bowel wall volume and the initial slope of increase (ISI) – were
calculated for all evaluated segments.
These features have been scored by four radiologists independently [1]. All samples
of the four readers were used for model development, without averaging over the
readers, since our model was intended to be applied on single readers’ outcomes.
All generalized linear regression models have been trained using R statistical
language (v3.1.2) [2].
Proefschrift2018-new2.indb 98 13/9/18 10:06
98
Chapter 4
APPENDIX A: SUPPLEMENTAL METHODS
Model development
For development of the scoring systems, an independent cohort was used, consisting
of 27 patients with known Crohn’s disease undergoing MR enterography (MRE)
and ileocolonoscopy (with segmental CDEIS scoring) within four weeks. Prior to
MRE, a standardized small bowel preparation was used consisting of 4 hours fasting
and 1600 mL 2.5% Mannitol solution ingested over 1 hour before the scan. This
cohort was recruited for a previous study [1]. Three patients were excluded from the
original cohort, because no informed consent could be obtained for future research.
Scans from the development cohort were all individually evaluated by four observers
(C.Y.N., D.P., J.S., J.M.) resulting in four evaluations per dataset [1]. All readers were
unaware of the findings at the initial reading (> 1 year before study reading for all
cases) and the findings from ileocolonoscopy, but were aware of patients’ surgical
history. MRI examinations from were evaluated using online viewer software (3Dnet
Suite, Biotronics3D, London, UK). Features common to three previously validated
MRI activity scores (MaRIA, London score and CDMI) were evaluated in all segments
of the dataset and included in the selection process for model development. By
reducing the number of features to include only the most essential, the potential
validity of the developed model is increased. The included features comprised three
categories: 1. mural thickness, 2. contrast enhancement (either subjectively graded
or quantified using RCE) and 3. T2 mural signal intensity (classified in the MaRIA as
mural edema) (see Score calculation, Appendix A). Additional features, for example
perimural T2 signal and ulceration, were not included as they are not common to
all MRI activity scores. Semi-automatic measurements – maximum and mean bowel
wall thickness, excess bowel wall volume and the initial slope of increase (ISI) – were
calculated for all evaluated segments.
These features have been scored by four radiologists independently [1]. All samples
of the four readers were used for model development, without averaging over the
readers, since our model was intended to be applied on single readers’ outcomes.
All generalized linear regression models have been trained using R statistical
language (v3.1.2) [2].
Proefschrift2018-new2.indb 98 13/9/18 10:06
99
Crohn disease: Semiautomatic MRI assessment
Two models were developed based on the previously mentioned three categories
(mural thickness, contrast enhancement, T2 mural signal). For the first model,
semi-automatic wall thickness and contrast enhancement parameters were
included in the development process. For the second model, the semi-automatic
measurements were excluded, relying only on subjective radiologist scores alone
for mural thickness, contrast enhancement and T2 mural signal intensity (See MRI
features and grading categories, Appendix A).
From both the semi-automatic and the subjective models the ‘best’ model was
selected using a previously described exhaustive search method for biomarker
discovery [3]. In summary, this method evaluated all possible combinations of MRI
features as candidate models for predicting CDEIS, under the above constraint of
having at least one feature per category. Specifically, the rank correlation to CDEIS
of each putative model was determined in the retrospective data using a 50-fold
bootstrap cross-validation [4]. Eventually, this procedure delivered two models: the
top ranking semi-automatic model and the top ranking subjective model. These
were termed the “VIGOR score” and the “subjective score”, respectively.
MRI protocol
Plane Slice thick-ness (mm)
FOV TR (ms) TE (ms) Flip angle
Balanced GE Coronal 5 380x380 2.5 1.25 60
BTFE dynamic Coronal 10 380x380 2-2.1 1 45
T2-SSFSE Coronal 4 380x380 628-660 60 90
T2-SSFSE Axial 4 400x400 759 119 90
T2-w SSFSE fat saturation Axial 7 380x380 967-1314 50 90
DCE sequence Coronal 2.5 380x380-439
2.9 1.8 15
3D T1-w SPGE fat satu-ration
Coronal 2 380x380-459
2.2-2.4 1.0-1.1 10
3D T1-w SPGE fat satu-ration
Axial 2 380x380 2.1-2.3 1.0-1.1 10
BTFE, balanced turbo field-echo; DCE, dynamic contrast enhanced; FOV, field of view; GE, gradient echo; SPGE, spoiled gradient-echo; SSFSE, single-shot fast spin echo; TE, echo time; TR, repetition time.
The DCE sequence consisted of 300 consecutive volumetric acquisitions at a
temporal resolution of 1.2 seconds/volume. Intravenous gadolinium contrast
was administered 60 seconds after the start of the DCE sequence block using
the standard contrast agent in the participating centres: gadobutrol (Gadovist
Proefschrift2018-new2.indb 99 13/9/18 10:06
99
Crohn disease: Semiautomatic MRI assessment
Two models were developed based on the previously mentioned three categories
(mural thickness, contrast enhancement, T2 mural signal). For the first model,
semi-automatic wall thickness and contrast enhancement parameters were
included in the development process. For the second model, the semi-automatic
measurements were excluded, relying only on subjective radiologist scores alone
for mural thickness, contrast enhancement and T2 mural signal intensity (See MRI
features and grading categories, Appendix A).
From both the semi-automatic and the subjective models the ‘best’ model was
selected using a previously described exhaustive search method for biomarker
discovery [3]. In summary, this method evaluated all possible combinations of MRI
features as candidate models for predicting CDEIS, under the above constraint of
having at least one feature per category. Specifically, the rank correlation to CDEIS
of each putative model was determined in the retrospective data using a 50-fold
bootstrap cross-validation [4]. Eventually, this procedure delivered two models: the
top ranking semi-automatic model and the top ranking subjective model. These
were termed the “VIGOR score” and the “subjective score”, respectively.
MRI protocol
Plane Slice thick-ness (mm)
FOV TR (ms) TE (ms) Flip angle
Balanced GE Coronal 5 380x380 2.5 1.25 60
BTFE dynamic Coronal 10 380x380 2-2.1 1 45
T2-SSFSE Coronal 4 380x380 628-660 60 90
T2-SSFSE Axial 4 400x400 759 119 90
T2-w SSFSE fat saturation Axial 7 380x380 967-1314 50 90
DCE sequence Coronal 2.5 380x380-439
2.9 1.8 15
3D T1-w SPGE fat satu-ration
Coronal 2 380x380-459
2.2-2.4 1.0-1.1 10
3D T1-w SPGE fat satu-ration
Axial 2 380x380 2.1-2.3 1.0-1.1 10
BTFE, balanced turbo field-echo; DCE, dynamic contrast enhanced; FOV, field of view; GE, gradient echo; SPGE, spoiled gradient-echo; SSFSE, single-shot fast spin echo; TE, echo time; TR, repetition time.
The DCE sequence consisted of 300 consecutive volumetric acquisitions at a
temporal resolution of 1.2 seconds/volume. Intravenous gadolinium contrast
was administered 60 seconds after the start of the DCE sequence block using
the standard contrast agent in the participating centres: gadobutrol (Gadovist
Proefschrift2018-new2.indb 99 13/9/18 10:06
100
Chapter 4
1.0 mmol/L, Bayer Schering Pharma, Berlin, Germany) or gadoterate meglumine
(Dotarem 0.5 mmol/L, Guerbet, Paris, France). Following the DCE series, coronal
and axial 3D T1-weighted spoiled gradient-echo (SPGE) images were acquired in
the delayed phase (approximately 7 minutes after contrast injection). To reduce
bowel peristalsis, three separate doses of 10 mg intravenous butylscopolamine
bromide (Buscopan, Boehringer Ingelheim, Ingelheim, Germany) were administered
during the examination.
MRI features and grading categories
Grading score
0 1 2 3
MRI featuresLondon/CDMI
Mural thicknessa 1–3 mm > 3–5 mm > 5–7 mm > 7 mm
Mural T2 signal Equivalent to nor-mal bowel wall
Minor increase in signal-bowel wall appears dark grey on fat saturated images
Moderate increase in signal-bowel wall appears light grey on fat satu-rated images
Marked increase in signal-bowel wall contains areas of white high signal approaching that of luminal content
Perimural T2 signal
Equivalent to nor-mal mesentery
Increase in mes-enteric signal but no fluid
Small fluid rim (≤ 2 mm)
Larger fluid rim (> 2 mm)
T1 enhancement Equivalent to nor-mal bowel wall
Minor enhance-ment - bowel wall signal greater than normal small bow-el but significantly less than nearby vascular structures
Moderate enhancement - bowel wall signal increased but somewhat less than nearby vas-cular structures
Marked enhance-ment - bowel wall signal approaches that of nearby vas-cular structures
MaRIA
Mural thickness in mma
RCE
Edema Absent Present
Ulcers Absent Present
a. Measured using electronic calipersMRI=magnetic resonance imaging, RCE=relative contrast enhancement
Overall scan quality was graded on a scale from 0 (non-diagnostic images) to 3
(diagnostic images without artefacts). Subsequently, the following five bowel
segments were evaluated individually: the terminal ileum (most distal 20 cm of the
ileum), ascending colon, transverse colon, descending/sigmoid colon and rectum.
Luminal distension, defined as the percentage of adequately distended bowel for
diagnostic evaluation, was graded for each segment from 0 to 4 (< 20%, 20–40%,
40–60%, 60–80%, > 80%).
Proefschrift2018-new2.indb 100 13/9/18 10:06
100
Chapter 4
1.0 mmol/L, Bayer Schering Pharma, Berlin, Germany) or gadoterate meglumine
(Dotarem 0.5 mmol/L, Guerbet, Paris, France). Following the DCE series, coronal
and axial 3D T1-weighted spoiled gradient-echo (SPGE) images were acquired in
the delayed phase (approximately 7 minutes after contrast injection). To reduce
bowel peristalsis, three separate doses of 10 mg intravenous butylscopolamine
bromide (Buscopan, Boehringer Ingelheim, Ingelheim, Germany) were administered
during the examination.
MRI features and grading categories
Grading score
0 1 2 3
MRI featuresLondon/CDMI
Mural thicknessa 1–3 mm > 3–5 mm > 5–7 mm > 7 mm
Mural T2 signal Equivalent to nor-mal bowel wall
Minor increase in signal-bowel wall appears dark grey on fat saturated images
Moderate increase in signal-bowel wall appears light grey on fat satu-rated images
Marked increase in signal-bowel wall contains areas of white high signal approaching that of luminal content
Perimural T2 signal
Equivalent to nor-mal mesentery
Increase in mes-enteric signal but no fluid
Small fluid rim (≤ 2 mm)
Larger fluid rim (> 2 mm)
T1 enhancement Equivalent to nor-mal bowel wall
Minor enhance-ment - bowel wall signal greater than normal small bow-el but significantly less than nearby vascular structures
Moderate enhancement - bowel wall signal increased but somewhat less than nearby vas-cular structures
Marked enhance-ment - bowel wall signal approaches that of nearby vas-cular structures
MaRIA
Mural thickness in mma
RCE
Edema Absent Present
Ulcers Absent Present
a. Measured using electronic calipersMRI=magnetic resonance imaging, RCE=relative contrast enhancement
Overall scan quality was graded on a scale from 0 (non-diagnostic images) to 3
(diagnostic images without artefacts). Subsequently, the following five bowel
segments were evaluated individually: the terminal ileum (most distal 20 cm of the
ileum), ascending colon, transverse colon, descending/sigmoid colon and rectum.
Luminal distension, defined as the percentage of adequately distended bowel for
diagnostic evaluation, was graded for each segment from 0 to 4 (< 20%, 20–40%,
40–60%, 60–80%, > 80%).
Proefschrift2018-new2.indb 100 13/9/18 10:06
101
Crohn disease: Semiautomatic MRI assessment
Score calculation
Calculation of the London score, the Magnetic Resonance Index of Activity (MaRIA)
and the relative contrast enhancement (RCE) using bowel wall signal intensity (SI)
measured in a region of interest:
London score = 1.79 + 1.34 x Wall thickness + 0.94 x mural T2 signal
CDMI = Wall thickness + T1 enhancement + mural T2 signal + perimural T2 signal
MaRIA = 1.5 x Wall thickness (mm) + 0.02 x RCE + 5 x oedema + 10 x ulceration
RCE = (SI postcontrast - SI precontrast)/(SI precontrast)
RCE calculation did not include a noise correction term, as was used by Rimola et
al [5], since inconsistent noise measurements were observed in our data, yielding
arbitrary RCE values. Signal intensity values were corrected using the method
described by Chenevert et al [6].
Excess bowel wall volume feature
The excess bowel wall volume was defined as the volume of the delineated region
exceeding normal thickness. Normal thickness was calculated as the mean automatic
thickness of healthy segments (no activity on MRI/endoscopy) in the development
cohort.
Sample size calculation for subset of patients
Employing an α of 0.05 and a β of 0.20, expected colonic sensitivity of 0.4 and
prevalence of 0.15, expected terminal ileum specificity of 0.8 and prevalence of 0.67,
the necessary number of terminal ileum and colonic segments was calculated to be
45 and 154 segments, respectively. Anticipating a segment exclusion of 10%, a total
of 50 patient datasets were required.
Diagnostic accuracy
Diagnostic accuracy for segmental disease activity (defined as a CDEIS ≥3 [7])
was assessed by applying segmental MRI scores to all bowel segments of the
Proefschrift2018-new2.indb 101 13/9/18 10:06
101
Crohn disease: Semiautomatic MRI assessment
Score calculation
Calculation of the London score, the Magnetic Resonance Index of Activity (MaRIA)
and the relative contrast enhancement (RCE) using bowel wall signal intensity (SI)
measured in a region of interest:
London score = 1.79 + 1.34 x Wall thickness + 0.94 x mural T2 signal
CDMI = Wall thickness + T1 enhancement + mural T2 signal + perimural T2 signal
MaRIA = 1.5 x Wall thickness (mm) + 0.02 x RCE + 5 x oedema + 10 x ulceration
RCE = (SI postcontrast - SI precontrast)/(SI precontrast)
RCE calculation did not include a noise correction term, as was used by Rimola et
al [5], since inconsistent noise measurements were observed in our data, yielding
arbitrary RCE values. Signal intensity values were corrected using the method
described by Chenevert et al [6].
Excess bowel wall volume feature
The excess bowel wall volume was defined as the volume of the delineated region
exceeding normal thickness. Normal thickness was calculated as the mean automatic
thickness of healthy segments (no activity on MRI/endoscopy) in the development
cohort.
Sample size calculation for subset of patients
Employing an α of 0.05 and a β of 0.20, expected colonic sensitivity of 0.4 and
prevalence of 0.15, expected terminal ileum specificity of 0.8 and prevalence of 0.67,
the necessary number of terminal ileum and colonic segments was calculated to be
45 and 154 segments, respectively. Anticipating a segment exclusion of 10%, a total
of 50 patient datasets were required.
Diagnostic accuracy
Diagnostic accuracy for segmental disease activity (defined as a CDEIS ≥3 [7])
was assessed by applying segmental MRI scores to all bowel segments of the
Proefschrift2018-new2.indb 101 13/9/18 10:06
102
Chapter 4
50 randomly selected patients. For evaluation of diagnostic accuracy, segmental
disease activity on MRI was defined using these predetermined cut-off values:
MaRIA, ≥7; London score, ≥4.1; CDMI, ≥3 [5,8]. For the VIGOR and subjective scores,
the optimal cut-off points for detection of active disease were determined using
receiver-operating characteristics (ROC) analyses performed on the development
cohorts’ datasets. Sensitivity, specificity, positive predictive value, negative
predictive value and diagnostic accuracy were then calculated for all segments of
the prospective subset.
Per-patient analysis
For the per-patient analysis, MRI activity scores and global CDEIS in the subset
were calculated as the sum of segmental scores divided by the number of evaluated
segments. A stenosis score was added to the per-patient CDEIS score if applicable
[9]. Subsequently, MRI scores (per-patient and per-segment) were correlated to
CDEIS and interobserver agreement was determined in all segments using the
conventional ICC.
APPENDIX B: SUPPLEMENTAL RESULTS
Development cohorts’ segment in- and exclusions
The retrospective development cohort consisted of 27 known Crohn’s disease
patients (127 segments evaluable by radiologist and endoscopist). Eighteen
segments (6 colon, 12 rectum) were excluded from the analysis, due to severe
artefacts (n=4), poor distension (n=7) and fecal residue (n=7). A further 42 segments
were excluded, as semi-automatic features could not be derived in these segments
for the following reasons: segment outside the DCE field-of-view (33/42), failed
DCE registration (8/42) or failed segmentation (1/42). Of the 33 segments outside
the DCE field-of-view, 91% were either colonic (16/33) or rectal (14/33), which was
expected for this retrospective cohort, as MRI preparation and sequences were not
intended for colonic evaluation. As such, 67 segments remained.
Proefschrift2018-new2.indb 102 13/9/18 10:06
102
Chapter 4
50 randomly selected patients. For evaluation of diagnostic accuracy, segmental
disease activity on MRI was defined using these predetermined cut-off values:
MaRIA, ≥7; London score, ≥4.1; CDMI, ≥3 [5,8]. For the VIGOR and subjective scores,
the optimal cut-off points for detection of active disease were determined using
receiver-operating characteristics (ROC) analyses performed on the development
cohorts’ datasets. Sensitivity, specificity, positive predictive value, negative
predictive value and diagnostic accuracy were then calculated for all segments of
the prospective subset.
Per-patient analysis
For the per-patient analysis, MRI activity scores and global CDEIS in the subset
were calculated as the sum of segmental scores divided by the number of evaluated
segments. A stenosis score was added to the per-patient CDEIS score if applicable
[9]. Subsequently, MRI scores (per-patient and per-segment) were correlated to
CDEIS and interobserver agreement was determined in all segments using the
conventional ICC.
APPENDIX B: SUPPLEMENTAL RESULTS
Development cohorts’ segment in- and exclusions
The retrospective development cohort consisted of 27 known Crohn’s disease
patients (127 segments evaluable by radiologist and endoscopist). Eighteen
segments (6 colon, 12 rectum) were excluded from the analysis, due to severe
artefacts (n=4), poor distension (n=7) and fecal residue (n=7). A further 42 segments
were excluded, as semi-automatic features could not be derived in these segments
for the following reasons: segment outside the DCE field-of-view (33/42), failed
DCE registration (8/42) or failed segmentation (1/42). Of the 33 segments outside
the DCE field-of-view, 91% were either colonic (16/33) or rectal (14/33), which was
expected for this retrospective cohort, as MRI preparation and sequences were not
intended for colonic evaluation. As such, 67 segments remained.
Proefschrift2018-new2.indb 102 13/9/18 10:06
103
Crohn disease: Semiautomatic MRI assessment
Clinical characteristics of the subset group from the prospective cohort
Total no. of patients 50
Female, n (%) 29 (58)
Age at MRI (years), median (IQR) 32 (27–47)
Previous surgery, n (%) 19 (38)
Concomitant treatments
Anti-TNF antibodies, n (%) 15 (30)
Steroids, n. (%) of patients 8 (16)
Thiopurines, no. (%) 8 (16)
5-ASA, no. (%) of patients 10 (20)
Methotrexate, no. (%) 3 (6)
CRP (mg/L), median (IQR) 4 (2–11)
HBI, median (IQR) 5 (2–8)
CDEIS, median (IQR) 4.0 (0.1–7.7)
Montreal classification
Age at diagnosis (years), median (IQR) 22 (19–28)
Disease location
L1 ileal, n (%) 21 (42)
L2 colonic, n (%) 9 (18)
L3 ileocolonic, n (%) 20 (40)
L4 upper GI tract involvement, n (%) 2 (4)
Disease behaviour
B1 inflammatory 25 (50)
B2 stricturing 16 (32)
B3 penetrating 9 (18)
Perianal involvement, n (%) 9 (18)
5-ASA, 5-acetylsalicylic acid; CDEIS, Crohn’s disease Endoscopic Index of Severity; CRP, C-reactive protein; GI, gastrointestinal; HBI, Harvey-Bradshaw Index; IQR, interquartile range; MRE, magnetic resonance enterography; TNF, tumour necrosis factor.
Proefschrift2018-new2.indb 103 13/9/18 10:06
103
Crohn disease: Semiautomatic MRI assessment
Clinical characteristics of the subset group from the prospective cohort
Total no. of patients 50
Female, n (%) 29 (58)
Age at MRI (years), median (IQR) 32 (27–47)
Previous surgery, n (%) 19 (38)
Concomitant treatments
Anti-TNF antibodies, n (%) 15 (30)
Steroids, n. (%) of patients 8 (16)
Thiopurines, no. (%) 8 (16)
5-ASA, no. (%) of patients 10 (20)
Methotrexate, no. (%) 3 (6)
CRP (mg/L), median (IQR) 4 (2–11)
HBI, median (IQR) 5 (2–8)
CDEIS, median (IQR) 4.0 (0.1–7.7)
Montreal classification
Age at diagnosis (years), median (IQR) 22 (19–28)
Disease location
L1 ileal, n (%) 21 (42)
L2 colonic, n (%) 9 (18)
L3 ileocolonic, n (%) 20 (40)
L4 upper GI tract involvement, n (%) 2 (4)
Disease behaviour
B1 inflammatory 25 (50)
B2 stricturing 16 (32)
B3 penetrating 9 (18)
Perianal involvement, n (%) 9 (18)
5-ASA, 5-acetylsalicylic acid; CDEIS, Crohn’s disease Endoscopic Index of Severity; CRP, C-reactive protein; GI, gastrointestinal; HBI, Harvey-Bradshaw Index; IQR, interquartile range; MRE, magnetic resonance enterography; TNF, tumour necrosis factor.
Proefschrift2018-new2.indb 103 13/9/18 10:06
104
Chapter 4
Interobserver scatter plots. Scatter plots for MRI activity scores between observer 1 (y-axis)
and observer 2 (x-axis). Active (overlapping; active for both observers) segments of the full
prospective cohort are shown in the left figures, while all (overlapping; included for both
observers) segments (active and remission) of the 50-patient subset are shown in the figures
on the right.
Proefschrift2018-new2.indb 104 13/9/18 10:06
104
Chapter 4
Interobserver scatter plots. Scatter plots for MRI activity scores between observer 1 (y-axis)
and observer 2 (x-axis). Active (overlapping; active for both observers) segments of the full
prospective cohort are shown in the left figures, while all (overlapping; included for both
observers) segments (active and remission) of the 50-patient subset are shown in the figures
on the right.
Proefschrift2018-new2.indb 104 13/9/18 10:06
105
Crohn disease: Semiautomatic MRI assessment
Proefschrift2018-new2.indb 105 13/9/18 10:06
105
Crohn disease: Semiautomatic MRI assessment
Proefschrift2018-new2.indb 105 13/9/18 10:06
106
Chapter 4
REFERENCES1. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading crohn disease activity with MRI:
Interobserver variability of MRI features, MRI scoring of severity, and correlation with
crohn disease endoscopic index of severity. Am. J. Roentgenol. 2013;201(6):1220–1228.
2. R Core Team. R: A Language and Environment for Statistical Computing. R Found. Stat.
Comput. 2014:{ISBN} 3-900051-07-0, http://www.R-project.org. Available at: http://
www.r-project.org/.
3. Schüffler PJ, Mahapatra D, Tielbeek JAW, et al. A Model Development Pipeline for Crohn’s
Disease Severity Assessment from Magnetic Resonance Images. Abdom. Imaging.
Comput. Clin. Appl. 2013;8198:1–10.
4. Hastie T, Tibshirani R, Friedman J. The Elements of Statistical Learning. Springer 2001.
2009;18(4):746.
5. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment of
disease activity and severity in ileocolonic Crohn’s disease. Gut. 2009;58:1113–1120.
6. Chenevert TL, Malyarenko DI, Newitt D, et al. Errors in Quantitative Image Analysis due to
Platform-Dependent Image Scaling. Transl. Oncol. 2014;7(1):65–71.
7. Daperno M, Castiglione F, de Ridder L, et al. Results of the 2nd part Scientific Workshop
of the ECCO (II): Measures and markers of prediction to achieve, detect, and monitor
intestinal healing in Inflammatory Bowel Disease. J. Crohn’s Colitis. 2011;5(5):484–498.
8. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: Derivation and histopathological validation of an MR-
based activity index. Eur. J. Radiol. 2012;81(9):2080–2088.
9. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Thérapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut. 1989;30(7):983–989.
Proefschrift2018-new2.indb 106 13/9/18 10:06
106
Chapter 4
REFERENCES1. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading crohn disease activity with MRI:
Interobserver variability of MRI features, MRI scoring of severity, and correlation with
crohn disease endoscopic index of severity. Am. J. Roentgenol. 2013;201(6):1220–1228.
2. R Core Team. R: A Language and Environment for Statistical Computing. R Found. Stat.
Comput. 2014:{ISBN} 3-900051-07-0, http://www.R-project.org. Available at: http://
www.r-project.org/.
3. Schüffler PJ, Mahapatra D, Tielbeek JAW, et al. A Model Development Pipeline for Crohn’s
Disease Severity Assessment from Magnetic Resonance Images. Abdom. Imaging.
Comput. Clin. Appl. 2013;8198:1–10.
4. Hastie T, Tibshirani R, Friedman J. The Elements of Statistical Learning. Springer 2001.
2009;18(4):746.
5. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment of
disease activity and severity in ileocolonic Crohn’s disease. Gut. 2009;58:1113–1120.
6. Chenevert TL, Malyarenko DI, Newitt D, et al. Errors in Quantitative Image Analysis due to
Platform-Dependent Image Scaling. Transl. Oncol. 2014;7(1):65–71.
7. Daperno M, Castiglione F, de Ridder L, et al. Results of the 2nd part Scientific Workshop
of the ECCO (II): Measures and markers of prediction to achieve, detect, and monitor
intestinal healing in Inflammatory Bowel Disease. J. Crohn’s Colitis. 2011;5(5):484–498.
8. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: Derivation and histopathological validation of an MR-
based activity index. Eur. J. Radiol. 2012;81(9):2080–2088.
9. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Thérapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut. 1989;30(7):983–989.
Proefschrift2018-new2.indb 106 13/9/18 10:06
Proefschrift2018-new2.indb 107 13/9/18 10:06 Proefschrift2018-new2.indb 107 13/9/18 10:06
Proefschrift2018-new2.indb 108 13/9/18 10:06 Proefschrift2018-new2.indb 108 13/9/18 10:06
CHAPTER 5
Comparison of contrast-enhanced and diffusion-weighted MRI in assessment of the terminal ileum in Crohn’s disease patients
Carl. A.J. Puylaert, Jeroen A.W. Tielbeek, Peter J. Schüffler, C. Yung Nio,
Karin Horsthuis, Banafsche Mearadji, Cyriel Y. Ponsioen, Frans M. Vos, Jaap Stoker
Proefschrift2018-new2.indb 109 13/9/18 10:06
CHAPTER 5
Comparison of contrast-enhanced and diffusion-weighted MRI in assessment of the terminal ileum in Crohn’s disease patients
Carl. A.J. Puylaert, Jeroen A.W. Tielbeek, Peter J. Schüffler, C. Yung Nio,
Karin Horsthuis, Banafsche Mearadji, Cyriel Y. Ponsioen, Frans M. Vos, Jaap Stoker
Proefschrift2018-new2.indb 109 13/9/18 10:06
110
Chapter 5
ABSTRACT
Objectives
To compare the performance of contrast-enhanced (CE)-MRI and diffusion-weighted
imaging (DW)-MRI in grading Crohn’s disease activity of the terminal ileum.
Materials and methods
Three readers evaluated CE-MRI, DW-MRI and their combinations (CE/DW-
MRI and DW/CE-MRI, depending on which protocol was used at the start of
evaluation). Disease severity grading scores were correlated to the Crohn’s Disease
Endoscopic Index of Severity (CDEIS). Diagnostic accuracy, severity grading and
levels of confidence were compared between imaging protocols and interobserver
agreement was calculated.
Results
Sixty-one patients were included (30 female, median age 36). Diagnostic accuracy
for active disease for CE-MRI, DW-MRI, CE/DW-MRI and DW/CE-MRI ranged
between 0.82–0.85, 0.75–0.83, 0.79–0.84 and 0.74–0.82, respectively. Severity
grading correlation to CDEIS ranged between 0.70–0.74, 0.66–0.70, 0.69–0.75
and 0.67–0.74, respectively. For each reader, CE-MRI values were consistently
higher than DW-MRI, albeit not significantly. Confidence levels for all readers were
significantly higher for CE-MRI compared to DW-MRI (P<0.001). Further increased
confidence was seen when using combined imaging protocols.
Conclusions
There was no significant difference of CE-MRI and DW-MRI in determining disease
activity, but the higher confidence levels may favour CE-MRI. DW-MRI is a good
alternative in cases with relative contraindications for the use of intravenous
contrast medium.
Proefschrift2018-new2.indb 110 13/9/18 10:06
110
Chapter 5
ABSTRACT
Objectives
To compare the performance of contrast-enhanced (CE)-MRI and diffusion-weighted
imaging (DW)-MRI in grading Crohn’s disease activity of the terminal ileum.
Materials and methods
Three readers evaluated CE-MRI, DW-MRI and their combinations (CE/DW-
MRI and DW/CE-MRI, depending on which protocol was used at the start of
evaluation). Disease severity grading scores were correlated to the Crohn’s Disease
Endoscopic Index of Severity (CDEIS). Diagnostic accuracy, severity grading and
levels of confidence were compared between imaging protocols and interobserver
agreement was calculated.
Results
Sixty-one patients were included (30 female, median age 36). Diagnostic accuracy
for active disease for CE-MRI, DW-MRI, CE/DW-MRI and DW/CE-MRI ranged
between 0.82–0.85, 0.75–0.83, 0.79–0.84 and 0.74–0.82, respectively. Severity
grading correlation to CDEIS ranged between 0.70–0.74, 0.66–0.70, 0.69–0.75
and 0.67–0.74, respectively. For each reader, CE-MRI values were consistently
higher than DW-MRI, albeit not significantly. Confidence levels for all readers were
significantly higher for CE-MRI compared to DW-MRI (P<0.001). Further increased
confidence was seen when using combined imaging protocols.
Conclusions
There was no significant difference of CE-MRI and DW-MRI in determining disease
activity, but the higher confidence levels may favour CE-MRI. DW-MRI is a good
alternative in cases with relative contraindications for the use of intravenous
contrast medium.
Proefschrift2018-new2.indb 110 13/9/18 10:06
111
Comparison of CE-MRI and DW-MRI
INTRODUCTION
Magnetic resonance imaging (MRI) has gained a strong role in evaluation of luminal
Crohn’s disease and is the preferred modality for evaluation of small bowel disease
[1]. The terminal ileum is the most common location of small bowel Crohn’s disease
and can be visualized by both MRI and ileocolonoscopy. A typical MRI protocol
for Crohn’s disease evaluation includes a non-enhanced T2-weighted sequence
with fat-suppression and T1-weighted sequences before and after intravenous
administration of a gadolinium chelated contrast agent [2,3]. MRI features such as
the degree and pattern of bowel wall enhancement after intravenous contrast have
shown to be linked to inflammation as assessed by endoscopic and histopathologic
reference standards [4–6].
Diffusion weighted (DW)-MRI has been used for large organs, such as the brain and
liver, where quantitative measurements can be made with relative ease. However,
the sensitivity of DW-MRI to motion artefacts has limited its applications for small
bowel diseases [7]. Despite these difficulties, technical improvements and recent
positive results of DW-MRI have encouraged new investigations into small bowel
applications [8].
A recent study by Kim et al. found that the addition of DW sequences to contrast-
enhanced (CE-)MRI did not provide a substantial benefit in terms of diagnostic
accuracy. Although sensitivity was increased (62% to 83%), the added detection
concerned mainly mild disease with doubtful clinical relevance, while a decrease
of specificity was seen (94% to 60%) [9]. However, a different study by Qi et al.,
using capsule endoscopy as their reference standard, did find an improvement of
diagnostic accuracy (79% to 92%) when DWI was added to CE-MRI [10]. A study
by Seo et al. included 44 patients of the same cohort and focused on substitution
of CE sequences with DW sequences in a conventional MRI protocol [11]. DW-MRI
and CE-MRI showed 91.8% agreement for dichotomous classification of segments
(inflammation or no inflammation) and comparable correlation to the Crohn’s
Disease Endoscopic Index of Activity (CDEIS) (r=0.61 and 0.71, p=0.11). A study by
Schmid-Tannwald et al. included 14 patients with internal fistulas and sinus tracts
of different etiologies. They found no significant difference in the detection rate of
fistulas and sinus tracts between CE-MRI (96%) and DW-MRI (76%) [12].
Proefschrift2018-new2.indb 111 13/9/18 10:06
111
Comparison of CE-MRI and DW-MRI
INTRODUCTION
Magnetic resonance imaging (MRI) has gained a strong role in evaluation of luminal
Crohn’s disease and is the preferred modality for evaluation of small bowel disease
[1]. The terminal ileum is the most common location of small bowel Crohn’s disease
and can be visualized by both MRI and ileocolonoscopy. A typical MRI protocol
for Crohn’s disease evaluation includes a non-enhanced T2-weighted sequence
with fat-suppression and T1-weighted sequences before and after intravenous
administration of a gadolinium chelated contrast agent [2,3]. MRI features such as
the degree and pattern of bowel wall enhancement after intravenous contrast have
shown to be linked to inflammation as assessed by endoscopic and histopathologic
reference standards [4–6].
Diffusion weighted (DW)-MRI has been used for large organs, such as the brain and
liver, where quantitative measurements can be made with relative ease. However,
the sensitivity of DW-MRI to motion artefacts has limited its applications for small
bowel diseases [7]. Despite these difficulties, technical improvements and recent
positive results of DW-MRI have encouraged new investigations into small bowel
applications [8].
A recent study by Kim et al. found that the addition of DW sequences to contrast-
enhanced (CE-)MRI did not provide a substantial benefit in terms of diagnostic
accuracy. Although sensitivity was increased (62% to 83%), the added detection
concerned mainly mild disease with doubtful clinical relevance, while a decrease
of specificity was seen (94% to 60%) [9]. However, a different study by Qi et al.,
using capsule endoscopy as their reference standard, did find an improvement of
diagnostic accuracy (79% to 92%) when DWI was added to CE-MRI [10]. A study
by Seo et al. included 44 patients of the same cohort and focused on substitution
of CE sequences with DW sequences in a conventional MRI protocol [11]. DW-MRI
and CE-MRI showed 91.8% agreement for dichotomous classification of segments
(inflammation or no inflammation) and comparable correlation to the Crohn’s
Disease Endoscopic Index of Activity (CDEIS) (r=0.61 and 0.71, p=0.11). A study by
Schmid-Tannwald et al. included 14 patients with internal fistulas and sinus tracts
of different etiologies. They found no significant difference in the detection rate of
fistulas and sinus tracts between CE-MRI (96%) and DW-MRI (76%) [12].
Proefschrift2018-new2.indb 111 13/9/18 10:06
112
Chapter 5
The effort to replace the use of intravenous contrast medium is motivated by the
occurrence of side-effects, mainly nephrogenic systemic fibrosis, and the need to
avoid contrast medium in certain groups of patients, such as children and pregnant
women [13]. It should be noted that almost all reported cases of nephrogenic
systemic fibrosis have occurred while using linear gadolinium agents in patients
with end-stage kidney disease [14]. Recent research has also brought forward
concerns over gadolinium depositions in intracranial neuronal tissue, which were
found to be dose-dependent but unrelated to renal function [15]. However, the
clinical implications of these findings are yet unclear. Although there are medical
and financial motivations to reduce the use of gadolinium contrast media, the
benefits of replacement should be thoroughly investigated to justify the omission
of the well-established use of contrast enhanced sequences. For a comprehensive
comparison, different aspects of image assessment should be considered, such as
image quality and evaluability, diagnosis of active disease, severity grading and
interobserver agreement. Furthermore, different combinations of incorporated
scan sequences could lead to differences in performance and should be evaluated.
The purpose of this study was to determine the diagnostic and grading performance
of CE-MRI, DW-MRI and combined protocols, for disease activity of the terminal
ileum in Crohn’s disease patients.
MATERIALS AND METHODS
Patients
Between October 2011 and September 2014, patients with known or suspected
Crohn’s disease were prospectively recruited as part of the VIGOR++ project
(FP7/2007-2013, 270379). The full cohort has been previously published, in a
study investigating the use of semi-automatic MRI measurements in Crohn’s
disease patients [16]. For the current study, MRI examinations from a single centre
(Academic Medical Centre, Amsterdam, the Netherlands) were re-examined. Each
patient underwent MRI and ileocolonoscopy within two weeks as part of their
clinical follow-up. Patients with no endoscopic intubation of the terminal ileum
or with missing essential MRI sequences were excluded from the analysis. Ethical
permission was obtained from the hospital’s medical ethics committee and all
included patients gave written informed consent.
Proefschrift2018-new2.indb 112 13/9/18 10:06
112
Chapter 5
The effort to replace the use of intravenous contrast medium is motivated by the
occurrence of side-effects, mainly nephrogenic systemic fibrosis, and the need to
avoid contrast medium in certain groups of patients, such as children and pregnant
women [13]. It should be noted that almost all reported cases of nephrogenic
systemic fibrosis have occurred while using linear gadolinium agents in patients
with end-stage kidney disease [14]. Recent research has also brought forward
concerns over gadolinium depositions in intracranial neuronal tissue, which were
found to be dose-dependent but unrelated to renal function [15]. However, the
clinical implications of these findings are yet unclear. Although there are medical
and financial motivations to reduce the use of gadolinium contrast media, the
benefits of replacement should be thoroughly investigated to justify the omission
of the well-established use of contrast enhanced sequences. For a comprehensive
comparison, different aspects of image assessment should be considered, such as
image quality and evaluability, diagnosis of active disease, severity grading and
interobserver agreement. Furthermore, different combinations of incorporated
scan sequences could lead to differences in performance and should be evaluated.
The purpose of this study was to determine the diagnostic and grading performance
of CE-MRI, DW-MRI and combined protocols, for disease activity of the terminal
ileum in Crohn’s disease patients.
MATERIALS AND METHODS
Patients
Between October 2011 and September 2014, patients with known or suspected
Crohn’s disease were prospectively recruited as part of the VIGOR++ project
(FP7/2007-2013, 270379). The full cohort has been previously published, in a
study investigating the use of semi-automatic MRI measurements in Crohn’s
disease patients [16]. For the current study, MRI examinations from a single centre
(Academic Medical Centre, Amsterdam, the Netherlands) were re-examined. Each
patient underwent MRI and ileocolonoscopy within two weeks as part of their
clinical follow-up. Patients with no endoscopic intubation of the terminal ileum
or with missing essential MRI sequences were excluded from the analysis. Ethical
permission was obtained from the hospital’s medical ethics committee and all
included patients gave written informed consent.
Proefschrift2018-new2.indb 112 13/9/18 10:06
113
Comparison of CE-MRI and DW-MRI
MRI protocol
Patients fasted for 4 hours prior to the examination. Oral contrast medium consisted
of 2400 mL 2.5% Mannitol solution (Baxter, Utrecht, the Netherlands) split in two
doses: 800 mL (3 hours before examination) and 1600 mL (1 hour before examination).
Patients were scanned on a 3.0-Tesla MRI unit (Intera/Ingenia; Philips Healthcare,
Best, the Netherlands) in the supine position using the protocol outlined in table 1.
An axial free-breathing DWI sequence (with b-values 0, 300 and 600 s/mm2) was
used for apparent diffusion coefficient (ADC) mapping. A coronal 3D T1-weighted
spoiled gradient echo (SPGE) sequence with fat suppression was performed before
injection of intravenous gadolinium contrast. A dynamic contrast-enhanced (DCE)
coronal 3D T1-weighted SPGE was performed. Sixty seconds after the start of DCE-
MRI, 0.1 mL/kg body weight of gadobutrol (Gadovist 1.0 mmol/mL, Bayer Schering
Pharma, Berlin, Germany) was administered intravenously by bolus injection (5
ml/s). Subsequently, coronal and axial 3D T1-weighted SPGE sequences with fat
suppression were performed in the delayed phase (± 7 minutes after injection). To
achieve spasmolysis, 10 mg of butylscopolamine bromide (Buscopan, Boehringer
Ingelheim, Germany) was administered intravenously three times at even intervals
during the examination (Table 1).
Table 1. Protocol for MRI acquisition
Plane Fat saturation
Slice thick-ness (mm)
TR/TE (ms)
Flip an-gle (°)
FOV (mm) Matrix
Balanced GE Coronal No 5 2.5/1.25 60 380x380
T2-w SSFSE Coronal No 4 633/60 90 380x380 384x384
T2-w SSFSE Axial No 4 759/119 90 400x400 528x528
T2-w SSFSE Axial Yes 7 1088/50 90 380x302 320x256
DWI Axial Yes 7 2972/60 90 380x260 288x196
3D T1-w SPGE Coronal Yes 2 2.39/1.07 10 380x380 192x192
3D T1-w SPGE Axial Yes 2 2.27/1.09 10 380x380 384x384
FFE Fast Field Echo, HASTE Half-Fourier Acquisition Single-Shot Turbo Spin-Echo, THRIVE T1 High-Resolution Isotropic Volume Excitation
Ileocolonoscopy
Ileocolonoscopy was performed using standard bowel preparation and equipment
(model CF-160L, Olympus) within two weeks of the MRI examination by either a
gastroenterologist or senior resident under direct supervision of a gastroenterologist.
Proefschrift2018-new2.indb 113 13/9/18 10:06
113
Comparison of CE-MRI and DW-MRI
MRI protocol
Patients fasted for 4 hours prior to the examination. Oral contrast medium consisted
of 2400 mL 2.5% Mannitol solution (Baxter, Utrecht, the Netherlands) split in two
doses: 800 mL (3 hours before examination) and 1600 mL (1 hour before examination).
Patients were scanned on a 3.0-Tesla MRI unit (Intera/Ingenia; Philips Healthcare,
Best, the Netherlands) in the supine position using the protocol outlined in table 1.
An axial free-breathing DWI sequence (with b-values 0, 300 and 600 s/mm2) was
used for apparent diffusion coefficient (ADC) mapping. A coronal 3D T1-weighted
spoiled gradient echo (SPGE) sequence with fat suppression was performed before
injection of intravenous gadolinium contrast. A dynamic contrast-enhanced (DCE)
coronal 3D T1-weighted SPGE was performed. Sixty seconds after the start of DCE-
MRI, 0.1 mL/kg body weight of gadobutrol (Gadovist 1.0 mmol/mL, Bayer Schering
Pharma, Berlin, Germany) was administered intravenously by bolus injection (5
ml/s). Subsequently, coronal and axial 3D T1-weighted SPGE sequences with fat
suppression were performed in the delayed phase (± 7 minutes after injection). To
achieve spasmolysis, 10 mg of butylscopolamine bromide (Buscopan, Boehringer
Ingelheim, Germany) was administered intravenously three times at even intervals
during the examination (Table 1).
Table 1. Protocol for MRI acquisition
Plane Fat saturation
Slice thick-ness (mm)
TR/TE (ms)
Flip an-gle (°)
FOV (mm) Matrix
Balanced GE Coronal No 5 2.5/1.25 60 380x380
T2-w SSFSE Coronal No 4 633/60 90 380x380 384x384
T2-w SSFSE Axial No 4 759/119 90 400x400 528x528
T2-w SSFSE Axial Yes 7 1088/50 90 380x302 320x256
DWI Axial Yes 7 2972/60 90 380x260 288x196
3D T1-w SPGE Coronal Yes 2 2.39/1.07 10 380x380 192x192
3D T1-w SPGE Axial Yes 2 2.27/1.09 10 380x380 384x384
FFE Fast Field Echo, HASTE Half-Fourier Acquisition Single-Shot Turbo Spin-Echo, THRIVE T1 High-Resolution Isotropic Volume Excitation
Ileocolonoscopy
Ileocolonoscopy was performed using standard bowel preparation and equipment
(model CF-160L, Olympus) within two weeks of the MRI examination by either a
gastroenterologist or senior resident under direct supervision of a gastroenterologist.
Proefschrift2018-new2.indb 113 13/9/18 10:06
114
Chapter 5
The endoscopist was blinded to results from MRI, with the exception of cases where
balloon-dilatation was indicated. For those cases, the stricture length on MRI was
used to determine the feasibility of balloon-dilatation. The segmental Crohn’s
Disease Endoscopic Index of Severity (CDEIS) was calculated for all endoscopically
intubated terminal ileum segments [17].
Image analysis
Three abdominal radiologists (C.Y.N., K.H., B.M.) with respectively 18, 8 and 11 years
of experience in IBD imaging, evaluated each case at two different time points.
Cases were initially evaluated using either CE-MRI or DW-MRI. Balanced GE and T2-
weighted SSFSE (with and without fat suppression) were included in both cases.
Directly after evaluation with the initial protocol, the omitted sequences were added
to form a combined protocol (CE/DW-MRI or DW/CE-MRI, depending on the initial
protocol).
Cases were randomly assigned an initial evaluation protocol at the first time point
and the assignment was reversed at the second time point. To reduce memory bias,
an interval of 6 weeks was used between the first and second time point, and case
numbering and order were again randomized.
Imaging sequences were separately graded for quality (0 – non diagnostic, 1 –
diagnostic, numerous artefacts, 2 – diagnostic, few artefacts, 3 – diagnostic, no
artefacts). Disease activity was graded using the MRI features based on T1-, T2-,
and diffusion-weighted sequences and grading criteria presented in table 2. Active
disease was defined as the presence of >0 grade on one or more disease features.
Using 11-point Likert scales, readers graded overall severity (0: no disease – 10:
very severe disease) and their confidence for grading (0: not confident – 10: fully
confident). After initial evaluation, the omitted sequences were added and the
following features were once more evaluated: T1 enhancement and pattern or DWI
mural signal (depending on the added sequences), stenosis, comb sign, fistulas,
abscess and enlarged lymph nodes. Optionally, overall severity grading and level
of confidence could be changed. Comb sign, enlarged lymph nodes, fistula and
abscess were deemed present when found by at least two of three readers.
Proefschrift2018-new2.indb 114 13/9/18 10:06
114
Chapter 5
The endoscopist was blinded to results from MRI, with the exception of cases where
balloon-dilatation was indicated. For those cases, the stricture length on MRI was
used to determine the feasibility of balloon-dilatation. The segmental Crohn’s
Disease Endoscopic Index of Severity (CDEIS) was calculated for all endoscopically
intubated terminal ileum segments [17].
Image analysis
Three abdominal radiologists (C.Y.N., K.H., B.M.) with respectively 18, 8 and 11 years
of experience in IBD imaging, evaluated each case at two different time points.
Cases were initially evaluated using either CE-MRI or DW-MRI. Balanced GE and T2-
weighted SSFSE (with and without fat suppression) were included in both cases.
Directly after evaluation with the initial protocol, the omitted sequences were added
to form a combined protocol (CE/DW-MRI or DW/CE-MRI, depending on the initial
protocol).
Cases were randomly assigned an initial evaluation protocol at the first time point
and the assignment was reversed at the second time point. To reduce memory bias,
an interval of 6 weeks was used between the first and second time point, and case
numbering and order were again randomized.
Imaging sequences were separately graded for quality (0 – non diagnostic, 1 –
diagnostic, numerous artefacts, 2 – diagnostic, few artefacts, 3 – diagnostic, no
artefacts). Disease activity was graded using the MRI features based on T1-, T2-,
and diffusion-weighted sequences and grading criteria presented in table 2. Active
disease was defined as the presence of >0 grade on one or more disease features.
Using 11-point Likert scales, readers graded overall severity (0: no disease – 10:
very severe disease) and their confidence for grading (0: not confident – 10: fully
confident). After initial evaluation, the omitted sequences were added and the
following features were once more evaluated: T1 enhancement and pattern or DWI
mural signal (depending on the added sequences), stenosis, comb sign, fistulas,
abscess and enlarged lymph nodes. Optionally, overall severity grading and level
of confidence could be changed. Comb sign, enlarged lymph nodes, fistula and
abscess were deemed present when found by at least two of three readers.
Proefschrift2018-new2.indb 114 13/9/18 10:06
115
Comparison of CE-MRI and DW-MRI
Table 2. MRI features and grading criteria
MRI features Score
0 1 2 3
Mural thickness (mm)
Mural T2 signal Normal bowel wall
Minor increase - bowel wall ap-pears dark gray
Moderate increase - bowel wall ap-pears light grey
Marked increase - bowel wall contains areas of white
Perimural T2 signal Equivalent to nor-mal mesentery
Increase in mes-enteric signal
Small fluid rim (≤ 2 mm)
Larger fluid rim (> 2 mm)
T1 enhancement Equivalent to nor-mal bowel wall
Minor enhance-ment - greater than normal bowel but signifi-cantly less than nearby vascular structures
Moderate enhancement - greater than normal bowel but somewhat less than nearby vas-cular structures
Marked enhance-ment - ap-proaches that of nearby vascular structures
Enhancement pattern
Not applicable Homogeneous Mucosal Layered
Mural DWI signal Equal intensity to adjacent normal bowel
Slightly higher than adjacent normal bowel
Definitely higher than adjacent normal bowel
Stenosis (% lumen reduction)
No stenosis (0–80%)
Stenosis (>80%) without preste-notic dilatation
Stenosis (>80%) with prestenotic dilatation
Comb sign Absent Present
Fistulas Absent Present
Abscess Absent Present
Lymph nodes (> 1 cm)
Absent Present
Statistical analysis
Active endoscopic disease was defined as a CDEIS ≥3 [18]. Parameters for the
diagnostic value of MRI for endoscopic active disease were calculated for all imaging
protocols and sensitivity, specificity and accuracy were compared using the McNemar
test. Individual features and severity grading scores were correlated to CDEIS
using the Spearman rank correlation. For comparison of correlation coefficients
between datasets the Steiger’s Z-test for dependent, overlapping correlations was
used [19]. Confidence scores were compared using the Wilcoxon signed rank test.
Interobserver agreement was calculated using Fleiss’ kappa coefficients for binomial
Proefschrift2018-new2.indb 115 13/9/18 10:06
115
Comparison of CE-MRI and DW-MRI
Table 2. MRI features and grading criteria
MRI features Score
0 1 2 3
Mural thickness (mm)
Mural T2 signal Normal bowel wall
Minor increase - bowel wall ap-pears dark gray
Moderate increase - bowel wall ap-pears light grey
Marked increase - bowel wall contains areas of white
Perimural T2 signal Equivalent to nor-mal mesentery
Increase in mes-enteric signal
Small fluid rim (≤ 2 mm)
Larger fluid rim (> 2 mm)
T1 enhancement Equivalent to nor-mal bowel wall
Minor enhance-ment - greater than normal bowel but signifi-cantly less than nearby vascular structures
Moderate enhancement - greater than normal bowel but somewhat less than nearby vas-cular structures
Marked enhance-ment - ap-proaches that of nearby vascular structures
Enhancement pattern
Not applicable Homogeneous Mucosal Layered
Mural DWI signal Equal intensity to adjacent normal bowel
Slightly higher than adjacent normal bowel
Definitely higher than adjacent normal bowel
Stenosis (% lumen reduction)
No stenosis (0–80%)
Stenosis (>80%) without preste-notic dilatation
Stenosis (>80%) with prestenotic dilatation
Comb sign Absent Present
Fistulas Absent Present
Abscess Absent Present
Lymph nodes (> 1 cm)
Absent Present
Statistical analysis
Active endoscopic disease was defined as a CDEIS ≥3 [18]. Parameters for the
diagnostic value of MRI for endoscopic active disease were calculated for all imaging
protocols and sensitivity, specificity and accuracy were compared using the McNemar
test. Individual features and severity grading scores were correlated to CDEIS
using the Spearman rank correlation. For comparison of correlation coefficients
between datasets the Steiger’s Z-test for dependent, overlapping correlations was
used [19]. Confidence scores were compared using the Wilcoxon signed rank test.
Interobserver agreement was calculated using Fleiss’ kappa coefficients for binomial
Proefschrift2018-new2.indb 115 13/9/18 10:06
116
Chapter 5
data and intraclass coefficients for continuous and ordinal data [20,21]. Kappa and
intraclass correlation coefficient (ICC) values were interpreted using the following
criteria: ≤0.20, poor; ≥0.21–0.40, fair; ≥0.41–0.60, moderate; ≥0.61–0.80, good; ≥0.81–
1.00,very good [22]. Interpretation of Spearman’s correlation coefficient was as
follows: 0–0.20, very weak; ≥0.20–0.40, weak; ≥0.40–0.60, moderate; ≥0.60–0.80,
strong; ≥0.80–1.00, very strong. A p-value of < 0.05 was considered significant. All
analyses were performed in SPSS 22 for Mac (SPSS, Chicago, Ill) and R Statistical
language (v3.1.2, Vienna, Austria).
RESULTS
Patients
From a total of 89 eligible patients, 28 were excluded as detailed in Figure 1.
Eventually, 61 patients were evaluated. Six of these patients had a final diagnosis
other than Crohn’s disease: ulcerative colitis (n=1), irritable bowel syndrome (n=4)
and unclear diagnosis (n=1). These patients were included in the analysis, as these
diagnoses were unknown before MRI and ileocolonoscopy, and prior indication for
MRI was the same as for other patients. Patient clinical characteristics are presented
in table 3.
Image quality and distension
For each evaluation protocol, a total of 183 evaluations were performed by the
three readers. CE and DW sequences showed good image quality (score 2 or 3)
in 98% and 93% of cases, respectively, with mean (SD) image quality scores of
2.9 (0.4) and 2.6 (0.7). In observer 1 and 2 no significant differences were seen,
although for observer 3, CE sequences showed a significantly higher image quality.
CE sequences were all rated as diagnostic, while six DW sequences were rated as
non-diagnostic due to severe artefacts (3%). The terminal ileum could be evaluated
on CE and DW sequences in 98% and 92% of cases, respectively. Adequate terminal
ileum distension (score 2 or 3) was seen in 88% of evaluations.
Evaluation of diagnostic accuracy
Sensitivity, specificity, positive predictive value, negative predictive value and
Proefschrift2018-new2.indb 116 13/9/18 10:06
116
Chapter 5
data and intraclass coefficients for continuous and ordinal data [20,21]. Kappa and
intraclass correlation coefficient (ICC) values were interpreted using the following
criteria: ≤0.20, poor; ≥0.21–0.40, fair; ≥0.41–0.60, moderate; ≥0.61–0.80, good; ≥0.81–
1.00,very good [22]. Interpretation of Spearman’s correlation coefficient was as
follows: 0–0.20, very weak; ≥0.20–0.40, weak; ≥0.40–0.60, moderate; ≥0.60–0.80,
strong; ≥0.80–1.00, very strong. A p-value of < 0.05 was considered significant. All
analyses were performed in SPSS 22 for Mac (SPSS, Chicago, Ill) and R Statistical
language (v3.1.2, Vienna, Austria).
RESULTS
Patients
From a total of 89 eligible patients, 28 were excluded as detailed in Figure 1.
Eventually, 61 patients were evaluated. Six of these patients had a final diagnosis
other than Crohn’s disease: ulcerative colitis (n=1), irritable bowel syndrome (n=4)
and unclear diagnosis (n=1). These patients were included in the analysis, as these
diagnoses were unknown before MRI and ileocolonoscopy, and prior indication for
MRI was the same as for other patients. Patient clinical characteristics are presented
in table 3.
Image quality and distension
For each evaluation protocol, a total of 183 evaluations were performed by the
three readers. CE and DW sequences showed good image quality (score 2 or 3)
in 98% and 93% of cases, respectively, with mean (SD) image quality scores of
2.9 (0.4) and 2.6 (0.7). In observer 1 and 2 no significant differences were seen,
although for observer 3, CE sequences showed a significantly higher image quality.
CE sequences were all rated as diagnostic, while six DW sequences were rated as
non-diagnostic due to severe artefacts (3%). The terminal ileum could be evaluated
on CE and DW sequences in 98% and 92% of cases, respectively. Adequate terminal
ileum distension (score 2 or 3) was seen in 88% of evaluations.
Evaluation of diagnostic accuracy
Sensitivity, specificity, positive predictive value, negative predictive value and
Proefschrift2018-new2.indb 116 13/9/18 10:06
117
Comparison of CE-MRI and DW-MRI
accuracy for each reader for detection of disease activity on MRI as compared to the
endoscopic reference standard can be found in table 4. No significant differences
were seen between imaging protocols (P>0.1).
Table 3. Patient characteristics
No. of patients, n 61
Female, n (%) 30 (49)
Age at MRI, median (IQR) 36 (26–47)
Age at diagnosis, median (IQR) 23 (19–30)
Disease duration (years), median (IQR) 11 (5–17)
Harvey-Bradshaw index, median (IQR) 6 (2–9)
Previous surgery, n (%) 30 (49)
Therapy at time of MRI
5-ASA, n (%) 6 (10)
Corticosteroids, n (%) 15 (25)
Thiopurines, n (%) 9 (15)
Methotrexate, n (%) 3 (5)
Anti-TNF antibodies, n (%) 20 (33)
CRP (mg/l), median (IQR) 3.7 (1.1–10.7)
CDEIS, median (IQR) 9 (0–16)
CDEIS <3.5/3.5–7/>7, n (%) 21 (34) / 3 (5) / 37 (61)
CDEIS Crohn's Disease Endoscopic Index of Activity; CRP C-reactive protein, IQR Inter-quartile range; MRI Magnetic Resonance Imaging
Assessment of disease severity and confidence
Correlations between severity grading and CDEIS, interobserver agreement for
severity grading and levels of confidence are presented in table 5. No significant
differences were found between the correlation coefficients of different imaging
protocols (P>0.05), although CE-MRI showed numerically higher coefficients than
DW-MRI for all readers. Interobserver agreement for severity grading was very good
for CE-MRI, CE/DW-MRI and DW/CE-MRI (ICC: 0.84, 0.82 and 0.85, respectively)
and good for DW-MRI (ICC: 0.79). Confidence levels were significantly higher for CE-
MRI than for DW-MRI (P<0.02). Combined imaging protocols showed significantly
improved confidence levels over DW-MRI for all observers (P<0.001) and over CE-
MRI for observer 1 and 2 (P<0.001), but not for observer 3 (P=0.06).
Proefschrift2018-new2.indb 117 13/9/18 10:06
117
Comparison of CE-MRI and DW-MRI
accuracy for each reader for detection of disease activity on MRI as compared to the
endoscopic reference standard can be found in table 4. No significant differences
were seen between imaging protocols (P>0.1).
Table 3. Patient characteristics
No. of patients, n 61
Female, n (%) 30 (49)
Age at MRI, median (IQR) 36 (26–47)
Age at diagnosis, median (IQR) 23 (19–30)
Disease duration (years), median (IQR) 11 (5–17)
Harvey-Bradshaw index, median (IQR) 6 (2–9)
Previous surgery, n (%) 30 (49)
Therapy at time of MRI
5-ASA, n (%) 6 (10)
Corticosteroids, n (%) 15 (25)
Thiopurines, n (%) 9 (15)
Methotrexate, n (%) 3 (5)
Anti-TNF antibodies, n (%) 20 (33)
CRP (mg/l), median (IQR) 3.7 (1.1–10.7)
CDEIS, median (IQR) 9 (0–16)
CDEIS <3.5/3.5–7/>7, n (%) 21 (34) / 3 (5) / 37 (61)
CDEIS Crohn's Disease Endoscopic Index of Activity; CRP C-reactive protein, IQR Inter-quartile range; MRI Magnetic Resonance Imaging
Assessment of disease severity and confidence
Correlations between severity grading and CDEIS, interobserver agreement for
severity grading and levels of confidence are presented in table 5. No significant
differences were found between the correlation coefficients of different imaging
protocols (P>0.05), although CE-MRI showed numerically higher coefficients than
DW-MRI for all readers. Interobserver agreement for severity grading was very good
for CE-MRI, CE/DW-MRI and DW/CE-MRI (ICC: 0.84, 0.82 and 0.85, respectively)
and good for DW-MRI (ICC: 0.79). Confidence levels were significantly higher for CE-
MRI than for DW-MRI (P<0.02). Combined imaging protocols showed significantly
improved confidence levels over DW-MRI for all observers (P<0.001) and over CE-
MRI for observer 1 and 2 (P<0.001), but not for observer 3 (P=0.06).
Proefschrift2018-new2.indb 117 13/9/18 10:06
118
Chapter 5
Tab
le 4
. Dia
gn
ost
ic a
ccu
racy v
alu
es
for
all
imag
ing
pro
toco
ls p
er
read
er
CE
-MR
ID
W-M
RI
CE
/DW
-MR
ID
W/C
E-M
RI
Ob
1O
b2
Ob
3O
b1
Ob
2O
b3
Ob
1O
b2
Ob
3O
b1
Ob
2O
b3
Sen
siti
vit
y0
.78
0.8
50
.85
0.6
90
.73
0.8
20
.78
0.8
50
.85
0.6
80
.83
0.7
9
Sp
ecif
icit
y0
.95
0.7
50
.85
0.9
00
.80
0.8
50
.95
0.6
50
.79
0.8
50
.80
0.8
5
PP
V0
.97
0.8
80
.92
0.9
30
.88
0.9
10
.97
0.8
30
.89
0.9
00
.89
0.9
1
NP
V0
.68
0.7
10
.74
0.6
00
.59
0.7
10
.68
0.6
80
.71
0.5
70
.70
0.6
8
Accu
racy
0.8
40
.82
0.8
50
.76
0.7
50
.83
0.8
40
.79
0.8
30
.74
0.8
20
.81
CE
co
ntr
ast
en
han
ced
, D
W d
iffu
sio
n-w
eig
hte
d, M
RI m
ag
neti
c r
eso
nan
ce im
ag
ing
, N
PV
neg
ati
ve p
red
icti
ve v
alu
e, O
b o
bse
rver, P
PV
po
siti
ve
pre
dic
tive v
alu
e
Tab
le 5
. Severi
ty g
rad
ing
co
rrela
tio
n t
o C
DE
IS a
nd
co
nfi
den
ce s
co
res
for
each
ob
serv
er
an
d im
ag
ing
pro
toco
l. S
everi
ty g
rad
ing
co
nfi
den
ce f
or
each
im
ag
ing
pro
toco
l
C
E-M
RI
DW
-MR
IC
E/D
W-M
RI
DW
/CE
-MR
I
Co
rrela
tio
n c
oeff
icie
nt
(to
CD
EIS
)O
bse
rver
10
.74
0.6
80
.75
0.6
7
Ob
serv
er
20
.70
.66
0.6
90
.74
Ob
serv
er
30
.74
0.7
0.7
20
.72
Inte
rob
serv
er
ag
reem
en
t,
ICC
(9
5%
CI)
0.8
4 (
0.7
6–0
.89
)0
.79
(0
.69
–0.8
6)
0.8
2 (
0.7
4–0
.88
)0
.85
(0
.76
–0.9
1)
Co
nfi
den
ce s
co
res,
mean
(S
D)
Ob
serv
er
18
.5 (
1.1)
7.5
(1.0
)8
.9 (
0.9
)9
.0 (
0.7
)
Ob
serv
er
28
.6 (
1.3
)8
.1 (
1.5
)9
.2 (
1.1)
9.3
(1.0
)
Ob
serv
er
37.
0 (
2.0
)6
.2 (
1.8
)7.
4 (
2.1)
7.4
(1.
9)
CD
EIS
Cro
hn
’s d
isease
en
do
sco
pic
in
dex o
f se
veri
ty, C
E c
on
trast
-en
han
ced
, C
I co
nfi
den
ce in
terv
al,
DW
dif
fusi
on
-weig
hte
d, IC
C in
tracla
ss
co
rrela
tio
n c
oeff
icie
nt,
MR
I m
ag
neti
c r
eso
nan
ce im
ag
ing
, S
D s
tan
dard
devia
tio
n
Proefschrift2018-new2.indb 118 13/9/18 10:06
118
Chapter 5
Tab
le 4
. Dia
gn
ost
ic a
ccu
racy v
alu
es
for
all
imag
ing
pro
toco
ls p
er
read
er
CE
-MR
ID
W-M
RI
CE
/DW
-MR
ID
W/C
E-M
RI
Ob
1O
b2
Ob
3O
b1
Ob
2O
b3
Ob
1O
b2
Ob
3O
b1
Ob
2O
b3
Sen
siti
vit
y0
.78
0.8
50
.85
0.6
90
.73
0.8
20
.78
0.8
50
.85
0.6
80
.83
0.7
9
Sp
ecif
icit
y0
.95
0.7
50
.85
0.9
00
.80
0.8
50
.95
0.6
50
.79
0.8
50
.80
0.8
5
PP
V0
.97
0.8
80
.92
0.9
30
.88
0.9
10
.97
0.8
30
.89
0.9
00
.89
0.9
1
NP
V0
.68
0.7
10
.74
0.6
00
.59
0.7
10
.68
0.6
80
.71
0.5
70
.70
0.6
8
Accu
racy
0.8
40
.82
0.8
50
.76
0.7
50
.83
0.8
40
.79
0.8
30
.74
0.8
20
.81
CE
co
ntr
ast
en
han
ced
, D
W d
iffu
sio
n-w
eig
hte
d, M
RI m
ag
neti
c r
eso
nan
ce im
ag
ing
, N
PV
neg
ati
ve p
red
icti
ve v
alu
e, O
b o
bse
rver, P
PV
po
siti
ve
pre
dic
tive v
alu
e
Tab
le 5
. Severi
ty g
rad
ing
co
rrela
tio
n t
o C
DE
IS a
nd
co
nfi
den
ce s
co
res
for
each
ob
serv
er
an
d im
ag
ing
pro
toco
l. S
everi
ty g
rad
ing
co
nfi
den
ce f
or
each
im
ag
ing
pro
toco
l
C
E-M
RI
DW
-MR
IC
E/D
W-M
RI
DW
/CE
-MR
I
Co
rrela
tio
n c
oeff
icie
nt
(to
CD
EIS
)O
bse
rver
10
.74
0.6
80
.75
0.6
7
Ob
serv
er
20
.70
.66
0.6
90
.74
Ob
serv
er
30
.74
0.7
0.7
20
.72
Inte
rob
serv
er
ag
reem
en
t,
ICC
(9
5%
CI)
0.8
4 (
0.7
6–0
.89
)0
.79
(0
.69
–0.8
6)
0.8
2 (
0.7
4–0
.88
)0
.85
(0
.76
–0.9
1)
Co
nfi
den
ce s
co
res,
mean
(S
D)
Ob
serv
er
18
.5 (
1.1)
7.5
(1.0
)8
.9 (
0.9
)9
.0 (
0.7
)
Ob
serv
er
28
.6 (
1.3
)8
.1 (
1.5
)9
.2 (
1.1)
9.3
(1.0
)
Ob
serv
er
37.
0 (
2.0
)6
.2 (
1.8
)7.
4 (
2.1)
7.4
(1.
9)
CD
EIS
Cro
hn
’s d
isease
en
do
sco
pic
in
dex o
f se
veri
ty, C
E c
on
trast
-en
han
ced
, C
I co
nfi
den
ce in
terv
al,
DW
dif
fusi
on
-weig
hte
d, IC
C in
tracla
ss
co
rrela
tio
n c
oeff
icie
nt,
MR
I m
ag
neti
c r
eso
nan
ce im
ag
ing
, S
D s
tan
dard
devia
tio
n
Proefschrift2018-new2.indb 118 13/9/18 10:06
119
Comparison of CE-MRI and DW-MRI
Individual MRI features
Strong correlation to CDEIS was seen for following features: wall thickness
(r=0.64–0.72), mural T2 signal (r=0.64–0.77), T1 enhancement (r=0.62–0.75) and
enhancement pattern (r=0.63–0.77). Mural DWI signal showed moderate-to-strong
correlation to CDEIS (r=0.58–0.71), while perimural T2 signal and stenosis showed
weak-to-moderate correlation to CDEIS (r=0.21–0.48 and r=0.33–0.55, respectively).
ICC values showed good interobserver agreement for wall thickness (0.75–0.78),
mural T2 signal (0.77–0.82), T1 enhancement (0.75–0.83), enhancement pattern
(0.75–0.77) and mural DWI signal (0.67–0.70), while lower agreement was seen for
perimural T2 signal (0.50–0.62) and stenosis (0.52–0.63).
Five fistulas were identified using CE-MRI, of which three were identified using
DW-MRI, while the other two were only identified by one reader using DW-MRI.
No additional fistulas were described using DW-MRI or the combined imaging
protocols. No abscesses were identified. CE-MRI showed good kappa values against
fair-to-moderate kappa values for DW-MRI: comb sign (0.66 vs. 0.45), fistula (0.71
vs. 0.53) and enlarged lymph nodes (0.61 vs. 0.22).
DISCUSSION
The results from our study indicate that CE-MRI and DW-MRI have comparable
accuracy in diagnosis and grading of disease activity, although readers had
significantly higher levels of grading confidence using CE-MRI. Despite the low
prevalence of fistulas, a discrepancy in detection rate was seen in favour of CE-
MRI. Good agreement was seen for detection of extramural disease features
(i.e. comb sign, fistula and enlarged lymph nodes) on CE-MRI, against fair-to-
moderate agreement using DW-MRI. Furthermore, a considerable proportion of
DWI sequences were considered non-evaluable (8%). Combined imaging protocols
showed no increase in diagnostic or grading performance, with the exception of
increased confidence levels over CE-MRI for two out of three readers.
A number of previous studies have compared separate aspects of the use of CE-
MRI and DWI-MRI for Crohn’s disease. Two pediatric studies compared the accuracy
for detection of small bowel lesions between DWI-MRI and CE-MRI, and found that
DWI-MRI provided similar or even better performance [23,24]. A recent longitudinal
Proefschrift2018-new2.indb 119 13/9/18 10:06
119
Comparison of CE-MRI and DW-MRI
Individual MRI features
Strong correlation to CDEIS was seen for following features: wall thickness
(r=0.64–0.72), mural T2 signal (r=0.64–0.77), T1 enhancement (r=0.62–0.75) and
enhancement pattern (r=0.63–0.77). Mural DWI signal showed moderate-to-strong
correlation to CDEIS (r=0.58–0.71), while perimural T2 signal and stenosis showed
weak-to-moderate correlation to CDEIS (r=0.21–0.48 and r=0.33–0.55, respectively).
ICC values showed good interobserver agreement for wall thickness (0.75–0.78),
mural T2 signal (0.77–0.82), T1 enhancement (0.75–0.83), enhancement pattern
(0.75–0.77) and mural DWI signal (0.67–0.70), while lower agreement was seen for
perimural T2 signal (0.50–0.62) and stenosis (0.52–0.63).
Five fistulas were identified using CE-MRI, of which three were identified using
DW-MRI, while the other two were only identified by one reader using DW-MRI.
No additional fistulas were described using DW-MRI or the combined imaging
protocols. No abscesses were identified. CE-MRI showed good kappa values against
fair-to-moderate kappa values for DW-MRI: comb sign (0.66 vs. 0.45), fistula (0.71
vs. 0.53) and enlarged lymph nodes (0.61 vs. 0.22).
DISCUSSION
The results from our study indicate that CE-MRI and DW-MRI have comparable
accuracy in diagnosis and grading of disease activity, although readers had
significantly higher levels of grading confidence using CE-MRI. Despite the low
prevalence of fistulas, a discrepancy in detection rate was seen in favour of CE-
MRI. Good agreement was seen for detection of extramural disease features
(i.e. comb sign, fistula and enlarged lymph nodes) on CE-MRI, against fair-to-
moderate agreement using DW-MRI. Furthermore, a considerable proportion of
DWI sequences were considered non-evaluable (8%). Combined imaging protocols
showed no increase in diagnostic or grading performance, with the exception of
increased confidence levels over CE-MRI for two out of three readers.
A number of previous studies have compared separate aspects of the use of CE-
MRI and DWI-MRI for Crohn’s disease. Two pediatric studies compared the accuracy
for detection of small bowel lesions between DWI-MRI and CE-MRI, and found that
DWI-MRI provided similar or even better performance [23,24]. A recent longitudinal
Proefschrift2018-new2.indb 119 13/9/18 10:06
120
Chapter 5
study by Huh et al. showed that DWI-MRI could be used to diagnose complete
remission and improved inflammation after medical therapy with 76% and 84%
accuracy, respectively [25]. Our results are similar to that of a previous study by
Seo et al., which found no significant difference in terms of correlation to CDEIS
between CE-MRI (r=0.71) and DW-MRI (r=0.61), although a similar discrepancy for
detection of penetrating complications was reported in favour of CE-MRI [11]. In
two studies by Schmid-Tannwald et al, CE-MRI showed no significant differences to
DW-MRI for sensitivity for the diagnosis of active inflammation (0.80 versus 0.67)
and detection of fistulas and sinus tracts (0.96 versus 0.76) [12,26]. Although these
studies, and our own, found no significant differences between CE-MRI and DW-
MRI, the consistency of numerical differences in favour of CE-MRI raises concern
whether these are not based on random variance and might indicate a clinically
relevant difference in diagnosis and grading. Preferably, a systematic review
and if possible a meta-analysis of the mentioned studies should be performed
to definitively conclude whether CE-MRI has superior diagnostic and grading
accuracy over DW-MRI. Furthermore, none of the previous studies evaluated levels
of confidence, which should also be taken into account.
Our study had several limitations that should be addressed. Due to the limited
field-of-view of the DWI sequence, which was positioned on the terminal ileum,
only these segments were analyzed in the current study. A similar study reported
a higher rate of false positives in colonic segments, compared to terminal ileum
segments [9]. Delayed contrast-enhanced sequences used in our study and portal-
venous phase sequences have both shown capable of detecting and grading
mural lesions [4,27], although a recent study found that an enhancement ratio
between early and delayed sequences could be used for further characterization
of inflammation and fibrosis [28]. In our study, DW-MRI was performed using three
b-values (0, 300, 600 s/mm2). A recent study and review advised the use of a
slightly higher b-value of 800 s/mm2 to provide the best diagnostic accuracy and
signal-to-noise ratio [8,10]. The use of quantitative DWI measurements, namely
the apparent diffusion coefficient (ADC), has been investigated in several research
studies, and has shown promise as a biomarker for bowel inflammation, although
concerns over its reproducibility should be addressed in future studies [29]. As
such, ADC measurement was not included in the current study. Quantification of
contrast enhancement can be obtained using the relative contrast enhancement
(RCE) feature, which is incorporated in the Magnetic Resonance Index of Activity
Proefschrift2018-new2.indb 120 13/9/18 10:06
120
Chapter 5
study by Huh et al. showed that DWI-MRI could be used to diagnose complete
remission and improved inflammation after medical therapy with 76% and 84%
accuracy, respectively [25]. Our results are similar to that of a previous study by
Seo et al., which found no significant difference in terms of correlation to CDEIS
between CE-MRI (r=0.71) and DW-MRI (r=0.61), although a similar discrepancy for
detection of penetrating complications was reported in favour of CE-MRI [11]. In
two studies by Schmid-Tannwald et al, CE-MRI showed no significant differences to
DW-MRI for sensitivity for the diagnosis of active inflammation (0.80 versus 0.67)
and detection of fistulas and sinus tracts (0.96 versus 0.76) [12,26]. Although these
studies, and our own, found no significant differences between CE-MRI and DW-
MRI, the consistency of numerical differences in favour of CE-MRI raises concern
whether these are not based on random variance and might indicate a clinically
relevant difference in diagnosis and grading. Preferably, a systematic review
and if possible a meta-analysis of the mentioned studies should be performed
to definitively conclude whether CE-MRI has superior diagnostic and grading
accuracy over DW-MRI. Furthermore, none of the previous studies evaluated levels
of confidence, which should also be taken into account.
Our study had several limitations that should be addressed. Due to the limited
field-of-view of the DWI sequence, which was positioned on the terminal ileum,
only these segments were analyzed in the current study. A similar study reported
a higher rate of false positives in colonic segments, compared to terminal ileum
segments [9]. Delayed contrast-enhanced sequences used in our study and portal-
venous phase sequences have both shown capable of detecting and grading
mural lesions [4,27], although a recent study found that an enhancement ratio
between early and delayed sequences could be used for further characterization
of inflammation and fibrosis [28]. In our study, DW-MRI was performed using three
b-values (0, 300, 600 s/mm2). A recent study and review advised the use of a
slightly higher b-value of 800 s/mm2 to provide the best diagnostic accuracy and
signal-to-noise ratio [8,10]. The use of quantitative DWI measurements, namely
the apparent diffusion coefficient (ADC), has been investigated in several research
studies, and has shown promise as a biomarker for bowel inflammation, although
concerns over its reproducibility should be addressed in future studies [29]. As
such, ADC measurement was not included in the current study. Quantification of
contrast enhancement can be obtained using the relative contrast enhancement
(RCE) feature, which is incorporated in the Magnetic Resonance Index of Activity
Proefschrift2018-new2.indb 120 13/9/18 10:06
121
Comparison of CE-MRI and DW-MRI
(MaRIA) [6]. However, manual region of interest placement and corrections of signal
intensity values on certain scanner types complicate the use of these measurements
[30].
A recent meta-analysis which investigated DW-MRI for the diagnosis of bowel
inflammation in Crohn’s disease revealed a sensitivity of 79% and specificity of 61%
[31]. They reported a high heterogeneity in the collected data and concluded that
accuracy was likely overestimated in some studies, due to issues such as lack of
blinding and use of contrast enhanced sequences as a reference standard. Reasons
for the high number of false positives in DW-MRI are yet to be investigated, but are
suggested to be caused by inadequate bowel distension and preparation [8].
Results from our study and other studies show no significant differences between
CE-MRI and DW-MRI in terms of diagnosis and grading of bowel inflammation.
However, we found higher levels of grading confidence for CE-MRI, a higher rate
of evaluable scans and a discrepancy in diagnosis of penetrating complications in
favour of CE-MRI. Based on our findings, we would recommend the use of CE-MRI
for routine examinations, while DW-MRI can be a good alternative in patients with
contraindications for intravenous contrast medium. Although combined imaging
protocols showed increased confidence scores, they did not perform better in terms
of diagnostic or grading capabilities and revealed no additional findings.
Proefschrift2018-new2.indb 121 13/9/18 10:06
121
Comparison of CE-MRI and DW-MRI
(MaRIA) [6]. However, manual region of interest placement and corrections of signal
intensity values on certain scanner types complicate the use of these measurements
[30].
A recent meta-analysis which investigated DW-MRI for the diagnosis of bowel
inflammation in Crohn’s disease revealed a sensitivity of 79% and specificity of 61%
[31]. They reported a high heterogeneity in the collected data and concluded that
accuracy was likely overestimated in some studies, due to issues such as lack of
blinding and use of contrast enhanced sequences as a reference standard. Reasons
for the high number of false positives in DW-MRI are yet to be investigated, but are
suggested to be caused by inadequate bowel distension and preparation [8].
Results from our study and other studies show no significant differences between
CE-MRI and DW-MRI in terms of diagnosis and grading of bowel inflammation.
However, we found higher levels of grading confidence for CE-MRI, a higher rate
of evaluable scans and a discrepancy in diagnosis of penetrating complications in
favour of CE-MRI. Based on our findings, we would recommend the use of CE-MRI
for routine examinations, while DW-MRI can be a good alternative in patients with
contraindications for intravenous contrast medium. Although combined imaging
protocols showed increased confidence scores, they did not perform better in terms
of diagnostic or grading capabilities and revealed no additional findings.
Proefschrift2018-new2.indb 121 13/9/18 10:06
122
Chapter 5
REFERENCES1. Panes J, Bouhnik Y, Reinisch W, et al. Imaging techniques for assessment of inflammatory
bowel disease: Joint ECCO and ESGAR evidence-based consensus guidelines. J Crohn’s
Colitis 2013;7:556–85.
2. Maccioni F, Bruni A, Viscido A, et al. MR imaging in patients with Crohn disease: value
of T2- versus T1-weighted gadolinium-enhanced MR sequences with use of an oral
superparamagnetic contrast agent. Radiology 2006;238:517–30.
3. Low RN, Sebrechts CP, Politoske DA, et al. Crohn disease with endoscopic correlation:
single-shot fast spin-echo and gadolinium-enhanced fat-suppressed spoiled gradient-
echo MR imaging. Radiology 2002;222:652–60.
4. Tielbeek JAW, Ziech MLW, Li Z, et al. Evaluation of conventional, dynamic contrast
enhanced and diffusion weighted MRI for quantitative Crohn’s disease assessment with
histopathology of surgical specimens. Eur Radiol 2014;24:619–29.
5. Ziech MLW, Lavini C, Caan MWA, et al. Dynamic contrast-enhanced MRI in patients with
luminal Crohn’s disease. Eur J Radiol 2012;81:3019–27.
6. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment of
disease activity and severity in ileocolonic Crohn’s disease. Gut 2009;58:1113–20.
7. Dohan A, Taylor S, Hoeffel C, et al. Diffusion-weighted MRI in Crohn’s disease: Current
status and recommendations. J Magn Reson Imaging Published Online First: June 2016.
8. Park SH. DWI at MR Enterography for Evaluating Bowel Inflammation in Crohn Disease.
AJR Am J Roentgenol 2016;:1–9.
9. Kim K-J, Lee Y, Park SH, et al. Diffusion-weighted MR enterography for evaluating Crohn’s
disease: how does it add diagnostically to conventional MR enterography? Inflamm Bowel
Dis 2015;21:101–9.
10. Qi F, Jun S, Qi QY, et al. Utility of the diffusion-weighted imaging for activity evaluation
in Crohn’s disease patients underwent magnetic resonance enterography. BMC
Gastroenterol 2015;15:12.
11. Seo N, Park SH, Kim K-J, et al. MR Enterography for the Evaluation of Small-Bowel
Inflammation in Crohn Disease by Using Diffusion-weighted Imaging without Intravenous
Contrast Material: A Prospective Noninferiority Study. Radiology 2016;278:762–72.
12. Schmid-Tannwald C, Agrawal G, Dahi F, et al. Diffusion-weighted MRI: Role in detecting
abdominopelvic internal fistulas and sinus tracts. J Magn Reson Imaging 2012;35:125–31.
13. Quaia E, Sozzi M, Gennari AG, et al. Impact of gadolinium-based contrast agent in the
assessment of Crohn’s disease activity: Is contrast agent injection necessary? J Magn
Reson Imaging 2015.
14. ACR Committee on drugs and contrast Media. ACR Manual on Contrast Media. 2013.
15. 15 McDonald RJ, McDonald JS, Kallmes DF, et al. Intracranial Gadolinium Deposition
after Contrast-enhanced MR Imaging. Radiology 2015;0:150025.
16. Puylaert CAJ, Schüffler PJ, Naziroglu RE, et al. Semiautomatic Assessment of the Terminal
Ileum and Colon in Patients with Crohn Disease Using MRI (the VIGOR++ Project). Acad
Radiol Published Online First: 2018.
17. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Thérapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut 1989;30:983–9.
18. Daperno M, Castiglione F, de Ridder L, et al. Results of the 2nd part Scientific Workshop
of the ECCO (II): Measures and markers of prediction to achieve, detect, and monitor
Proefschrift2018-new2.indb 122 13/9/18 10:06
122
Chapter 5
REFERENCES1. Panes J, Bouhnik Y, Reinisch W, et al. Imaging techniques for assessment of inflammatory
bowel disease: Joint ECCO and ESGAR evidence-based consensus guidelines. J Crohn’s
Colitis 2013;7:556–85.
2. Maccioni F, Bruni A, Viscido A, et al. MR imaging in patients with Crohn disease: value
of T2- versus T1-weighted gadolinium-enhanced MR sequences with use of an oral
superparamagnetic contrast agent. Radiology 2006;238:517–30.
3. Low RN, Sebrechts CP, Politoske DA, et al. Crohn disease with endoscopic correlation:
single-shot fast spin-echo and gadolinium-enhanced fat-suppressed spoiled gradient-
echo MR imaging. Radiology 2002;222:652–60.
4. Tielbeek JAW, Ziech MLW, Li Z, et al. Evaluation of conventional, dynamic contrast
enhanced and diffusion weighted MRI for quantitative Crohn’s disease assessment with
histopathology of surgical specimens. Eur Radiol 2014;24:619–29.
5. Ziech MLW, Lavini C, Caan MWA, et al. Dynamic contrast-enhanced MRI in patients with
luminal Crohn’s disease. Eur J Radiol 2012;81:3019–27.
6. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment of
disease activity and severity in ileocolonic Crohn’s disease. Gut 2009;58:1113–20.
7. Dohan A, Taylor S, Hoeffel C, et al. Diffusion-weighted MRI in Crohn’s disease: Current
status and recommendations. J Magn Reson Imaging Published Online First: June 2016.
8. Park SH. DWI at MR Enterography for Evaluating Bowel Inflammation in Crohn Disease.
AJR Am J Roentgenol 2016;:1–9.
9. Kim K-J, Lee Y, Park SH, et al. Diffusion-weighted MR enterography for evaluating Crohn’s
disease: how does it add diagnostically to conventional MR enterography? Inflamm Bowel
Dis 2015;21:101–9.
10. Qi F, Jun S, Qi QY, et al. Utility of the diffusion-weighted imaging for activity evaluation
in Crohn’s disease patients underwent magnetic resonance enterography. BMC
Gastroenterol 2015;15:12.
11. Seo N, Park SH, Kim K-J, et al. MR Enterography for the Evaluation of Small-Bowel
Inflammation in Crohn Disease by Using Diffusion-weighted Imaging without Intravenous
Contrast Material: A Prospective Noninferiority Study. Radiology 2016;278:762–72.
12. Schmid-Tannwald C, Agrawal G, Dahi F, et al. Diffusion-weighted MRI: Role in detecting
abdominopelvic internal fistulas and sinus tracts. J Magn Reson Imaging 2012;35:125–31.
13. Quaia E, Sozzi M, Gennari AG, et al. Impact of gadolinium-based contrast agent in the
assessment of Crohn’s disease activity: Is contrast agent injection necessary? J Magn
Reson Imaging 2015.
14. ACR Committee on drugs and contrast Media. ACR Manual on Contrast Media. 2013.
15. 15 McDonald RJ, McDonald JS, Kallmes DF, et al. Intracranial Gadolinium Deposition
after Contrast-enhanced MR Imaging. Radiology 2015;0:150025.
16. Puylaert CAJ, Schüffler PJ, Naziroglu RE, et al. Semiautomatic Assessment of the Terminal
Ileum and Colon in Patients with Crohn Disease Using MRI (the VIGOR++ Project). Acad
Radiol Published Online First: 2018.
17. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Thérapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut 1989;30:983–9.
18. Daperno M, Castiglione F, de Ridder L, et al. Results of the 2nd part Scientific Workshop
of the ECCO (II): Measures and markers of prediction to achieve, detect, and monitor
Proefschrift2018-new2.indb 122 13/9/18 10:06
123
Comparison of CE-MRI and DW-MRI
intestinal healing in Inflammatory Bowel Disease. J. Crohn’s Colitis. 2011;5:484–98.
19. Steiger JH. Tests for comparing elements of a correlation matrix. Psychol. Bull.
1980;87:245–51.
20. Fleiss JL. Measuring nominal scale agreement among many raters. Psychol. Bull.
1971;76:378–82.
21. Fleiss JL. The Equivalence of Weighted Kappa and the Intraclass Correlation Coefficient
as Measures of Reliability. Educ. Psychol. Meas. 1973;33:613–9.
22. Landis JR, Koch GG. The measurement of observer agreement for categorical data.
Biometrics 1977;33:159–74.
23. Dubron C, Avni F, Boutry N, et al. Prospective evaluation of free-breathing diffusionweighted
imaging for the detection of inflammatory bowel disease with MR enterography in
childhood population. Br J Radiol 2016;89.
24. Neubauer H, Pabst T, Dick A, et al. Small-bowel MRI in children and young adults with
Crohn disease: retrospective head-to-head comparison of contrast-enhanced and
diffusion-weighted MRI. Pediatr Radiol 2013;43:103–14.
25. Huh J, Kim KJ, Park SH, et al. Diffusion-Weighted MR Enterography to Monitor Bowel
Inflammation after Medical Therapy in Crohn’s Disease: A Prospective Longitudinal Study.
Korean J Radiol 2017;18:162–72.
26. Schmid-Tannwald C, Schmid-Tannwald CM, Morelli JN, et al. The role of diffusion-weighted
MRI in assessment of inflammatory bowel disease. Abdom Radiol (New York) Published
Online First: April 2016.
27. Zappa M, Stefanescu C, Cazals-Hatem D, et al. Which magnetic resonance imaging
findings accurately evaluate inflammation in small bowel Crohn’s disease? A retrospective
comparison with surgical pathologic analysis. Inflamm Bowel Dis 2011;17:984–93.
28. Rimola J, Planell N, Rodríguez S, et al. Characterization of Inflammation and Fibrosis in
Crohn’s Disease Lesions by Magnetic Resonance Imaging. Am J Gastroenterol 2015;:1–9.
29. Kim SY, Park SH. Reply to What is the Role of Diffusion-weighted Imaging in Ileocolonic
Crohn’s Disease? Inflamm. Bowel Dis. 2015;21:E9–10.
30. Chenevert TL, Malyarenko DI, Newitt D, et al. Errors in Quantitative Image Analysis due to
Platform-Dependent Image Scaling. Transl Oncol 2014;7:65–71.
31. Choi SH, Kim KW, Lee JY, et al. Diffusion-weighted Magnetic Resonance Enterography
for Evaluating Bowel Inflammation in Crohn’s Disease: A Systematic Review and Meta-
analysis. Inflamm Bowel Dis 2016;22:669–79.
Proefschrift2018-new2.indb 123 13/9/18 10:06
123
Comparison of CE-MRI and DW-MRI
intestinal healing in Inflammatory Bowel Disease. J. Crohn’s Colitis. 2011;5:484–98.
19. Steiger JH. Tests for comparing elements of a correlation matrix. Psychol. Bull.
1980;87:245–51.
20. Fleiss JL. Measuring nominal scale agreement among many raters. Psychol. Bull.
1971;76:378–82.
21. Fleiss JL. The Equivalence of Weighted Kappa and the Intraclass Correlation Coefficient
as Measures of Reliability. Educ. Psychol. Meas. 1973;33:613–9.
22. Landis JR, Koch GG. The measurement of observer agreement for categorical data.
Biometrics 1977;33:159–74.
23. Dubron C, Avni F, Boutry N, et al. Prospective evaluation of free-breathing diffusionweighted
imaging for the detection of inflammatory bowel disease with MR enterography in
childhood population. Br J Radiol 2016;89.
24. Neubauer H, Pabst T, Dick A, et al. Small-bowel MRI in children and young adults with
Crohn disease: retrospective head-to-head comparison of contrast-enhanced and
diffusion-weighted MRI. Pediatr Radiol 2013;43:103–14.
25. Huh J, Kim KJ, Park SH, et al. Diffusion-Weighted MR Enterography to Monitor Bowel
Inflammation after Medical Therapy in Crohn’s Disease: A Prospective Longitudinal Study.
Korean J Radiol 2017;18:162–72.
26. Schmid-Tannwald C, Schmid-Tannwald CM, Morelli JN, et al. The role of diffusion-weighted
MRI in assessment of inflammatory bowel disease. Abdom Radiol (New York) Published
Online First: April 2016.
27. Zappa M, Stefanescu C, Cazals-Hatem D, et al. Which magnetic resonance imaging
findings accurately evaluate inflammation in small bowel Crohn’s disease? A retrospective
comparison with surgical pathologic analysis. Inflamm Bowel Dis 2011;17:984–93.
28. Rimola J, Planell N, Rodríguez S, et al. Characterization of Inflammation and Fibrosis in
Crohn’s Disease Lesions by Magnetic Resonance Imaging. Am J Gastroenterol 2015;:1–9.
29. Kim SY, Park SH. Reply to What is the Role of Diffusion-weighted Imaging in Ileocolonic
Crohn’s Disease? Inflamm. Bowel Dis. 2015;21:E9–10.
30. Chenevert TL, Malyarenko DI, Newitt D, et al. Errors in Quantitative Image Analysis due to
Platform-Dependent Image Scaling. Transl Oncol 2014;7:65–71.
31. Choi SH, Kim KW, Lee JY, et al. Diffusion-weighted Magnetic Resonance Enterography
for Evaluating Bowel Inflammation in Crohn’s Disease: A Systematic Review and Meta-
analysis. Inflamm Bowel Dis 2016;22:669–79.
Proefschrift2018-new2.indb 123 13/9/18 10:06
Proefschrift2018-new2.indb 124 13/9/18 10:06 Proefschrift2018-new2.indb 124 13/9/18 10:06
CHAPTER 6
Quantified terminal ileal motility during MR enterography as a biomarker of Crohn's disease activity: a prospective study
Alex Menys, Carl. A.J. Puylaert, Charlotte J. Tutein Nolthenius, Andrew A. Plumb,
Jesica C. Makanyanga, Doug A. Pendse, Cyriel Y. Ponsioen, Frans M. Vos, Jaap
Stoker, Stuart A. Taylor
Proefschrift2018-new2.indb 125 13/9/18 10:06
CHAPTER 6
Quantified terminal ileal motility during MR enterography as a biomarker of Crohn's disease activity: a prospective study
Alex Menys, Carl. A.J. Puylaert, Charlotte J. Tutein Nolthenius, Andrew A. Plumb,
Jesica C. Makanyanga, Doug A. Pendse, Cyriel Y. Ponsioen, Frans M. Vos, Jaap
Stoker, Stuart A. Taylor
Proefschrift2018-new2.indb 125 13/9/18 10:06
Chapter 6
126
ABSTRACT
Purpose
To evaluate the accuracy of MRI quantified small bowel motility for Crohn’s disease
(CD) activity against endoscopic and histopathological reference standards.
Materials and Methods
For this dual site prospective study, 82 subjects (median age 31, range 16 to 70;
42 male, median age 31, range 17 to 70, 40 female median age 31, range 16 to 59)
underwent colonoscopy and MR enterography within 14 days (October 2011 to March
2014). The Crohn’s Disease Endoscopic Activity Index (CDEIS), histopathological
activity score (eAIS) and magnetic resonance index of activity (MaRIA) were scored
in the terminal ileum (TI). TI motility was quantified using an image registration
based motility algorithm. Sensitivity and specificity of Motility (> 0.3AU) and MaRIA
(≥11) for disease activity (CDEIS >=4 or eAIS≥1) was compared using McNemar’s test,
and Receiver Operating Characteristic (ROC) curves constructed and area under
the curve (AUC) compared (Bootstrap method, R statistical software, Vienna).
Motility was correlated with reference standards using Spearman’s rank estimates.
Results
Against CDEIS and compared to MaRIA, motility had higher sensitivity (92.7%
vs. 78.1%, p=0.03), but lower specificity (61.0% vs. 80.5%, p=0.04). Against eAIS
and compared to MaRIA motility had higher sensitivity (91.7% vs. 75%, p=0.03)
but similar specificity (70.5% vs. 73.5%, P=1.0). Motility had moderate negative
correlations with eAIS r=-0.61, 95% CI -0.73 to -0.45) and CDEIS (r= -0.59 95% CI
-0.72 to -0.43)) and a ROC AUC of 0.86 (CDEIS), 0.87 (eAIS) respectively.
Conclusion
Quantified motility appears a valid biomarker for endoscopic and histopathological
activity in Crohn’s disease.
Proefschrift2018-new2.indb 126 13/9/18 10:06
Chapter 6
126
ABSTRACT
Purpose
To evaluate the accuracy of MRI quantified small bowel motility for Crohn’s disease
(CD) activity against endoscopic and histopathological reference standards.
Materials and Methods
For this dual site prospective study, 82 subjects (median age 31, range 16 to 70;
42 male, median age 31, range 17 to 70, 40 female median age 31, range 16 to 59)
underwent colonoscopy and MR enterography within 14 days (October 2011 to March
2014). The Crohn’s Disease Endoscopic Activity Index (CDEIS), histopathological
activity score (eAIS) and magnetic resonance index of activity (MaRIA) were scored
in the terminal ileum (TI). TI motility was quantified using an image registration
based motility algorithm. Sensitivity and specificity of Motility (> 0.3AU) and MaRIA
(≥11) for disease activity (CDEIS >=4 or eAIS≥1) was compared using McNemar’s test,
and Receiver Operating Characteristic (ROC) curves constructed and area under
the curve (AUC) compared (Bootstrap method, R statistical software, Vienna).
Motility was correlated with reference standards using Spearman’s rank estimates.
Results
Against CDEIS and compared to MaRIA, motility had higher sensitivity (92.7%
vs. 78.1%, p=0.03), but lower specificity (61.0% vs. 80.5%, p=0.04). Against eAIS
and compared to MaRIA motility had higher sensitivity (91.7% vs. 75%, p=0.03)
but similar specificity (70.5% vs. 73.5%, P=1.0). Motility had moderate negative
correlations with eAIS r=-0.61, 95% CI -0.73 to -0.45) and CDEIS (r= -0.59 95% CI
-0.72 to -0.43)) and a ROC AUC of 0.86 (CDEIS), 0.87 (eAIS) respectively.
Conclusion
Quantified motility appears a valid biomarker for endoscopic and histopathological
activity in Crohn’s disease.
Proefschrift2018-new2.indb 126 13/9/18 10:06
Terminal ileal motility during MR enterography
127
INTRODUCTION
Magnetic resonance enterography (MRE) is increasingly implemented for the
diagnosis and monitoring of Crohn’s disease [1]. In particular, MR activity scores
based on features such as wall thickness, T2 signal, ulceration and contrast
enhancement have been validated against a variety of reference standards including
endoscopy, histology and biochemical markers such as calprotectin [2–5], and can
effectively depict treatment response [6–8]. A limitation of current activity scores
is that treatment changes may lag behind clinical response [9] and although inter
and intra agreement data is reasonable [10], scoring is time consuming which limits
use in routine clinical practice.
An alternative to assessing activity using bowel structure is to evaluate function,
specifically segmental motility. Recent software innovations now allow rapid
quantitation of segmental bowel motility [11–14] with minimal user input. There is
emerging evidence that reduced segmental motility in affected bowel is directly
associated with Crohn’s disease inflammatory activity [15,16]. Furthermore, initial
data suggests improvements in motility in response to treatment may be better
able to predict early treatment outcome than standard MRI activity scores [6].
To date, most studies evaluating bowel motility as a biomarker of Crohn’s disease
activity have been single site and retrospective [15–18]. Such studies are vital to
create the evidence based for any imaging biomarker, but thereafter efficacy must
be proved in prospective cohort studies, ideally multisite, to show generalizability.
The purpose of our study was to evaluate the accuracy of MRI quantified small bowel
motility for Crohn’s disease (CD) activity against endoscopic and histopathological
reference standards.
MATERIALS AND METHODS
Subjects
For our dual site prospective study, subjects were recruited from two European
centres (Center 1 – University College London, UK and Academisch Medisch
Centrum, Netherlands) as part of the VIGOR++ study (funded by European Union's
Proefschrift2018-new2.indb 127 13/9/18 10:06
Terminal ileal motility during MR enterography
127
INTRODUCTION
Magnetic resonance enterography (MRE) is increasingly implemented for the
diagnosis and monitoring of Crohn’s disease [1]. In particular, MR activity scores
based on features such as wall thickness, T2 signal, ulceration and contrast
enhancement have been validated against a variety of reference standards including
endoscopy, histology and biochemical markers such as calprotectin [2–5], and can
effectively depict treatment response [6–8]. A limitation of current activity scores
is that treatment changes may lag behind clinical response [9] and although inter
and intra agreement data is reasonable [10], scoring is time consuming which limits
use in routine clinical practice.
An alternative to assessing activity using bowel structure is to evaluate function,
specifically segmental motility. Recent software innovations now allow rapid
quantitation of segmental bowel motility [11–14] with minimal user input. There is
emerging evidence that reduced segmental motility in affected bowel is directly
associated with Crohn’s disease inflammatory activity [15,16]. Furthermore, initial
data suggests improvements in motility in response to treatment may be better
able to predict early treatment outcome than standard MRI activity scores [6].
To date, most studies evaluating bowel motility as a biomarker of Crohn’s disease
activity have been single site and retrospective [15–18]. Such studies are vital to
create the evidence based for any imaging biomarker, but thereafter efficacy must
be proved in prospective cohort studies, ideally multisite, to show generalizability.
The purpose of our study was to evaluate the accuracy of MRI quantified small bowel
motility for Crohn’s disease (CD) activity against endoscopic and histopathological
reference standards.
MATERIALS AND METHODS
Subjects
For our dual site prospective study, subjects were recruited from two European
centres (Center 1 – University College London, UK and Academisch Medisch
Centrum, Netherlands) as part of the VIGOR++ study (funded by European Union's
Proefschrift2018-new2.indb 127 13/9/18 10:06
Chapter 6
128
Seventh Framework Programme, project number 270379) which ran between June
2011 and March 2014. Staff costs for the current study were in part met by National
Institute for Health Research University College London Hospitals Biomedical
Research Centre. Ethical permission was obtained from both institutions medical
ethics committee and written informed consent was obtained from each patient.
In summary, the VIGOR++ study was designed to developed software analysis
tools for bowel wall structure such as wall thickness and contrast enhancement on
order to part automate evaluation of disease activity using MRI [19–21]. Recruited
subjects had known or clinically suspected Crohn’s disease and underwent MRE
and colonoscopy with biopsy sampling within two weeks of each other (as a
convenience series). The funding agency had no influence on how the study was
performed or reported.
The VIGOR++ study identified a total of 158 recruited subjects (69 at Center 1, 89 at
Center 2). Of these, 52 subjects were excluded for the following reasons: diagnosis
other than Crohn's disease (n=18), > 14 days between MRI and colonoscopy (n=7),
failure to comply with the oral contrast protocol (n=6), cancelled or aborted
ileocolonoscopy (n=5), incomplete MRI protocol (n=14; e.g. missing sequences and
incomplete imaging), insufficient bowel cleansing (n=1) and non-compliance to
breathing commands due to language barrier (n=1) leaving 106 subjects suitable
for analysis (37 at center 1 and 69 at center 2). Our study includes a subset of this
VIGOR study cohort. For inclusion in our study, subjects had confirmed diagnosis of
Crohn’s disease based on established criteria, good quality motility images through
the terminal ileum (see motility assessment section below) as judged by our study
coordinator (post-doctoral research fellow with 8 years of experience in MRE) and
study radiologist (AM & GB [consultant gastrointestinal radiologist with 9 years of
experience) in consensus) and a terminal ileum biopsy available [22].
MRI Protocol
An identical MRI protocol was used at both sites. Subjects fasted for 4h before
ingesting 800ML 2% mannitol 3h prior to the start of the scan to fill the colon. A
further 1600ML 2% mannitol was provided 1h before the scan start time and subjects
instructed to drink to tolerance to distend the small bowel.
Subjects were scanned in the supine position on 3T Systems (Philips Achieva,
Proefschrift2018-new2.indb 128 13/9/18 10:06
Chapter 6
128
Seventh Framework Programme, project number 270379) which ran between June
2011 and March 2014. Staff costs for the current study were in part met by National
Institute for Health Research University College London Hospitals Biomedical
Research Centre. Ethical permission was obtained from both institutions medical
ethics committee and written informed consent was obtained from each patient.
In summary, the VIGOR++ study was designed to developed software analysis
tools for bowel wall structure such as wall thickness and contrast enhancement on
order to part automate evaluation of disease activity using MRI [19–21]. Recruited
subjects had known or clinically suspected Crohn’s disease and underwent MRE
and colonoscopy with biopsy sampling within two weeks of each other (as a
convenience series). The funding agency had no influence on how the study was
performed or reported.
The VIGOR++ study identified a total of 158 recruited subjects (69 at Center 1, 89 at
Center 2). Of these, 52 subjects were excluded for the following reasons: diagnosis
other than Crohn's disease (n=18), > 14 days between MRI and colonoscopy (n=7),
failure to comply with the oral contrast protocol (n=6), cancelled or aborted
ileocolonoscopy (n=5), incomplete MRI protocol (n=14; e.g. missing sequences and
incomplete imaging), insufficient bowel cleansing (n=1) and non-compliance to
breathing commands due to language barrier (n=1) leaving 106 subjects suitable
for analysis (37 at center 1 and 69 at center 2). Our study includes a subset of this
VIGOR study cohort. For inclusion in our study, subjects had confirmed diagnosis of
Crohn’s disease based on established criteria, good quality motility images through
the terminal ileum (see motility assessment section below) as judged by our study
coordinator (post-doctoral research fellow with 8 years of experience in MRE) and
study radiologist (AM & GB [consultant gastrointestinal radiologist with 9 years of
experience) in consensus) and a terminal ileum biopsy available [22].
MRI Protocol
An identical MRI protocol was used at both sites. Subjects fasted for 4h before
ingesting 800ML 2% mannitol 3h prior to the start of the scan to fill the colon. A
further 1600ML 2% mannitol was provided 1h before the scan start time and subjects
instructed to drink to tolerance to distend the small bowel.
Subjects were scanned in the supine position on 3T Systems (Philips Achieva,
Proefschrift2018-new2.indb 128 13/9/18 10:06
Terminal ileal motility during MR enterography
129
center 1 & Philips Ingenia, center 2; Philips, Best, The Netherlands) using the
manufacturer’s external body coils. Specific details of the scan parameters are
provided in Appendix 1. In summary, small bowel motility was captured using a 2D,
coronal, Balanced Turbo Field Echo (BTFE) sequences acquired during a 22 second
breath-hold. The temporal resolution of the dynamic images was 1.1 seconds per
slice with a slice thickness of 10mm. After each breath hold acquisition, radiographic
technicians repositioned the acquisition block from anterior to posterior, to cover
the whole of the small bowel volume, ensuring at least one of the dynamic series
was acquired through the terminal ileum. Following motility sequences (Appendix
1), a spasmolytic agent (Hyoscine butylbromide, Boehringer Ingelheim, Germany)
was administered intravenously (IV) and before structural MRI images including T2-
weighted and contrast-enhanced sequences were acquired.
Colonoscopy and endoscopic assessment of inflammation
Ileocolonoscopy was performed using standard bowel preparation and equipment
within two weeks of the MRI examination by either a consultant gastroenterologist
(with at least 10 years’ experience) or senior gastroenterology trainee under direct
supervision of the consultant gastroenterologist. The endoscopist was blinded to
results from MRI except when balloon-dilatation was considered for short strictures
and MR images were used to determine the stricture length. The segmental Crohn's
Disease Endoscopic Index of Severity (CDEIS) was scored for all endoscopically
intubated terminal ileum segments using conventional definitions [23], within 20cm
of the ileocecal valve. Two to four biopsies were also taken from the last 5cm of the
terminal ileum. When colonoscopy was performed by a senior trainee, the CDEIS
was assigned by the supervising consultant present. The CDEIS was selected as
a reference standard as it is a widely used quantitative metric for Crohn’s disease
activity.
Histopathological assessment of inflammation
Each terminal ileal biopsy was stained with haematoxylin-eosin and reviewed in
face-to-face consensus by two experienced pathologists (MRJ and LAB) with 10
and 15 years respectively, blinded to clinical information other than the diagnosis
of CD. The endoscopic biopsy of acute inflammation score (eAIS) was used to
semi-objectively evaluate typical morphological features associated with Crohn’s
disease activity as previously described [24,25]. The eAIS score includes measures
Proefschrift2018-new2.indb 129 13/9/18 10:06
Terminal ileal motility during MR enterography
129
center 1 & Philips Ingenia, center 2; Philips, Best, The Netherlands) using the
manufacturer’s external body coils. Specific details of the scan parameters are
provided in Appendix 1. In summary, small bowel motility was captured using a 2D,
coronal, Balanced Turbo Field Echo (BTFE) sequences acquired during a 22 second
breath-hold. The temporal resolution of the dynamic images was 1.1 seconds per
slice with a slice thickness of 10mm. After each breath hold acquisition, radiographic
technicians repositioned the acquisition block from anterior to posterior, to cover
the whole of the small bowel volume, ensuring at least one of the dynamic series
was acquired through the terminal ileum. Following motility sequences (Appendix
1), a spasmolytic agent (Hyoscine butylbromide, Boehringer Ingelheim, Germany)
was administered intravenously (IV) and before structural MRI images including T2-
weighted and contrast-enhanced sequences were acquired.
Colonoscopy and endoscopic assessment of inflammation
Ileocolonoscopy was performed using standard bowel preparation and equipment
within two weeks of the MRI examination by either a consultant gastroenterologist
(with at least 10 years’ experience) or senior gastroenterology trainee under direct
supervision of the consultant gastroenterologist. The endoscopist was blinded to
results from MRI except when balloon-dilatation was considered for short strictures
and MR images were used to determine the stricture length. The segmental Crohn's
Disease Endoscopic Index of Severity (CDEIS) was scored for all endoscopically
intubated terminal ileum segments using conventional definitions [23], within 20cm
of the ileocecal valve. Two to four biopsies were also taken from the last 5cm of the
terminal ileum. When colonoscopy was performed by a senior trainee, the CDEIS
was assigned by the supervising consultant present. The CDEIS was selected as
a reference standard as it is a widely used quantitative metric for Crohn’s disease
activity.
Histopathological assessment of inflammation
Each terminal ileal biopsy was stained with haematoxylin-eosin and reviewed in
face-to-face consensus by two experienced pathologists (MRJ and LAB) with 10
and 15 years respectively, blinded to clinical information other than the diagnosis
of CD. The endoscopic biopsy of acute inflammation score (eAIS) was used to
semi-objectively evaluate typical morphological features associated with Crohn’s
disease activity as previously described [24,25]. The eAIS score includes measures
Proefschrift2018-new2.indb 129 13/9/18 10:06
Chapter 6
130
of epithelial damage, architectural changes in the mucosa, epithelial neutrophils,
erosion/ulceration and the presence of granulomas (Appendix 2). The biopsy with
the highest score was used to grade the level of inflammation for each patient. The
eAIS was selected as reference standard as it is an alternative quantitative metric to
CDEIS for measuring disease activity.
Motility assessment
Terminal ileal motility was quantified using the methodology described by Menys et
al. [15]. Specifically, the dynamic series for each patient were processed by our study
coordinator using a previously validated registration algorithm designed for the
assessment of bowel motility [13]. Deformation fields generated by the registration
were used to produce a surrogate motility metric defined as the standard deviation
of the deformation fields’ Jacobian determinant (SD Jacobian) in the terminal ileum,
with a score of 0 representing no motility. A single observer (GB –independent to the
observers who derived the MaRIA score) was presented with a single ‘registration
target image’ (selected automatically by the registration technique) containing the
terminal ileum and was otherwise blind to the motility data (including the cine series
& motility maps) and all clinical data. The observer used the anatomical images for
reference and manually placed a polygonal region of interest within the last 5cm of
terminal ileum in each patient to encompass the bowel wall and lumen (Figure 1).
The ROI was automatically applied to the parametric motility map of SD Jacobian
values and the average pixel value taken from that map to create the terminal ileal
motility score for that patient in arbitrary units (AU). Because previous data has
already shown good intra observer agreement for this observer [26], this was not
repeated for the current study.
Magnetic Resonance Index of Activity (MaRIA)
The terminal ileal MaRIA score within 5cm of the ileocecal valve was calculated
independently by two radiologists for subjects scanned at their respective
institutions (33 subjects by DP, 10 years’ experience at Center 1 & 49 subjects by
CN, 4 years’ experience at Center 2) using the standard anatomical images, as
described by Rimola et al. [4] and using the following formula:
MaRIA = 1.5 x wall thickness (mm) + 0.02 x Relative contrast enhancement + 5 x
edema + 10 x ulcers
Proefschrift2018-new2.indb 130 13/9/18 10:06
Chapter 6
130
of epithelial damage, architectural changes in the mucosa, epithelial neutrophils,
erosion/ulceration and the presence of granulomas (Appendix 2). The biopsy with
the highest score was used to grade the level of inflammation for each patient. The
eAIS was selected as reference standard as it is an alternative quantitative metric to
CDEIS for measuring disease activity.
Motility assessment
Terminal ileal motility was quantified using the methodology described by Menys et
al. [15]. Specifically, the dynamic series for each patient were processed by our study
coordinator using a previously validated registration algorithm designed for the
assessment of bowel motility [13]. Deformation fields generated by the registration
were used to produce a surrogate motility metric defined as the standard deviation
of the deformation fields’ Jacobian determinant (SD Jacobian) in the terminal ileum,
with a score of 0 representing no motility. A single observer (GB –independent to the
observers who derived the MaRIA score) was presented with a single ‘registration
target image’ (selected automatically by the registration technique) containing the
terminal ileum and was otherwise blind to the motility data (including the cine series
& motility maps) and all clinical data. The observer used the anatomical images for
reference and manually placed a polygonal region of interest within the last 5cm of
terminal ileum in each patient to encompass the bowel wall and lumen (Figure 1).
The ROI was automatically applied to the parametric motility map of SD Jacobian
values and the average pixel value taken from that map to create the terminal ileal
motility score for that patient in arbitrary units (AU). Because previous data has
already shown good intra observer agreement for this observer [26], this was not
repeated for the current study.
Magnetic Resonance Index of Activity (MaRIA)
The terminal ileal MaRIA score within 5cm of the ileocecal valve was calculated
independently by two radiologists for subjects scanned at their respective
institutions (33 subjects by DP, 10 years’ experience at Center 1 & 49 subjects by
CN, 4 years’ experience at Center 2) using the standard anatomical images, as
described by Rimola et al. [4] and using the following formula:
MaRIA = 1.5 x wall thickness (mm) + 0.02 x Relative contrast enhancement + 5 x
edema + 10 x ulcers
Proefschrift2018-new2.indb 130 13/9/18 10:06
Terminal ileal motility during MR enterography
131
Figure 1. Anatomical reference images (A – Healthy & C - Diseased) as seen with MRI and
corresponding motility maps (B & D). The region of interest at the terminal ileum is indicated
along with the motility score depicting high (B) and low (D) motility.
Statistical analysis
The primary analysis tested the ability of motility scores below a predefined cut
off to detect active inflammation using both and endoscopic and histological
standard of reference. A secondary analysis assessed the ability of motility values
to quantify the severity of inflammation (again, judged using both endoscopy and
histological scoring systems). The same analyses were calculated for the MaRIA
score and compared to those of the motility metric. A pre-defined motility metric
Proefschrift2018-new2.indb 131 13/9/18 10:06
Terminal ileal motility during MR enterography
131
Figure 1. Anatomical reference images (A – Healthy & C - Diseased) as seen with MRI and
corresponding motility maps (B & D). The region of interest at the terminal ileum is indicated
along with the motility score depicting high (B) and low (D) motility.
Statistical analysis
The primary analysis tested the ability of motility scores below a predefined cut
off to detect active inflammation using both and endoscopic and histological
standard of reference. A secondary analysis assessed the ability of motility values
to quantify the severity of inflammation (again, judged using both endoscopy and
histological scoring systems). The same analyses were calculated for the MaRIA
score and compared to those of the motility metric. A pre-defined motility metric
Proefschrift2018-new2.indb 131 13/9/18 10:06
Chapter 6
132
cut-off of 0.30 AU was applied to define the presence of active inflammation. This
threshold was pre-specified using the optimal cut-point from a previously-published
retrospective cohort [15] (Appendix 3).
The sensitivity and specificity of a motility score of <0.30 AU for the presence
of active inflammation, defined as either an CDEIS of >=4 (endoscopic standard
of reference) or eAIS score of >= 1 (histopathological standard of reference) was
calculated. A CDEIS cut off of >=4 was chosen given is utility as marker of mucosal
healing [7]. The above analysis was repeated using the MaRIA score. A MaRIA score
of <7 has been previously associated with mucosal healing (>=7 representing active
disease) and a score of <11 has been associated with ulcer healing (>= 11 representing
the presence of ulceration) [4]. Both cut-offs were tested in the analysis.
The differences in sensitivity and specificity between motility and the MaRIA scores
was assessed by McNemar’s test where the null hypothesis was no difference
existed between the two scores. Thereafter, linear correlation between motility and
the MaRIA scores, and both eAIS and CDEIS was performed with Spearman’s Rho.
Finally, receiver operating characteristic (ROC) curves of motility and the MaRIA
score to detect inflammation based on both eAIS and CDEIs were constructed and
area under the ROC curve (AUC) calculated. The AUC for motility and MaRIA
against both standard of reference were compared (DeLong method). The optimal
motility and MaRIA cut offs were derived from the ROC curves for the current study
data automatically (Top-Left method). All statistical analysis was performed in R
2018 (Vienna, Austria) and ROC analysis was performed with the pROC package in
R [27].
The sensitivity and specificity of both the motility score and MaRIA against the
endoscopic and histopathological standard was assessed separately according to
recruitment site.
RESULTS
Overall 24 subjects were further excluded from our study because of missed
terminal ileum on dynamic sequences (n = 11), poor motility scan quality (n = 6) or
recent use of motility altering medication; pro-kinetics and opioid analgesia (n = 7)
Proefschrift2018-new2.indb 132 13/9/18 10:06
Chapter 6
132
cut-off of 0.30 AU was applied to define the presence of active inflammation. This
threshold was pre-specified using the optimal cut-point from a previously-published
retrospective cohort [15] (Appendix 3).
The sensitivity and specificity of a motility score of <0.30 AU for the presence
of active inflammation, defined as either an CDEIS of >=4 (endoscopic standard
of reference) or eAIS score of >= 1 (histopathological standard of reference) was
calculated. A CDEIS cut off of >=4 was chosen given is utility as marker of mucosal
healing [7]. The above analysis was repeated using the MaRIA score. A MaRIA score
of <7 has been previously associated with mucosal healing (>=7 representing active
disease) and a score of <11 has been associated with ulcer healing (>= 11 representing
the presence of ulceration) [4]. Both cut-offs were tested in the analysis.
The differences in sensitivity and specificity between motility and the MaRIA scores
was assessed by McNemar’s test where the null hypothesis was no difference
existed between the two scores. Thereafter, linear correlation between motility and
the MaRIA scores, and both eAIS and CDEIS was performed with Spearman’s Rho.
Finally, receiver operating characteristic (ROC) curves of motility and the MaRIA
score to detect inflammation based on both eAIS and CDEIs were constructed and
area under the ROC curve (AUC) calculated. The AUC for motility and MaRIA
against both standard of reference were compared (DeLong method). The optimal
motility and MaRIA cut offs were derived from the ROC curves for the current study
data automatically (Top-Left method). All statistical analysis was performed in R
2018 (Vienna, Austria) and ROC analysis was performed with the pROC package in
R [27].
The sensitivity and specificity of both the motility score and MaRIA against the
endoscopic and histopathological standard was assessed separately according to
recruitment site.
RESULTS
Overall 24 subjects were further excluded from our study because of missed
terminal ileum on dynamic sequences (n = 11), poor motility scan quality (n = 6) or
recent use of motility altering medication; pro-kinetics and opioid analgesia (n = 7)
Proefschrift2018-new2.indb 132 13/9/18 10:06
Terminal ileal motility during MR enterography
133
(Figure 2). 82 subjects were eligible for our study (33 Center 1 and 49 Center 2),
median age 31 (range 16 to 70) with demographics shown in Table 1.
Figure 2. Flow of participants
Detection of infl ammation
Endoscopic reference
The sensitivity and specificity of motility and the MaRIA for identifying endoscopic
activity (CDEIS >= 4) is shown in Table 2. The median CDEIS across the cohort was
6 (range 0 to 36).
Motility achieved a greater sensitivity for active disease (92.7%) than MaRIA (78.0%)
using a cut-off of >11 (P = 0.03). The sensitivity of motility was not different to
MaRIA (85.3%) when using a cut off >7 (P = 0.26). Overall, motility identified 38 of
41 subjects with active disease based on CDEIS compared to 35 and 32 of 41 for
MaRIA >7 and >11 respectively. The specificity of motility (61.0%) was lower than
MaRIA (80.5%) most notably at the <11 cut-off (P = 0.04) and also at the >7 cut-off
(75.6%, P = 0.05).
Histopathological reference
The sensitivity and specificity of motility and the MaRIA for identifying
histopathological inflammation (eAIS ≥1) is shown in Table 3. The median eAIS
across the cohort was 1 (range 0 to 4). Motility achieved higher sensitivity (91.7)
Proefschrift2018-new2.indb 133 13/9/18 10:06
Terminal ileal motility during MR enterography
133
(Figure 2). 82 subjects were eligible for our study (33 Center 1 and 49 Center 2),
median age 31 (range 16 to 70) with demographics shown in Table 1.
Figure 2. Flow of participants
Detection of infl ammation
Endoscopic reference
The sensitivity and specificity of motility and the MaRIA for identifying endoscopic
activity (CDEIS >= 4) is shown in Table 2. The median CDEIS across the cohort was
6 (range 0 to 36).
Motility achieved a greater sensitivity for active disease (92.7%) than MaRIA (78.0%)
using a cut-off of >11 (P = 0.03). The sensitivity of motility was not different to
MaRIA (85.3%) when using a cut off >7 (P = 0.26). Overall, motility identified 38 of
41 subjects with active disease based on CDEIS compared to 35 and 32 of 41 for
MaRIA >7 and >11 respectively. The specificity of motility (61.0%) was lower than
MaRIA (80.5%) most notably at the <11 cut-off (P = 0.04) and also at the >7 cut-off
(75.6%, P = 0.05).
Histopathological reference
The sensitivity and specificity of motility and the MaRIA for identifying
histopathological inflammation (eAIS ≥1) is shown in Table 3. The median eAIS
across the cohort was 1 (range 0 to 4). Motility achieved higher sensitivity (91.7)
Proefschrift2018-new2.indb 133 13/9/18 10:06
Chapter 6
134
than MaRIA (>7; 75.0% %, P = 0.03) and >11; 70.8%, P = 0.006) at both cut offs.
The specificity of motility (70.5%) was similar to MaRIA >7 (73.5%, P = 1) and at
the >11 cut-off (82.3%, P = 0.25). Overall, motility identified 44 of 48 subjects with
active disease based on eAIS compared to 36 and 34 of 48 for MaRIA >7 and >11
respectively).
Table 1. Patient demographics
Center 1 (n = 33) Center 2 (n = 49)Males (%) 12 (36%) 30 (50%)
Median age (range) 27 years (16 to 63 years) 35 years (19 to 70 years)
Median age females (range) 29 years (16 to 63 years) 35 years (20 to 59 years)
Median age males (range) 25 years (17 to 43 years) 36 years (19 to 70 years)
Median age at diagnosis (range) 22 years (6 to 53 years) 24 years (13 to 56 years)
Median disease duration (range) 4 years (0 to 30 years) 8 years (0 to 42 years)
Montreal Disease Location
L1 (%) 7 (21%) 27 (55%)
L2 (%) 5 (15%) 5 (10%)
L3 (%) 21 (64%) 17 (35%)
Number of previous surgical procedures
0 (%) 26 (84%) 30 (61%)
1 (%) 5 (16%) 1 (2%)
2 (%) - 2 (4%)
3+ (%) - 16 (33%)
Medication
Steroid 2 (7%) 10 (20%)
5-aminosalicylic (ASA) acid or Immuno-modulator*
15 (50%) 12 (25%)
Biologic 6 (20%) 16 (33%)
Median C-Reactive Protein (range) (Normal <5)
2 (0 to 68) 4 (0 to 40)
Median Harvey-Bradshaw Index (range) 4 (0 to 9) 6 (0 to 38)
Median CDEIS** (range) 0 (0 to 31) 9 (0 to 36)
0 24 21
>7 2 9
>11 7 19
Median eAIS*** (range) 0 (0 to 4) 2 (0 to 4)
0 17 17
1 6 2
2 6 17
3 1 6
4 3 7
Anti-Diarrhoeal (Loperimide) 1 3
* Immunomodulators including methotrexate, azathioprine and 6-mercaptopurine. ** Crohn’s Disease Endoscopic Index of Severity. *** Endoscopic Acute Index of Severity
Proefschrift2018-new2.indb 134 13/9/18 10:06
Chapter 6
134
than MaRIA (>7; 75.0% %, P = 0.03) and >11; 70.8%, P = 0.006) at both cut offs.
The specificity of motility (70.5%) was similar to MaRIA >7 (73.5%, P = 1) and at
the >11 cut-off (82.3%, P = 0.25). Overall, motility identified 44 of 48 subjects with
active disease based on eAIS compared to 36 and 34 of 48 for MaRIA >7 and >11
respectively).
Table 1. Patient demographics
Center 1 (n = 33) Center 2 (n = 49)Males (%) 12 (36%) 30 (50%)
Median age (range) 27 years (16 to 63 years) 35 years (19 to 70 years)
Median age females (range) 29 years (16 to 63 years) 35 years (20 to 59 years)
Median age males (range) 25 years (17 to 43 years) 36 years (19 to 70 years)
Median age at diagnosis (range) 22 years (6 to 53 years) 24 years (13 to 56 years)
Median disease duration (range) 4 years (0 to 30 years) 8 years (0 to 42 years)
Montreal Disease Location
L1 (%) 7 (21%) 27 (55%)
L2 (%) 5 (15%) 5 (10%)
L3 (%) 21 (64%) 17 (35%)
Number of previous surgical procedures
0 (%) 26 (84%) 30 (61%)
1 (%) 5 (16%) 1 (2%)
2 (%) - 2 (4%)
3+ (%) - 16 (33%)
Medication
Steroid 2 (7%) 10 (20%)
5-aminosalicylic (ASA) acid or Immuno-modulator*
15 (50%) 12 (25%)
Biologic 6 (20%) 16 (33%)
Median C-Reactive Protein (range) (Normal <5)
2 (0 to 68) 4 (0 to 40)
Median Harvey-Bradshaw Index (range) 4 (0 to 9) 6 (0 to 38)
Median CDEIS** (range) 0 (0 to 31) 9 (0 to 36)
0 24 21
>7 2 9
>11 7 19
Median eAIS*** (range) 0 (0 to 4) 2 (0 to 4)
0 17 17
1 6 2
2 6 17
3 1 6
4 3 7
Anti-Diarrhoeal (Loperimide) 1 3
* Immunomodulators including methotrexate, azathioprine and 6-mercaptopurine. ** Crohn’s Disease Endoscopic Index of Severity. *** Endoscopic Acute Index of Severity
Proefschrift2018-new2.indb 134 13/9/18 10:06
Terminal ileal motility during MR enterography
135
Table 2. Sensitivity and specificity of Motility and MaRIA for active disease based on an
endoscopic standard of reference (CDEIS) at the two cut-offs. TP = Tue Positive, FN = False
Negative, TN = True Negative, FP = False Positive).
Sensitivityn=82(TP+FN)
95% Con-fidence interval
Significance (P) against motility
Specificity n=82(TN+FP)
95% Con-fidence Interval
Significance (P) against motility
Motility <0.30 92.7 (38+3) 80.5 to 98.5 - 61.0 (25+16) 44.5 to 75.8 -
MaRIA >7 85.3 (35+6) 70.8 to 94.4 0.26 75.6 (31+10) 59.7 to 87.6 0.05
MaRIA >11 78.0 (32+9) 62.4 to 89.4 0.03 80.5 (33+8) 65.1 to 91.2 0.04
Table 3. Sensitivity and specificity of Motility and MaRIA for active disease based on a
histopathological standard of reference (eAIS ≥1). TP = Tue Positive, FN = False Negative, TN
= True Negative, FP = False Positive).
Sensitivityn=82(TP+FN)
95% Con-fidence interval
Significance (P) against motility
Specificity n=82(TN+FP)
95% Con-fidence Interval
Significance (P) against motility
Motility <0.30 91.7 (44+4) 80.0 to 97.7 - 70.5 (24+10) 55.6 to 87.1 -
MaRIA >7 75.0 (36+12) 60.4 to 86.3 0.03 73.5 (25+9) 55.6 to 77.1 1.0
MaRIA >11 70.8 (34+14) 55.9 to 83.1 0.006 82.3 (28+6) 65.5 to 93.2 0.25
Quantification of inflammation
Correlation of CDEIS and eAIS produced a strong positive correlation (Spearman’s
r = 0.75 [95% CI 0.64 to 0.83]) (Appendix 6). Motility demonstrated a moderate
negative correlation against the CDEIS score (r = -0.59, 95% CI -0.72 to -0.43), while
MaRIA showed a strong positive correlation with CDEIS (r = 0.71, 95% CI 0.58 to
0.80) (Figure 3). Motility had a moderate negative relationship with the eAIS score
(r = -0.61, 95% CI -0.73 to -0.45), while MaRIA had a moderate positive correlation
with eAIS (r = 0.54, 95% CI 0.37 to 0.68) (Figure 4). ROC curves generated for
endoscopic grading (CDEIS >=4) and histopathological grading of activity (eAIS
≥1) presented in Figure 5A and 5B respectively. Against CDEIS, motility produced
an AUC of 0.84 (95% CI 0.72 to 0.92) and MaRIA generated an AUC of 0.84 (95%
CI 0.75 to 0.93). The two ROC curves were not significantly different (p = 0.72).
Against eAIS, motility produced an AUC value of 0.87 (95% CI 0.79 to 0.96) and
MaRIA 0.78 (95% CI 0.68 to 0.89). The two ROC curves were not significantly
different (p = 0.125).In this prospective data, the optimal cut off point for motility
Proefschrift2018-new2.indb 135 13/9/18 10:06
Terminal ileal motility during MR enterography
135
Table 2. Sensitivity and specificity of Motility and MaRIA for active disease based on an
endoscopic standard of reference (CDEIS) at the two cut-offs. TP = Tue Positive, FN = False
Negative, TN = True Negative, FP = False Positive).
Sensitivityn=82(TP+FN)
95% Con-fidence interval
Significance (P) against motility
Specificity n=82(TN+FP)
95% Con-fidence Interval
Significance (P) against motility
Motility <0.30 92.7 (38+3) 80.5 to 98.5 - 61.0 (25+16) 44.5 to 75.8 -
MaRIA >7 85.3 (35+6) 70.8 to 94.4 0.26 75.6 (31+10) 59.7 to 87.6 0.05
MaRIA >11 78.0 (32+9) 62.4 to 89.4 0.03 80.5 (33+8) 65.1 to 91.2 0.04
Table 3. Sensitivity and specificity of Motility and MaRIA for active disease based on a
histopathological standard of reference (eAIS ≥1). TP = Tue Positive, FN = False Negative, TN
= True Negative, FP = False Positive).
Sensitivityn=82(TP+FN)
95% Con-fidence interval
Significance (P) against motility
Specificity n=82(TN+FP)
95% Con-fidence Interval
Significance (P) against motility
Motility <0.30 91.7 (44+4) 80.0 to 97.7 - 70.5 (24+10) 55.6 to 87.1 -
MaRIA >7 75.0 (36+12) 60.4 to 86.3 0.03 73.5 (25+9) 55.6 to 77.1 1.0
MaRIA >11 70.8 (34+14) 55.9 to 83.1 0.006 82.3 (28+6) 65.5 to 93.2 0.25
Quantification of inflammation
Correlation of CDEIS and eAIS produced a strong positive correlation (Spearman’s
r = 0.75 [95% CI 0.64 to 0.83]) (Appendix 6). Motility demonstrated a moderate
negative correlation against the CDEIS score (r = -0.59, 95% CI -0.72 to -0.43), while
MaRIA showed a strong positive correlation with CDEIS (r = 0.71, 95% CI 0.58 to
0.80) (Figure 3). Motility had a moderate negative relationship with the eAIS score
(r = -0.61, 95% CI -0.73 to -0.45), while MaRIA had a moderate positive correlation
with eAIS (r = 0.54, 95% CI 0.37 to 0.68) (Figure 4). ROC curves generated for
endoscopic grading (CDEIS >=4) and histopathological grading of activity (eAIS
≥1) presented in Figure 5A and 5B respectively. Against CDEIS, motility produced
an AUC of 0.84 (95% CI 0.72 to 0.92) and MaRIA generated an AUC of 0.84 (95%
CI 0.75 to 0.93). The two ROC curves were not significantly different (p = 0.72).
Against eAIS, motility produced an AUC value of 0.87 (95% CI 0.79 to 0.96) and
MaRIA 0.78 (95% CI 0.68 to 0.89). The two ROC curves were not significantly
different (p = 0.125).In this prospective data, the optimal cut off point for motility
Proefschrift2018-new2.indb 135 13/9/18 10:06
Chapter 6
136
detecting disease activity against CDEIS is 0.23 and against eAIS is 0.22 based on
the current study data shown in Appendix 4. The optimal cut off point for MaRIA
against CDEIS was 9.9 and against eAIS was also 9.9. The sensitivity of Motility and
MaRIA against CDEIS and eAIS according to recruitment site is shown in Appendix
5.
Figure 3. Correlation (Spearman’s Rank) between motility and CDEIS (A) and MaRIA and
CDEIS (B).
Figure 4. Correlation (Spearman’s Rank) between motility and eAIS (A) and MaRIA and eAIS
(B).
Proefschrift2018-new2.indb 136 13/9/18 10:06
Chapter 6
136
detecting disease activity against CDEIS is 0.23 and against eAIS is 0.22 based on
the current study data shown in Appendix 4. The optimal cut off point for MaRIA
against CDEIS was 9.9 and against eAIS was also 9.9. The sensitivity of Motility and
MaRIA against CDEIS and eAIS according to recruitment site is shown in Appendix
5.
Figure 3. Correlation (Spearman’s Rank) between motility and CDEIS (A) and MaRIA and
CDEIS (B).
Figure 4. Correlation (Spearman’s Rank) between motility and eAIS (A) and MaRIA and eAIS
(B).
Proefschrift2018-new2.indb 136 13/9/18 10:06
Terminal ileal motility during MR enterography
137
Figure 5. Motility (solid black line) and MaRIA (dotted line) ROC curves are presented for
endoscopic y (A) and histopathological activity (B).
DISCUSSION
In our study, we examined the relationship between terminal ileal motility and
Crohn’s disease activity assessed by both an endoscopic severity index (CDEIS)
and histopathology score (eAIS). Importantly, we used a predefined motility cut off
for activity based on previous work, and tested it prospectively in subjects from two
separate centers. We found that reduced motility (of <0.30AU) was highly sensitive
(92%/91.7%) at the cost of modest specificity (61.0%/70.5%) for both endoscopically-
defined and histologically-defined active inflammation respectively. As noted, this
diagnostic performance was achieved with a pre-specified threshold for a positive
test (derived from a previous retrospective cohort), and therefore an important
step in validating the technique. Indeed, the sensitivity of motility reduction for
inflammation was greater than that in the previous retrospective evaluation [15],
likely due to our prospective study design using a consistent bowel preparation,
and supervision of the imaging protocol by a study scientist. Motility reduction was
significantly more sensitive than a previously-validated activity score (MaRIA) for
histopathologic inflammation using previously-published thresholds7, suggesting
it may serve as a rapid, reproducible and simple means of excluding terminal ileal
inflammation in patients with known CD. Against an endoscopic activity score,
motility had similar sensitivity (92.7%) to the MaRIA score (using a cut off of <7)
Proefschrift2018-new2.indb 137 13/9/18 10:06
Terminal ileal motility during MR enterography
137
Figure 5. Motility (solid black line) and MaRIA (dotted line) ROC curves are presented for
endoscopic y (A) and histopathological activity (B).
DISCUSSION
In our study, we examined the relationship between terminal ileal motility and
Crohn’s disease activity assessed by both an endoscopic severity index (CDEIS)
and histopathology score (eAIS). Importantly, we used a predefined motility cut off
for activity based on previous work, and tested it prospectively in subjects from two
separate centers. We found that reduced motility (of <0.30AU) was highly sensitive
(92%/91.7%) at the cost of modest specificity (61.0%/70.5%) for both endoscopically-
defined and histologically-defined active inflammation respectively. As noted, this
diagnostic performance was achieved with a pre-specified threshold for a positive
test (derived from a previous retrospective cohort), and therefore an important
step in validating the technique. Indeed, the sensitivity of motility reduction for
inflammation was greater than that in the previous retrospective evaluation [15],
likely due to our prospective study design using a consistent bowel preparation,
and supervision of the imaging protocol by a study scientist. Motility reduction was
significantly more sensitive than a previously-validated activity score (MaRIA) for
histopathologic inflammation using previously-published thresholds7, suggesting
it may serve as a rapid, reproducible and simple means of excluding terminal ileal
inflammation in patients with known CD. Against an endoscopic activity score,
motility had similar sensitivity (92.7%) to the MaRIA score (using a cut off of <7)
Proefschrift2018-new2.indb 137 13/9/18 10:06
Chapter 6
138
although lower specificity (61%). Comparison of ROC curves of motility and MaRIA
in against the endoscopic and histopathological standards respectively showed no
significant differences in AUC value suggesting diagnostic performance for disease
activity is governed by choice of cut off value. For example, based on the data in
the current study, against CDEIS, a motility cut off of 0.23 AU gave 92% sensitivity
for active disease and 76% specificity, where as a MaRIA cut off of 9.9 gave 83%
sensitivity and 81 % specificity (Appendix 3). These cuts off differ a little from those
we prospectively defined for testing (motility AU <0.30AU and MaRIA <7 and <11).
Binary detection of activity is only one facet of CD assessment, and quantifying
severity is also important. Indeed, it is important that imaging indicators of activity
are reliable across the whole range of disease activity. We found a moderate
negative correlation between the eAIS and motility score that was similar to that
described in Menys et al and Cullman et al, and also similar to that achieved by
the MaRIA score [15,16]. To the best of our knowledge, this is the first time these
observations have been reproduced prospectively in a multi-institution setting. It
is perhaps intuitive that as the bowel wall thickens in response to inflammatory
activity, the contractile potential of the bowel should decrease. Indeed, this has
previously been observed in retrospective studies against histopathological and
structural markers of disease activity [15]. Motility also had a moderate negative
correlation with endoscopic scored activity, although a little lower than that of
MaRIA. The differing diagnostic performance of the two MRI techniques may reflect
their derivation. MaRIA has been developed and validated again endoscopy, while
motility has been predominantly validated against histopathology. The tested
motility threshold for activity was defined based on previous a comparison with the
eAIS, and its diagnostic performance against CDEIS is improved by a small change
in this threshold to <0.23 (based on the current study data). Nonetheless, these
data strongly suggest that motility can be used both for disease detection and
activity quantification. In this study we used a CDEIS cut off of 4 to define active
disease as such as score has been advocated as the best predictor of mucosal
healing [7]. However, we acknowledge that a CDEIS of below 3 has also been used
to define inactive disease. Overall in our cohort just a single patient had a CDEIS of
3 so our cut off is will not impact on the overall study findings.
Motility assessment has two key attributes that make it a potentially valuable tool.
Firstly, it describes a functional aspect of gut physiology that is not conveyed through
Proefschrift2018-new2.indb 138 13/9/18 10:06
Chapter 6
138
although lower specificity (61%). Comparison of ROC curves of motility and MaRIA
in against the endoscopic and histopathological standards respectively showed no
significant differences in AUC value suggesting diagnostic performance for disease
activity is governed by choice of cut off value. For example, based on the data in
the current study, against CDEIS, a motility cut off of 0.23 AU gave 92% sensitivity
for active disease and 76% specificity, where as a MaRIA cut off of 9.9 gave 83%
sensitivity and 81 % specificity (Appendix 3). These cuts off differ a little from those
we prospectively defined for testing (motility AU <0.30AU and MaRIA <7 and <11).
Binary detection of activity is only one facet of CD assessment, and quantifying
severity is also important. Indeed, it is important that imaging indicators of activity
are reliable across the whole range of disease activity. We found a moderate
negative correlation between the eAIS and motility score that was similar to that
described in Menys et al and Cullman et al, and also similar to that achieved by
the MaRIA score [15,16]. To the best of our knowledge, this is the first time these
observations have been reproduced prospectively in a multi-institution setting. It
is perhaps intuitive that as the bowel wall thickens in response to inflammatory
activity, the contractile potential of the bowel should decrease. Indeed, this has
previously been observed in retrospective studies against histopathological and
structural markers of disease activity [15]. Motility also had a moderate negative
correlation with endoscopic scored activity, although a little lower than that of
MaRIA. The differing diagnostic performance of the two MRI techniques may reflect
their derivation. MaRIA has been developed and validated again endoscopy, while
motility has been predominantly validated against histopathology. The tested
motility threshold for activity was defined based on previous a comparison with the
eAIS, and its diagnostic performance against CDEIS is improved by a small change
in this threshold to <0.23 (based on the current study data). Nonetheless, these
data strongly suggest that motility can be used both for disease detection and
activity quantification. In this study we used a CDEIS cut off of 4 to define active
disease as such as score has been advocated as the best predictor of mucosal
healing [7]. However, we acknowledge that a CDEIS of below 3 has also been used
to define inactive disease. Overall in our cohort just a single patient had a CDEIS of
3 so our cut off is will not impact on the overall study findings.
Motility assessment has two key attributes that make it a potentially valuable tool.
Firstly, it describes a functional aspect of gut physiology that is not conveyed through
Proefschrift2018-new2.indb 138 13/9/18 10:06
Terminal ileal motility during MR enterography
139
other existing parameters based on bowel structure. For example, endoscopy
evaluates ulceration and mucosal damage, histopathology looks at structural
and cellular changes in the mucosa, and cross-sectional imaging primarily details
morphological changes driven by fibrosis and inflammation (such as bowel wall
thickening and contrast enhancement). Motility is an important aspect of normal
physiology and our data suggests it provides at least comparable accuracy as MRI
assessment of bowel structure. Indeed, recent data demonstrates its potential value
as a biomarker to indicate early response to biological therapy [6] and beyond this,
a link between aberrant motility in normal bowel and patient symptom load has
been described [28,29]. Crucially in terms of clinical uptake, motility sequences
can be easily added to existing MRE protocols with a small time penalty. This raises
the intriguing possibility of combining structural and functional assessments into a
single, combined index, which would draw on the strengths of both (for example,
using elements of structural MRI activity scores in combination with quantified
motility assessment). Secondly, from a clinical and research perspective, segmental
motility assessment is quantitative, objective and relatively easy to perform,
requiring a single ROI at the area of interest, placement of which generally takes a
few seconds. A high level of inter-observer agreement for the technique has been
demonstrated [6,12]. We acquired motility data at 1.1s temporal resolution. Although
it is technically feasible to acquire data as a lower temporal resolution, in fact
recent work has shown that in terms of motility metrics, such rapid acquisitions
hold no advantage over data acquired at 1 image per second [30]. Existing MRI
activity scores based on bowel structure, although reproducible [10] are time
consuming, particularly those requiring drawing of multiple ROIs, which limit use in
routine clinical practice. A disadvantage of motility analysis is that specialized post
processing software is needed to analyze the data, although this is increasingly
available to the clinical community.
Our study does have limitations. We used a breath-hold protocol that may only
capture a ‘snap shot’ of bowel motility. Whilst motility is reduced in active disease,
low motility does not always represent inflammatory burden (i.e. specificity was
moderate), which in part might be due to imaging occurring during a physiological
“quiescent” phase of contractility. A future approach might be to acquire for an
extended duration to average out transient variability in motility patterns. Although
our study was prospective, a time delay was still present between biopsy and MRI.
Although the majority of the scans were within 5 days of the biopsy, the acute
Proefschrift2018-new2.indb 139 13/9/18 10:06
Terminal ileal motility during MR enterography
139
other existing parameters based on bowel structure. For example, endoscopy
evaluates ulceration and mucosal damage, histopathology looks at structural
and cellular changes in the mucosa, and cross-sectional imaging primarily details
morphological changes driven by fibrosis and inflammation (such as bowel wall
thickening and contrast enhancement). Motility is an important aspect of normal
physiology and our data suggests it provides at least comparable accuracy as MRI
assessment of bowel structure. Indeed, recent data demonstrates its potential value
as a biomarker to indicate early response to biological therapy [6] and beyond this,
a link between aberrant motility in normal bowel and patient symptom load has
been described [28,29]. Crucially in terms of clinical uptake, motility sequences
can be easily added to existing MRE protocols with a small time penalty. This raises
the intriguing possibility of combining structural and functional assessments into a
single, combined index, which would draw on the strengths of both (for example,
using elements of structural MRI activity scores in combination with quantified
motility assessment). Secondly, from a clinical and research perspective, segmental
motility assessment is quantitative, objective and relatively easy to perform,
requiring a single ROI at the area of interest, placement of which generally takes a
few seconds. A high level of inter-observer agreement for the technique has been
demonstrated [6,12]. We acquired motility data at 1.1s temporal resolution. Although
it is technically feasible to acquire data as a lower temporal resolution, in fact
recent work has shown that in terms of motility metrics, such rapid acquisitions
hold no advantage over data acquired at 1 image per second [30]. Existing MRI
activity scores based on bowel structure, although reproducible [10] are time
consuming, particularly those requiring drawing of multiple ROIs, which limit use in
routine clinical practice. A disadvantage of motility analysis is that specialized post
processing software is needed to analyze the data, although this is increasingly
available to the clinical community.
Our study does have limitations. We used a breath-hold protocol that may only
capture a ‘snap shot’ of bowel motility. Whilst motility is reduced in active disease,
low motility does not always represent inflammatory burden (i.e. specificity was
moderate), which in part might be due to imaging occurring during a physiological
“quiescent” phase of contractility. A future approach might be to acquire for an
extended duration to average out transient variability in motility patterns. Although
our study was prospective, a time delay was still present between biopsy and MRI.
Although the majority of the scans were within 5 days of the biopsy, the acute
Proefschrift2018-new2.indb 139 13/9/18 10:06
Chapter 6
140
nature of Crohn’s disease means there is a theoretical possibility of activity changes
between the two time-points. By matching motility and biopsy data to within 5cm
of the ICV and use of multiple biopsies we mitigated sampling errors to a reasonable
extent. We only used one observer to place ROIs for motility measurements. However
as noted above, there is good data supporting the reproducibility of motility
quantification between observers [12,26]. Although the current study was planned
as a sub study of the main VIGOR++ study, we were not able to perform the motility
analysis in some recruited subjects. This was due to poor bowel distension (n=6) or
the motility sequences being missed (n=15). Reassuringly however there were no
technical failures of the algorithm to process the data. In future work, standardised
volumetric protocols together with explicit radiographic technician training may
reduce the number of missing or incomplete motility sequences. The lack of a single
reference standard for CD disease activity is a problem for all work in the field. We
use two reasonable reference standards well described in the literature-histology
and endoscopy, although both have limitations. For example, histological scores are
subject to sampling errors from the site of biopsy. Furthermore, both scores assess
mainly the mucosal changes CD affects the full bowel wall thickness. It is very likely
fibrosis will affect motility as it does existing structural MRI activity scores. Indeed,
the influence of fibrosis on motility is an important consideration which we were
not able to address in the current study. In the absence of validated non-invasive
biomarkers of fibrosis, it is only possible to assess its influence by comparing to full
thickness histological sampling of surgical resection specimens. Collation of such
data will be important as development of motility analysis continues. However,
it should be acknowledged that use of surgical resection specimens introduces
spectrum bias as by definition patients are symptomatic and usually have failed
long term medical therapy. The aim of the current study was to evaluate motility
as a marker of disease activity and to this end we used recognised independent
scores of activity as our standard of reference (similar to how the MaRIA score
was validated). It maybe that the main clinical utility of motility will be in treatment
response assessment, where patients act as their own control and any change in
motility is due to changes in inflammatory activity [26]. The strong performance
against two separate disease activity reference standards is reassuring that
motility is indeed a biomarker of inflammation. Because data was obtained from a
larger study, a formal power calculation was not conducted. Whist we included a
reasonable number of datasets, future work should include consideration of study
power and affect size. Indeed, the current study provides important data to inform
Proefschrift2018-new2.indb 140 13/9/18 10:06
Chapter 6
140
nature of Crohn’s disease means there is a theoretical possibility of activity changes
between the two time-points. By matching motility and biopsy data to within 5cm
of the ICV and use of multiple biopsies we mitigated sampling errors to a reasonable
extent. We only used one observer to place ROIs for motility measurements. However
as noted above, there is good data supporting the reproducibility of motility
quantification between observers [12,26]. Although the current study was planned
as a sub study of the main VIGOR++ study, we were not able to perform the motility
analysis in some recruited subjects. This was due to poor bowel distension (n=6) or
the motility sequences being missed (n=15). Reassuringly however there were no
technical failures of the algorithm to process the data. In future work, standardised
volumetric protocols together with explicit radiographic technician training may
reduce the number of missing or incomplete motility sequences. The lack of a single
reference standard for CD disease activity is a problem for all work in the field. We
use two reasonable reference standards well described in the literature-histology
and endoscopy, although both have limitations. For example, histological scores are
subject to sampling errors from the site of biopsy. Furthermore, both scores assess
mainly the mucosal changes CD affects the full bowel wall thickness. It is very likely
fibrosis will affect motility as it does existing structural MRI activity scores. Indeed,
the influence of fibrosis on motility is an important consideration which we were
not able to address in the current study. In the absence of validated non-invasive
biomarkers of fibrosis, it is only possible to assess its influence by comparing to full
thickness histological sampling of surgical resection specimens. Collation of such
data will be important as development of motility analysis continues. However,
it should be acknowledged that use of surgical resection specimens introduces
spectrum bias as by definition patients are symptomatic and usually have failed
long term medical therapy. The aim of the current study was to evaluate motility
as a marker of disease activity and to this end we used recognised independent
scores of activity as our standard of reference (similar to how the MaRIA score
was validated). It maybe that the main clinical utility of motility will be in treatment
response assessment, where patients act as their own control and any change in
motility is due to changes in inflammatory activity [26]. The strong performance
against two separate disease activity reference standards is reassuring that
motility is indeed a biomarker of inflammation. Because data was obtained from a
larger study, a formal power calculation was not conducted. Whist we included a
reasonable number of datasets, future work should include consideration of study
power and affect size. Indeed, the current study provides important data to inform
Proefschrift2018-new2.indb 140 13/9/18 10:06
Terminal ileal motility during MR enterography
141
such considerations going forward. Finally, where we analysed data according
to recruitment site, the results suggested higher specificity for both motility and
MaRIA at Center 2 (AMC). This likely reflects the spectrum of disease across the two
sites. In particular, patients recruited from AMC tended to have more active disease
(median CDEIS 9, range 0 to 36) than those at Centre 1 (UCLH) (median CDEIS 0,
(range 0 to 31).
In summary, quantified motility is an objective biomarker of endoscopic and
histopathological inflammatory activity in Crohn’s disease and comparable
to previously validated MRI activity scores. It potentially has advantages over
existing MRI activity scores based on evaluation of bowel structure and contrast
enhancement.
Proefschrift2018-new2.indb 141 13/9/18 10:06
Terminal ileal motility during MR enterography
141
such considerations going forward. Finally, where we analysed data according
to recruitment site, the results suggested higher specificity for both motility and
MaRIA at Center 2 (AMC). This likely reflects the spectrum of disease across the two
sites. In particular, patients recruited from AMC tended to have more active disease
(median CDEIS 9, range 0 to 36) than those at Centre 1 (UCLH) (median CDEIS 0,
(range 0 to 31).
In summary, quantified motility is an objective biomarker of endoscopic and
histopathological inflammatory activity in Crohn’s disease and comparable
to previously validated MRI activity scores. It potentially has advantages over
existing MRI activity scores based on evaluation of bowel structure and contrast
enhancement.
Proefschrift2018-new2.indb 141 13/9/18 10:06
Chapter 6
142
REFERENCES1. Panes J, Bouhnik Y, Reinisch W, et al. Imaging techniques for assessment of inflammatory
bowel disease: joint ECCO and ESGAR evidence-based consensus guidelines. J Crohns
Colitis 2013;7(7):556–85.
2. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: derivation and histopathological validation of an MR-
based activity index. Eur J Radiol 2012;81(9):2080–8.
3. Makanyanga J, Dikaios N, Helbren E, et al. Evaluation of Crohn’s disease activity: Initial
validation of a magnetic resonance enterography global score (MEGS) against faecal
calprotectin. Eur. Radiol. 2014.
4. Rimola J, Ordás I, Rodriguez S, et al. Magnetic resonance imaging for evaluation of Crohn’s
disease: validation of parameters of severity and quantitative index of activity. Inflamm
Bowel Dis 2011;17(8):1759–68.
5. Taylor SA, Punwani S, Rodriguez-Justo M, et al. Mural Crohn disease: correlation of
dynamic contrast-enhanced MR imaging findings with angiogenesis and inflammation at
histologic examination--pilot study. Radiology 2009;251(2):369–79.
6. Plumb AA, Menys A, Russo E, et al. Magnetic resonance imaging-quantified small bowel
motility is a sensitive marker of response to medical therapy in Crohn’s disease. Aliment
Pharmacol Ther 2015;42(3):343-55.
7. Ordás I, Rimola J, Rodríguez S, et al. Accuracy of magnetic resonance enterography in
assessing response to therapy and mucosal healing in patients with Crohn’s disease.
Gastroenterology 2014;146(2):374–821.
8. Prezzi D, Bhatnagar G, Vega R, Makanyanga J, Halligan S, Taylor SA. Monitoring Crohn’s
disease during anti-TNF-therapy: validation of the magnetic resonance enterography
global score (MEGS) against a combined clinical reference standard. Eur Radiol
2015;26(7):2107-17
9. Tolan DJM, Greenhalgh R, Zealley IA, Halligan S, Taylor SA. MR Enterographic Manifestations
of Small Bowel Crohn Disease1. 2010;30(2):367–84.
10. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading Crohn disease activity with MRI:
interobserver variability of MRI features, MRI scoring of severity, and correlation with
Crohn disease endoscopic index of severity. AJR Am J Roentgenol 2013;201(6):1220–8.
11. Menys A, Hamy V, Makanyanga J, et al. Dual registration of abdominal motion for motility
assessment in free-breathing data sets acquired using dynamic MRI. Phys Med Biol
2014;59(16):4603–19.
12. Menys A, Taylor SA, Emmanuel A, et al. Global Small Bowel Motility: Assessment with
Dynamic MR Imaging. 2013;269(2):443–50.
13. Odille F, Menys A, Ahmed A, Punwani S, Taylor SA, Atkinson D. Quantitative assessment
of small bowel motility by nonrigid registration of dynamic MR images. Magn Reson Med
2012;68(3):783–93.
14. Hahnemann ML, Nensa F, Kinner S, Gerken G, Lauenstein TC. Motility mapping as
evaluation tool for bowel motility: Initial results on the development of an automated
color-coding algorithm in cine MRI. J Magn Reson Imaging 2014;41(2):354-60.
15. Menys A, Atkinson D, Odille F, et al. Quantified terminal ileal motility during MR
enterography as a potential biomarker of Crohn’s disease activity: a preliminary study.
Eur Radiol 2012;22(11):2494–501.
16. Cullmann JL, Bickelhaupt S, Froehlich JM, et al. MR imaging in Crohn’s disease: correlation
Proefschrift2018-new2.indb 142 13/9/18 10:06
Chapter 6
142
REFERENCES1. Panes J, Bouhnik Y, Reinisch W, et al. Imaging techniques for assessment of inflammatory
bowel disease: joint ECCO and ESGAR evidence-based consensus guidelines. J Crohns
Colitis 2013;7(7):556–85.
2. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: derivation and histopathological validation of an MR-
based activity index. Eur J Radiol 2012;81(9):2080–8.
3. Makanyanga J, Dikaios N, Helbren E, et al. Evaluation of Crohn’s disease activity: Initial
validation of a magnetic resonance enterography global score (MEGS) against faecal
calprotectin. Eur. Radiol. 2014.
4. Rimola J, Ordás I, Rodriguez S, et al. Magnetic resonance imaging for evaluation of Crohn’s
disease: validation of parameters of severity and quantitative index of activity. Inflamm
Bowel Dis 2011;17(8):1759–68.
5. Taylor SA, Punwani S, Rodriguez-Justo M, et al. Mural Crohn disease: correlation of
dynamic contrast-enhanced MR imaging findings with angiogenesis and inflammation at
histologic examination--pilot study. Radiology 2009;251(2):369–79.
6. Plumb AA, Menys A, Russo E, et al. Magnetic resonance imaging-quantified small bowel
motility is a sensitive marker of response to medical therapy in Crohn’s disease. Aliment
Pharmacol Ther 2015;42(3):343-55.
7. Ordás I, Rimola J, Rodríguez S, et al. Accuracy of magnetic resonance enterography in
assessing response to therapy and mucosal healing in patients with Crohn’s disease.
Gastroenterology 2014;146(2):374–821.
8. Prezzi D, Bhatnagar G, Vega R, Makanyanga J, Halligan S, Taylor SA. Monitoring Crohn’s
disease during anti-TNF-therapy: validation of the magnetic resonance enterography
global score (MEGS) against a combined clinical reference standard. Eur Radiol
2015;26(7):2107-17
9. Tolan DJM, Greenhalgh R, Zealley IA, Halligan S, Taylor SA. MR Enterographic Manifestations
of Small Bowel Crohn Disease1. 2010;30(2):367–84.
10. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading Crohn disease activity with MRI:
interobserver variability of MRI features, MRI scoring of severity, and correlation with
Crohn disease endoscopic index of severity. AJR Am J Roentgenol 2013;201(6):1220–8.
11. Menys A, Hamy V, Makanyanga J, et al. Dual registration of abdominal motion for motility
assessment in free-breathing data sets acquired using dynamic MRI. Phys Med Biol
2014;59(16):4603–19.
12. Menys A, Taylor SA, Emmanuel A, et al. Global Small Bowel Motility: Assessment with
Dynamic MR Imaging. 2013;269(2):443–50.
13. Odille F, Menys A, Ahmed A, Punwani S, Taylor SA, Atkinson D. Quantitative assessment
of small bowel motility by nonrigid registration of dynamic MR images. Magn Reson Med
2012;68(3):783–93.
14. Hahnemann ML, Nensa F, Kinner S, Gerken G, Lauenstein TC. Motility mapping as
evaluation tool for bowel motility: Initial results on the development of an automated
color-coding algorithm in cine MRI. J Magn Reson Imaging 2014;41(2):354-60.
15. Menys A, Atkinson D, Odille F, et al. Quantified terminal ileal motility during MR
enterography as a potential biomarker of Crohn’s disease activity: a preliminary study.
Eur Radiol 2012;22(11):2494–501.
16. Cullmann JL, Bickelhaupt S, Froehlich JM, et al. MR imaging in Crohn’s disease: correlation
Proefschrift2018-new2.indb 142 13/9/18 10:06
Terminal ileal motility during MR enterography
143
of MR motility measurement with histopathology in the terminal ileum. Neurogastroenterol
Motil 2013;25(9):749-e577.
17. Bickelhaupt S, Froehlich JM, Cattin R, et al. Differentiation between active and chronic
Crohn’s disease using MRI small-bowel motility examinations — Initial experience. Clin
Radiol 2013;68(12):1247–53.
18. Hahnemann ML, Nensa F, Kinner S, et al. Quantitative assessment of small bowel motility
in patients with Crohn’s disease using dynamic MRI. Neurogastroenterol Motil [Internet]
2015;27(6):841–8.
19. Li Z, Tielbeek JAW, Caan MWA, et al. Expiration-Phase Template-Based Motion Correction
of Free-Breathing Abdominal Dynamic Contrast Enhanced MRI. IEEE Trans Biomed Eng
2015;62(4):1215–25.
20. Naziroglu RE, Puylaert CAJ, Tielbeek JAW, et al. Semi-automatic bowel wall thickness
measurements on MR enterography in patients with Crohn’s disease. Br J Radiol [Internet]
2017;90(1074):20160654.
21. Puylaert CAJ, Schüffler PJ, Naziroglu RE, et al. Semiautomatic Assessment of the Terminal
Ileum and Colon in Patients with Crohn Disease Using MRI (the VIGOR++ Project). Acad
Radiol 2018.
22. Gomollón F, Dignass A, Annese V, et al. 3rd European Evidence-based Consensus on
the Diagnosis and Management of Crohn’s Disease 2016: Part 1: Diagnosis and Medical
Management. J Crohn’s Colitis. 2017;11(1):3–25.
23. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Therapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut 1989;30(7):983–9.
24. Stange EF, Travis SPL, Vermeire S, et al. European evidence based consensus on the
diagnosis and management of Crohn’s disease: definitions and diagnosis. Gut 2006;55
Suppl 1:i1-15.
25. Dignass A, Van Assche G, Lindsay JO, et al. The second European evidence-based
Consensus on the diagnosis and management of Crohn’s disease: Current management.
J Crohns Colitis 2010;4(1):28–62.
26. Plumb AA, Menys A, Russo E, et al. Magnetic resonance imaging-quantified small bowel
motility is a sensitive marker of response to medical therapy in Crohn’s disease. Aliment
Pharmacol Ther 2015;42(3):343–55.
27. Robin X, Turck N, Hainard A, et al. pROC: an open-source package for R and S+ to analyze
and compare ROC curves. BMC Bioinformatics 2011;12(1):77.
28. Bickelhaupt S, Pazahr S, Chuck N, et al. Crohn’s disease: small bowel motility impairment
correlates with inflammatory-related markers C-reactive protein and calprotectin.
Neurogastroenterol Motil 2013;25(6):467–73.
29. Menys A, Makanyanga J, Plumb A, et al. Aberrant Motility in Unaffected Small Bowel is
Linked to Inflammatory Burden and Patient Symptoms in Crohn’s Disease. Inflamm Bowel
Dis 2016;22(2):424–32.
30. de Jonge CS, Gollifer RM, Nederveen AJ, et al. Dynamic MRI for bowel motility imaging–
how fast and how long? Br J Radiol 2018;20170845.
Proefschrift2018-new2.indb 143 13/9/18 10:06
Terminal ileal motility during MR enterography
143
of MR motility measurement with histopathology in the terminal ileum. Neurogastroenterol
Motil 2013;25(9):749-e577.
17. Bickelhaupt S, Froehlich JM, Cattin R, et al. Differentiation between active and chronic
Crohn’s disease using MRI small-bowel motility examinations — Initial experience. Clin
Radiol 2013;68(12):1247–53.
18. Hahnemann ML, Nensa F, Kinner S, et al. Quantitative assessment of small bowel motility
in patients with Crohn’s disease using dynamic MRI. Neurogastroenterol Motil [Internet]
2015;27(6):841–8.
19. Li Z, Tielbeek JAW, Caan MWA, et al. Expiration-Phase Template-Based Motion Correction
of Free-Breathing Abdominal Dynamic Contrast Enhanced MRI. IEEE Trans Biomed Eng
2015;62(4):1215–25.
20. Naziroglu RE, Puylaert CAJ, Tielbeek JAW, et al. Semi-automatic bowel wall thickness
measurements on MR enterography in patients with Crohn’s disease. Br J Radiol [Internet]
2017;90(1074):20160654.
21. Puylaert CAJ, Schüffler PJ, Naziroglu RE, et al. Semiautomatic Assessment of the Terminal
Ileum and Colon in Patients with Crohn Disease Using MRI (the VIGOR++ Project). Acad
Radiol 2018.
22. Gomollón F, Dignass A, Annese V, et al. 3rd European Evidence-based Consensus on
the Diagnosis and Management of Crohn’s Disease 2016: Part 1: Diagnosis and Medical
Management. J Crohn’s Colitis. 2017;11(1):3–25.
23. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity
for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Therapeutiques
des Affections Inflammatoires du Tube Digestif (GETAID). Gut 1989;30(7):983–9.
24. Stange EF, Travis SPL, Vermeire S, et al. European evidence based consensus on the
diagnosis and management of Crohn’s disease: definitions and diagnosis. Gut 2006;55
Suppl 1:i1-15.
25. Dignass A, Van Assche G, Lindsay JO, et al. The second European evidence-based
Consensus on the diagnosis and management of Crohn’s disease: Current management.
J Crohns Colitis 2010;4(1):28–62.
26. Plumb AA, Menys A, Russo E, et al. Magnetic resonance imaging-quantified small bowel
motility is a sensitive marker of response to medical therapy in Crohn’s disease. Aliment
Pharmacol Ther 2015;42(3):343–55.
27. Robin X, Turck N, Hainard A, et al. pROC: an open-source package for R and S+ to analyze
and compare ROC curves. BMC Bioinformatics 2011;12(1):77.
28. Bickelhaupt S, Pazahr S, Chuck N, et al. Crohn’s disease: small bowel motility impairment
correlates with inflammatory-related markers C-reactive protein and calprotectin.
Neurogastroenterol Motil 2013;25(6):467–73.
29. Menys A, Makanyanga J, Plumb A, et al. Aberrant Motility in Unaffected Small Bowel is
Linked to Inflammatory Burden and Patient Symptoms in Crohn’s Disease. Inflamm Bowel
Dis 2016;22(2):424–32.
30. de Jonge CS, Gollifer RM, Nederveen AJ, et al. Dynamic MRI for bowel motility imaging–
how fast and how long? Br J Radiol 2018;20170845.
Proefschrift2018-new2.indb 143 13/9/18 10:06
Chapter 6
144
SUPPLEMENTARY MATERIALS
Appendix 1. The scan sequences used at the two imaging sites
3T Philips Achieva (Center 1) 3T Philips Ingenia (Center 2)
Sequence name BTFE BTFE
Plane Coronal Coronal
Field of view (mm) Variable Variable
No. of slices 20 20
Stacks Variable Variable
Repetition time 3.85 3.85
Echo time 1.91 1.91
Image matrix 256 x 184
Slice thickness (mm) 10 10
Slice gap (mm) 10 10
Averages 1 1
Flip angle 61 61
Temporal resolution (images per second
1.1 1.1
Dynamic scan time (sec-onds)
22 22
Appendix 2. Histopathology grading for endoscopic acute inflammation score (eAIS)
Histological variable Grade
Erosion or ulceration 0 = No, 1 = Yes
Polymorphs in the lamina propria 0 = No, 1 = Yes
Cryptitis 0 = No, 1 = Yes
Crypt abscess formation 0 = No, 1 = Yes
Inflammatory exudates 0 = No, 1 = Yes
Granulomas 0 = No, 1 = Yes
Proefschrift2018-new2.indb 144 13/9/18 10:06
Chapter 6
144
SUPPLEMENTARY MATERIALS
Appendix 1. The scan sequences used at the two imaging sites
3T Philips Achieva (Center 1) 3T Philips Ingenia (Center 2)
Sequence name BTFE BTFE
Plane Coronal Coronal
Field of view (mm) Variable Variable
No. of slices 20 20
Stacks Variable Variable
Repetition time 3.85 3.85
Echo time 1.91 1.91
Image matrix 256 x 184
Slice thickness (mm) 10 10
Slice gap (mm) 10 10
Averages 1 1
Flip angle 61 61
Temporal resolution (images per second
1.1 1.1
Dynamic scan time (sec-onds)
22 22
Appendix 2. Histopathology grading for endoscopic acute inflammation score (eAIS)
Histological variable Grade
Erosion or ulceration 0 = No, 1 = Yes
Polymorphs in the lamina propria 0 = No, 1 = Yes
Cryptitis 0 = No, 1 = Yes
Crypt abscess formation 0 = No, 1 = Yes
Inflammatory exudates 0 = No, 1 = Yes
Granulomas 0 = No, 1 = Yes
Proefschrift2018-new2.indb 144 13/9/18 10:06
Terminal ileal motility during MR enterography
145
Appendix 3. A receiver operating characteristic (ROC) curve was generated based on the
data from the 28 patients15 who had both motility and eAIS scores. ROC analysis revealed an
AUC of 0.85 (95% CI 0.70 – 0.99, DeLong method). An optimal motility cut-off value of 0.30
was chosen empirically that demonstrated an acceptable balance between sensitivity (0.77)
and specificity (0.73)
Appendix 4. Derivation and evaluation of ROC optimal operating points
Operating point (“Top-Left” method)
Sensitivity Specifi city
Motility against CDEIS
0.23 0.92 0.76
MaRIA against CDEIS
9.9 0.83 0.81
Motility against eAIS 0.22 0.87 0.85
MaRIA against eAIS 9.9 0.75 0.82
Proefschrift2018-new2.indb 145 13/9/18 10:06
Terminal ileal motility during MR enterography
145
Appendix 3. A receiver operating characteristic (ROC) curve was generated based on the
data from the 28 patients15 who had both motility and eAIS scores. ROC analysis revealed an
AUC of 0.85 (95% CI 0.70 – 0.99, DeLong method). An optimal motility cut-off value of 0.30
was chosen empirically that demonstrated an acceptable balance between sensitivity (0.77)
and specificity (0.73)
Appendix 4. Derivation and evaluation of ROC optimal operating points
Operating point (“Top-Left” method)
Sensitivity Specifi city
Motility against CDEIS
0.23 0.92 0.76
MaRIA against CDEIS
9.9 0.83 0.81
Motility against eAIS 0.22 0.87 0.85
MaRIA against eAIS 9.9 0.75 0.82
Proefschrift2018-new2.indb 145 13/9/18 10:06
Chapter 6
146
Appendix 5. Per-hospital analysis
CDEIS Standard at AMC
Sensitivityn=49(TP+FN)
95% Con-fi dence interval
Signifi cance (P) against motility
Specifi city n=49(TN+FP)
95% Con-fi dence Interval
Signifi cance (P) against motility
Motility <0.30
90.0 (27 + 3)
73.5 to 98 NA 68.4 (13 + 6) 43.5 to 87.4 NA
MaRIA >7 83.3 (25 + 5) 65.3 to 94.4 0.62 89.5 (17 + 2) 66.9 to 98.7 0.22
MaRIA >11 76.7 (23 + 7) 57.7 to 90.1 0.45 89.5 (17 + 2) 66.9 to 98.7 0.22
eAIS Standard at AMC
Sensitivityn=49(TP+FN)
95% Con-fi dence interval
Signifi cance (P) against motility
Specifi city n=49(TN+FP)
95% Con-fi dence Interval
Signifi cance (P) against motility
Motility <0.30
90.6 (29 + 3)
75.9 to 98.0 NA 76.5 (13 + 4) 50.1 to 93.2 NA
MaRIA >7 78.1 (25 + 7) 60.0 to 70.1 0.22 88.2 (15 + 2) 63.6 to 98.5 0.62
MaRIA >11 75.0 (24 + 8)
56.6 to 88.5 0.125 94.1 (16 + 1) 71.3 to 99.9 0.38
CDEIS Standard at UCL
Sensitivityn=33(TP+FN)
95% Con-fi dence interval
Signifi cance (P) against motility
Specifi city n=33(TN+FP)
95% Con-fi dence Interval
Signifi cance (P) against motility
Motility <0.30
100 (11 + 0) 71.5 to 100 NA 50 (11 + 11) 28.2 to 71.8 NA
MaRIA >7 90.1 (10 + 1) 58.7 to 99.8 1 63.6 (14 + 8) 40.7 to 82.8 0.45
MaRIA >11 81.8 (9 + 2) 48.2 to 97.7 1 72.7 (16 + 6) 72.7 (49.9 to 89.3)
0.125
eAIS Standard at UCL
Sensitivityn=33(TP+FN)
95% Con-fi dence interval
Signifi cance (P) against motility
Specifi city n=33(TN+FP)
95% Con-fi dence Interval
Signifi cance (P) against motility
Motility <0.30
93.8 (15 +1) 69.8 to 99.8 NA 58.8 (10 + 7) 32.9 to 81.6 NA
MaRIA >7 68.8 (11 + 5) 41.3 to 88.9 0.13 58.8 (10 + 7) 32.9 to 81.6 1
MaRIA >11 62.5 (10 + 6) 35.4 to 84.8 0.06 70.6 (10 + 6) 44.0 to 89.7 0.62
Proefschrift2018-new2.indb 146 13/9/18 10:06
Chapter 6
146
Appendix 5. Per-hospital analysis
CDEIS Standard at AMC
Sensitivityn=49(TP+FN)
95% Con-fi dence interval
Signifi cance (P) against motility
Specifi city n=49(TN+FP)
95% Con-fi dence Interval
Signifi cance (P) against motility
Motility <0.30
90.0 (27 + 3)
73.5 to 98 NA 68.4 (13 + 6) 43.5 to 87.4 NA
MaRIA >7 83.3 (25 + 5) 65.3 to 94.4 0.62 89.5 (17 + 2) 66.9 to 98.7 0.22
MaRIA >11 76.7 (23 + 7) 57.7 to 90.1 0.45 89.5 (17 + 2) 66.9 to 98.7 0.22
eAIS Standard at AMC
Sensitivityn=49(TP+FN)
95% Con-fi dence interval
Signifi cance (P) against motility
Specifi city n=49(TN+FP)
95% Con-fi dence Interval
Signifi cance (P) against motility
Motility <0.30
90.6 (29 + 3)
75.9 to 98.0 NA 76.5 (13 + 4) 50.1 to 93.2 NA
MaRIA >7 78.1 (25 + 7) 60.0 to 70.1 0.22 88.2 (15 + 2) 63.6 to 98.5 0.62
MaRIA >11 75.0 (24 + 8)
56.6 to 88.5 0.125 94.1 (16 + 1) 71.3 to 99.9 0.38
CDEIS Standard at UCL
Sensitivityn=33(TP+FN)
95% Con-fi dence interval
Signifi cance (P) against motility
Specifi city n=33(TN+FP)
95% Con-fi dence Interval
Signifi cance (P) against motility
Motility <0.30
100 (11 + 0) 71.5 to 100 NA 50 (11 + 11) 28.2 to 71.8 NA
MaRIA >7 90.1 (10 + 1) 58.7 to 99.8 1 63.6 (14 + 8) 40.7 to 82.8 0.45
MaRIA >11 81.8 (9 + 2) 48.2 to 97.7 1 72.7 (16 + 6) 72.7 (49.9 to 89.3)
0.125
eAIS Standard at UCL
Sensitivityn=33(TP+FN)
95% Con-fi dence interval
Signifi cance (P) against motility
Specifi city n=33(TN+FP)
95% Con-fi dence Interval
Signifi cance (P) against motility
Motility <0.30
93.8 (15 +1) 69.8 to 99.8 NA 58.8 (10 + 7) 32.9 to 81.6 NA
MaRIA >7 68.8 (11 + 5) 41.3 to 88.9 0.13 58.8 (10 + 7) 32.9 to 81.6 1
MaRIA >11 62.5 (10 + 6) 35.4 to 84.8 0.06 70.6 (10 + 6) 44.0 to 89.7 0.62
Proefschrift2018-new2.indb 146 13/9/18 10:06
Terminal ileal motility during MR enterography
147
Appendix 6. Scatter plot of CDEIS & eAIS.
Proefschrift2018-new2.indb 147 13/9/18 10:06
Terminal ileal motility during MR enterography
147
Appendix 6. Scatter plot of CDEIS & eAIS.
Proefschrift2018-new2.indb 147 13/9/18 10:06
Proefschrift2018-new2.indb 148 13/9/18 10:06 Proefschrift2018-new2.indb 148 13/9/18 10:06
CHAPTER 7
Long-term performance of readers trained in grading Crohn’s disease activity using MRI
Carl. A.J. Puylaert, Jeroen A.W. Tielbeek, Shandra Bipat, Thierry. N. Boellaard, C.
Yung Nio, Jaap Stoker
Proefschrift2018-new2.indb 149 13/9/18 10:06
CHAPTER 7
Long-term performance of readers trained in grading Crohn’s disease activity using MRI
Carl. A.J. Puylaert, Jeroen A.W. Tielbeek, Shandra Bipat, Thierry. N. Boellaard, C.
Yung Nio, Jaap Stoker
Proefschrift2018-new2.indb 149 13/9/18 10:06
Chapter 7
150
ABSTRACT
Purpose
We aim to evaluate the long-term performance of readers who had participated
in previous magnetic resonance imaging (MRI) reader training in grading Crohn
disease activity.
Materials and Methods
Fourteen readers (8 women; 12 radiologists, 2 residents; mean age 40; range 31–59),
who had participated in a previous MRI reader training, participated in a follow-up
evaluation after a mean interval of 29 months (range 25–34 months). Follow-up
evaluation comprised 25 MRI cases of suspected or known Crohn disease patients
with direct feedback; cases were identical to the evaluation set used in the initial
reader training (of which readers were unaware). Grading accuracy, overstaging, and
understaging were compared between training and follow-up using a consensus
score by two experienced abdominal radiologists as the reference standard.
Results
In the follow-up evaluation, overall grading accuracy was 73% (95% confidence
interval [CI]: 62%–81%), which was comparable to reader training grading accuracy
(72%, 95% CI: 61%–80%) (P = .66). Overstaging decreased significantly from 19%
(95% CI: 12%–27%) to 13% (95% CI: 8%–21%) between training and follow-up (P = .03),
whereas understaging increased significantly from 9% (95% CI: 4%–21%) to 14%
(95% CI: 7%–26%) (P < .01).
Conclusions
Readers have consistent long-term accuracy for grading Crohn disease activity
after case-based reader training with direct feedback.
Proefschrift2018-new2.indb 150 13/9/18 10:06
Chapter 7
150
ABSTRACT
Purpose
We aim to evaluate the long-term performance of readers who had participated
in previous magnetic resonance imaging (MRI) reader training in grading Crohn
disease activity.
Materials and Methods
Fourteen readers (8 women; 12 radiologists, 2 residents; mean age 40; range 31–59),
who had participated in a previous MRI reader training, participated in a follow-up
evaluation after a mean interval of 29 months (range 25–34 months). Follow-up
evaluation comprised 25 MRI cases of suspected or known Crohn disease patients
with direct feedback; cases were identical to the evaluation set used in the initial
reader training (of which readers were unaware). Grading accuracy, overstaging, and
understaging were compared between training and follow-up using a consensus
score by two experienced abdominal radiologists as the reference standard.
Results
In the follow-up evaluation, overall grading accuracy was 73% (95% confidence
interval [CI]: 62%–81%), which was comparable to reader training grading accuracy
(72%, 95% CI: 61%–80%) (P = .66). Overstaging decreased significantly from 19%
(95% CI: 12%–27%) to 13% (95% CI: 8%–21%) between training and follow-up (P = .03),
whereas understaging increased significantly from 9% (95% CI: 4%–21%) to 14%
(95% CI: 7%–26%) (P < .01).
Conclusions
Readers have consistent long-term accuracy for grading Crohn disease activity
after case-based reader training with direct feedback.
Proefschrift2018-new2.indb 150 13/9/18 10:06
Long-term performance of trained readers
151
INTRODUCTION
Magnetic resonance imaging (MRI) is the preferred technique for assessment of
disease activity in Crohn disease because of its capability to evaluate intra- and
extramural disease of both small bowel and colon, without the use of ionizing
radiation. A previous study showed that inexperienced radiologists and residents
could be successful- ly trained to grade Crohn disease activity using case-based
training (1). In that study, readers graded 100 MRI cases of patients with suspected
or known Crohn disease, with the readers receiving direct feedback after each case.
Significant improvement of grading accuracy was seen (66% to 75%, P = .003),
whereas understaging decreased significantly (15% to 7%, P < .001).
Although case-based reader training can provide short- term improvement, to
our knowledge, no studies have investigated long-term performance, as reader
evaluation was always performed at the time of training. Furthermore, there is no
knowledge of how daily practice impacts reader’s level of experience after training.
The primary objective of this study was to compare readers’ results in grading
Crohn disease activity between reader training and a follow-up evaluation of the
same set of MRI cases. The secondary objective was to determine the influence of
readers’ interim experience on grading results at follow-up evaluation.
MATERIALS AND METHODS
Thirty-one readers from different hospitals, who had previously participated in a
reader training at a single tertiary center (1), were invited for a follow-up evaluation.
Before the start of initial reader training, readers had no (n = 14) or limited (n = 17)
MRI experience (<25 magnetic resonance enterography/enteroclysis examinations)
for evaluating Crohn disease. During the initial training readers had graded 100 MRI
cases of patients with suspected or known Crohn disease, with the readers receiving
direct feedback after each case (ie, expert radiologic report and endoscopic or
surgical findings). Before reading those 100 cases, the readers had graded 25
MRI cases without direct feedback, serving as a baseline. These 25 baseline cases
were reassessed as the last 25 cases of the aforementioned 100 cases but in a
differently randomized order; this time the readers received feedback case by case.
Proefschrift2018-new2.indb 151 13/9/18 10:06
Long-term performance of trained readers
151
INTRODUCTION
Magnetic resonance imaging (MRI) is the preferred technique for assessment of
disease activity in Crohn disease because of its capability to evaluate intra- and
extramural disease of both small bowel and colon, without the use of ionizing
radiation. A previous study showed that inexperienced radiologists and residents
could be successful- ly trained to grade Crohn disease activity using case-based
training (1). In that study, readers graded 100 MRI cases of patients with suspected
or known Crohn disease, with the readers receiving direct feedback after each case.
Significant improvement of grading accuracy was seen (66% to 75%, P = .003),
whereas understaging decreased significantly (15% to 7%, P < .001).
Although case-based reader training can provide short- term improvement, to
our knowledge, no studies have investigated long-term performance, as reader
evaluation was always performed at the time of training. Furthermore, there is no
knowledge of how daily practice impacts reader’s level of experience after training.
The primary objective of this study was to compare readers’ results in grading
Crohn disease activity between reader training and a follow-up evaluation of the
same set of MRI cases. The secondary objective was to determine the influence of
readers’ interim experience on grading results at follow-up evaluation.
MATERIALS AND METHODS
Thirty-one readers from different hospitals, who had previously participated in a
reader training at a single tertiary center (1), were invited for a follow-up evaluation.
Before the start of initial reader training, readers had no (n = 14) or limited (n = 17)
MRI experience (<25 magnetic resonance enterography/enteroclysis examinations)
for evaluating Crohn disease. During the initial training readers had graded 100 MRI
cases of patients with suspected or known Crohn disease, with the readers receiving
direct feedback after each case (ie, expert radiologic report and endoscopic or
surgical findings). Before reading those 100 cases, the readers had graded 25
MRI cases without direct feedback, serving as a baseline. These 25 baseline cases
were reassessed as the last 25 cases of the aforementioned 100 cases but in a
differently randomized order; this time the readers received feedback case by case.
Proefschrift2018-new2.indb 151 13/9/18 10:06
Chapter 7
152
This setup facilitated comparing grading accuracy, understaging, and overstaging
before and after the case-by-case reader training. Fourteen of these 31 readers from
13 hospitals attended the follow-up evaluation and were included in the present
study. The mean interval between reader training and the follow-up evaluation
was 29 months (range: 25–34). Readers were asked to describe their interim
experience with magnetic resonance enterography/enteroclysis and abdominal MRI
examinations between reader training and follow-up evaluation (described as the
average number of examinations seen per month).
Figure 1. Flowchart of reader training and follow-up evaluation.
Proefschrift2018-new2.indb 152 13/9/18 10:06
Chapter 7
152
This setup facilitated comparing grading accuracy, understaging, and overstaging
before and after the case-by-case reader training. Fourteen of these 31 readers from
13 hospitals attended the follow-up evaluation and were included in the present
study. The mean interval between reader training and the follow-up evaluation
was 29 months (range: 25–34). Readers were asked to describe their interim
experience with magnetic resonance enterography/enteroclysis and abdominal MRI
examinations between reader training and follow-up evaluation (described as the
average number of examinations seen per month).
Figure 1. Flowchart of reader training and follow-up evaluation.
Proefschrift2018-new2.indb 152 13/9/18 10:06
Long-term performance of trained readers
153
Table 1. Patient and MRI Characteristics
Parameter
Gender Male or female 13 / 12
Age at time of imaging Years (mean/SD) 37 (13)
Previous surgery n (%) 11 (44%)
MRI technique 1.5 T 20
3.0 T 5
Preparation Enterography 16
Enteroclysis 9
Disease activity of the patients per examination (reference standard)
None 7
Mild 6
Moderate 5
Severe 7
MRI, magnetic resonance imaging; SD, standard deviation.
Follow-Up Evaluation
During follow-up evaluation, readers evaluated 25 MRI cases (16 enterography and
9 enteroclysis), which were identical to the baseline set and the reassessed last
25 cases from the reader training, with direct feedback after each case (Fig 1).
The readers were not informed that these were identical cases from the previous
training session. Patient and MRI characteristics for these cases are presented in
Table 1. All examinations were performed at a 1.5 T unit (Avanto, Siemens Healthcare,
Erlangen, Germany) or a 3.0 T unit (Philips Medical Systems, Best, The Netherlands)
in the supine position, and the scan protocol included at least a coronal and axial
T2-weighted sequence, a coronal fat-saturated T1-weighted unenhanced sequence,
and coronal and axial fat-saturated postcontrast T1-weighted sequences (1). Follow-
up evaluation started with a 1-hour refresher course on how to use the grading
system (including two instruction cases) and technical instructions for our picture
archiving and communication system. Subsequently, readers individually evaluated
all 25 MRI cases using the same online scoring form used in reader training to grade
disease activity (http://mrentero.webklik.nl). This scoring system is a modification of
a validated qualitative scoring system, with the addition of pattern of enhancement,
comb sign, disease length, and assessment of extra-enteric complications (2) (Table
2). As an extension to this scoring system, an arbitrary classification was used to
classify patients in four categories: none, mild, moderate, or severe disease (Table
3). This modified scoring system was used, as clinical important imaging findings
Proefschrift2018-new2.indb 153 13/9/18 10:06
Long-term performance of trained readers
153
Table 1. Patient and MRI Characteristics
Parameter
Gender Male or female 13 / 12
Age at time of imaging Years (mean/SD) 37 (13)
Previous surgery n (%) 11 (44%)
MRI technique 1.5 T 20
3.0 T 5
Preparation Enterography 16
Enteroclysis 9
Disease activity of the patients per examination (reference standard)
None 7
Mild 6
Moderate 5
Severe 7
MRI, magnetic resonance imaging; SD, standard deviation.
Follow-Up Evaluation
During follow-up evaluation, readers evaluated 25 MRI cases (16 enterography and
9 enteroclysis), which were identical to the baseline set and the reassessed last
25 cases from the reader training, with direct feedback after each case (Fig 1).
The readers were not informed that these were identical cases from the previous
training session. Patient and MRI characteristics for these cases are presented in
Table 1. All examinations were performed at a 1.5 T unit (Avanto, Siemens Healthcare,
Erlangen, Germany) or a 3.0 T unit (Philips Medical Systems, Best, The Netherlands)
in the supine position, and the scan protocol included at least a coronal and axial
T2-weighted sequence, a coronal fat-saturated T1-weighted unenhanced sequence,
and coronal and axial fat-saturated postcontrast T1-weighted sequences (1). Follow-
up evaluation started with a 1-hour refresher course on how to use the grading
system (including two instruction cases) and technical instructions for our picture
archiving and communication system. Subsequently, readers individually evaluated
all 25 MRI cases using the same online scoring form used in reader training to grade
disease activity (http://mrentero.webklik.nl). This scoring system is a modification of
a validated qualitative scoring system, with the addition of pattern of enhancement,
comb sign, disease length, and assessment of extra-enteric complications (2) (Table
2). As an extension to this scoring system, an arbitrary classification was used to
classify patients in four categories: none, mild, moderate, or severe disease (Table
3). This modified scoring system was used, as clinical important imaging findings
Proefschrift2018-new2.indb 153 13/9/18 10:06
Chapter 7
154
(eg, fistula, abscess) are not integrated in the published scoring systems. The
following MRI features were recorded for each patient for the most severe lesion:
mural thickness, mural T2 signal, T1 enhancement, mural enhancement pattern,
total length of disease, and presence of comb sign (vascular enlargement of the
vasa recta). The following complications were described to be present: infiltrate
(tethering and kinking of bowel loops), abscess, fistula, and severe stenosis (defined
as 80% lumen reduction with prestenotic dilatation and a moderate-to-severe
increase in mural T2 signal) (Fig 2). Additionally, readers were asked to grade their
confidence for severity grading (0–10). The scoring system and classification system
were identical to the initial reader training. Surgical history was provided for each
case if applicable. Cases were reviewed in the same order as during the initial case-
by-case reader training, and direct feedback was given after each case.
Figure 2. MR enterography (3.0 T) of a 41-year-old man with known Crohn disease. Coronal T1-weighted image after intravenous gadolini-um contrast medium (a) and axial T2-weighted HASTE images (b) show a lesion (arrow) in the terminal ileum with an increased bowel wall thickness up to 10 mm, a marked increase in postcontrast enhancement with mixed homogeneous or layered appearance (layered pattern visible in contigu-ous images (not shown)). A prestenotic dilatation (arrowheads) (a,c) and marked increase in mural T2 signal are observed on cor-onal and axial fat-saturated T2-weighted images (arrow) (c,d), confirming the pres-ence of severe stenosis with active inflammation. Ten out of 14 readers correctly graded this case as severe disease, either due to the presence of severe stenosis (n = 7) or a total score ≥14 (n = 3). HASTE, half-Fourier acquisition single-shot turbo spin-echo; MR, magnetic resonance.
Proefschrift2018-new2.indb 154 13/9/18 10:06
Chapter 7
154
(eg, fistula, abscess) are not integrated in the published scoring systems. The
following MRI features were recorded for each patient for the most severe lesion:
mural thickness, mural T2 signal, T1 enhancement, mural enhancement pattern,
total length of disease, and presence of comb sign (vascular enlargement of the
vasa recta). The following complications were described to be present: infiltrate
(tethering and kinking of bowel loops), abscess, fistula, and severe stenosis (defined
as 80% lumen reduction with prestenotic dilatation and a moderate-to-severe
increase in mural T2 signal) (Fig 2). Additionally, readers were asked to grade their
confidence for severity grading (0–10). The scoring system and classification system
were identical to the initial reader training. Surgical history was provided for each
case if applicable. Cases were reviewed in the same order as during the initial case-
by-case reader training, and direct feedback was given after each case.
Figure 2. MR enterography (3.0 T) of a 41-year-old man with known Crohn disease. Coronal T1-weighted image after intravenous gadolini-um contrast medium (a) and axial T2-weighted HASTE images (b) show a lesion (arrow) in the terminal ileum with an increased bowel wall thickness up to 10 mm, a marked increase in postcontrast enhancement with mixed homogeneous or layered appearance (layered pattern visible in contigu-ous images (not shown)). A prestenotic dilatation (arrowheads) (a,c) and marked increase in mural T2 signal are observed on cor-onal and axial fat-saturated T2-weighted images (arrow) (c,d), confirming the pres-ence of severe stenosis with active inflammation. Ten out of 14 readers correctly graded this case as severe disease, either due to the presence of severe stenosis (n = 7) or a total score ≥14 (n = 3). HASTE, half-Fourier acquisition single-shot turbo spin-echo; MR, magnetic resonance.
Proefschrift2018-new2.indb 154 13/9/18 10:06
Long-term performance of trained readers
155
Table 2. MRI activity grading, score of features and complications per patient (2).
Score of MRI features
MRI Features 0 1 2 3
Mural thickness 1–3 mm >3–5 mm >5–7 mm >7 mm
Mural T2 signal Normal bowel Minor increase Moderate increase
Severe increase
T1 enhancement Normal bowel Minor increase Moderate increase
Severe increase
Mural enhancement pattern
NA Homogeneous Mucosal Layered
Total length of disease 0 cm >0–5 cm >5–15 cm >15 cm
Comb sign No Yes
Presence of complications
Infiltrate* No Yes
Abscess No Yes
Fistula No Yes
Severe stenosis** No Yes
MRI, magnetic resonance imaging; NA, not applicable.Lesions in the small bowel and colon were assessed. In case of involvement of more than one lesion, the lesion with the highest activity score was leading in the final assessment.* Tethering and kinking of bowel loops was considered as an infiltrate.** Lumen reduction (>80%) with prestenotic dilatation, wall thickening (> 3 mm) and in-creased (moderate/severe) mural T2 signal was considered as a severe stenosis with active disease.
Table 3. Description of disease activity based on MRI features and complications in Table 2.
Crohn Disease Activity
None No signs of Crohn disease activity
Mild Signs of activity. No features with score 3. No complications. Total score ≤ 8.
Moderate Score 9–13 or contains a feature with score 3. No signs of complications
Severe Presence of at least one complication or a total score of ≥ 14.
MRI, magnetic resonance imaging.
Reference Standard
All MRI examinations had been scored by two abdominal radiologists (CYN and
JS), with 18 and 21 years of experience in imaging of inflammatory bowel disease,
followed by a consensus reading. The score after consensus was used as the
reference standard to form 4 x 4 tables for each trainee and calculate reader grading
accuracy, understaging, and overstaging.
Proefschrift2018-new2.indb 155 13/9/18 10:06
Long-term performance of trained readers
155
Table 2. MRI activity grading, score of features and complications per patient (2).
Score of MRI features
MRI Features 0 1 2 3
Mural thickness 1–3 mm >3–5 mm >5–7 mm >7 mm
Mural T2 signal Normal bowel Minor increase Moderate increase
Severe increase
T1 enhancement Normal bowel Minor increase Moderate increase
Severe increase
Mural enhancement pattern
NA Homogeneous Mucosal Layered
Total length of disease 0 cm >0–5 cm >5–15 cm >15 cm
Comb sign No Yes
Presence of complications
Infiltrate* No Yes
Abscess No Yes
Fistula No Yes
Severe stenosis** No Yes
MRI, magnetic resonance imaging; NA, not applicable.Lesions in the small bowel and colon were assessed. In case of involvement of more than one lesion, the lesion with the highest activity score was leading in the final assessment.* Tethering and kinking of bowel loops was considered as an infiltrate.** Lumen reduction (>80%) with prestenotic dilatation, wall thickening (> 3 mm) and in-creased (moderate/severe) mural T2 signal was considered as a severe stenosis with active disease.
Table 3. Description of disease activity based on MRI features and complications in Table 2.
Crohn Disease Activity
None No signs of Crohn disease activity
Mild Signs of activity. No features with score 3. No complications. Total score ≤ 8.
Moderate Score 9–13 or contains a feature with score 3. No signs of complications
Severe Presence of at least one complication or a total score of ≥ 14.
MRI, magnetic resonance imaging.
Reference Standard
All MRI examinations had been scored by two abdominal radiologists (CYN and
JS), with 18 and 21 years of experience in imaging of inflammatory bowel disease,
followed by a consensus reading. The score after consensus was used as the
reference standard to form 4 x 4 tables for each trainee and calculate reader grading
accuracy, understaging, and overstaging.
Proefschrift2018-new2.indb 155 13/9/18 10:06
Chapter 7
156
Statistical Analysis
Grading accuracy, understaging, and overstaging were compared between reader
training and follow-up evaluation. Readers evaluated all cases independently;
therefore, we regarded these measurements as independent repetitions. However,
to correct for repeated measurements among readers, we used generalized
estimating equations to estimate grading accuracy, understaging, and overstaging.
Grading accuracy, understaging, and overstaging were compared by adding the
datasets as covariates in the abovementioned model: (1) reassessed set from reader
training and (2) the follow-up evaluation set; a significant regression coefficient
was considered to show a difference between these sets. Grading accuracy,
understaging, and overstaging values with corresponding confidence intervals
(CIs) were calculated by antilogit transformation of the logit values and standard
errors of the models. Linear weighted kappa values were calculated for each reader
to evaluate intrarater reliability between reader training and follow-up evaluation.
For the subanalysis on interim experience, we ranked readers based on interim
experience with magnetic resonance enterography/enteroclysis (average number
of examinations per month) and subsequently divided them into two equally
large groups, categorized as low or high experience. The same division was made
based on interim experience with abdominal MRI. Mean confidence scores were
first calculated per case and subsequently compared between reader training and
followup evaluation using the paired t test. For all analyses, a P value of <.05 was
considered to indicate a statistically significant difference. SPSS statistics 22 (IBM
Corporation, Armonk, NY) and GraphPad Prism 5.01 (La Jolla, CA) were used. The
institutional review board waived requirement of informed consent for the use of
MRI examinations. All readers gave written informed consent.
RESULTS
Readers
Fourteen readers (8 women; mean age 40; range 31–59) participated in one of
three follow-up evaluation sessions. Ten other readers had enlisted for the follow-
up evaluation but were unavailable at the time of the sessions. From the 31 readers
of the initial reader training, the 14 participants of the follow-up evaluation and
the 17 nonparticipants showed equal grading accuracy at the initial reader training
Proefschrift2018-new2.indb 156 13/9/18 10:06
Chapter 7
156
Statistical Analysis
Grading accuracy, understaging, and overstaging were compared between reader
training and follow-up evaluation. Readers evaluated all cases independently;
therefore, we regarded these measurements as independent repetitions. However,
to correct for repeated measurements among readers, we used generalized
estimating equations to estimate grading accuracy, understaging, and overstaging.
Grading accuracy, understaging, and overstaging were compared by adding the
datasets as covariates in the abovementioned model: (1) reassessed set from reader
training and (2) the follow-up evaluation set; a significant regression coefficient
was considered to show a difference between these sets. Grading accuracy,
understaging, and overstaging values with corresponding confidence intervals
(CIs) were calculated by antilogit transformation of the logit values and standard
errors of the models. Linear weighted kappa values were calculated for each reader
to evaluate intrarater reliability between reader training and follow-up evaluation.
For the subanalysis on interim experience, we ranked readers based on interim
experience with magnetic resonance enterography/enteroclysis (average number
of examinations per month) and subsequently divided them into two equally
large groups, categorized as low or high experience. The same division was made
based on interim experience with abdominal MRI. Mean confidence scores were
first calculated per case and subsequently compared between reader training and
followup evaluation using the paired t test. For all analyses, a P value of <.05 was
considered to indicate a statistically significant difference. SPSS statistics 22 (IBM
Corporation, Armonk, NY) and GraphPad Prism 5.01 (La Jolla, CA) were used. The
institutional review board waived requirement of informed consent for the use of
MRI examinations. All readers gave written informed consent.
RESULTS
Readers
Fourteen readers (8 women; mean age 40; range 31–59) participated in one of
three follow-up evaluation sessions. Ten other readers had enlisted for the follow-
up evaluation but were unavailable at the time of the sessions. From the 31 readers
of the initial reader training, the 14 participants of the follow-up evaluation and
the 17 nonparticipants showed equal grading accuracy at the initial reader training
Proefschrift2018-new2.indb 156 13/9/18 10:06
Long-term performance of trained readers
157
of 72% (95% CI: 62%–81%) and 72% (95% CI: 63%–80%), respectively (P = .89).
The participating readers included twelve radiologists and two residents from four
academic hospitals, four teaching hospitals, and five general hospitals. Between
reader training and follow-up evaluation, readers had examined an average of
6.2 (range 1–20) magnetic resonance enterography/enteroclysis and 30.9 (range
5–100) abdominal MRI examinations per month.
Grading Accuracy, Understaging, and Overstaging
Grading accuracy, understaging, and overstaging for reader training and follow-
up evaluation with corresponding P values and grading accuracies per category
of severity are presented in Table 4. Grading accuracies at reader training and the
follow-up evaluation were comparable (P = .66), whereas overstaging decreased
significantly (P = .03) and understaging increased significantly (P < .01). Accuracies
for grading of individual features during follow-up evaluation are presented in Table
5.
Intra-Rater Reliability
Weighted kappa values for agreement between reader training and follow-up
evaluation for each reader ranged from 0.42 to 0.91, with a mean of 0.68.
Table 4. Grading accuracy, understaging and overstaging for reader training and follow-up
evaluation with corresponding P values and grading accuracies per category of severity.
Reader Training Follow-Up P Value
Grading accuracy (95% CI)
72% (95% CI: 61%–80%) 73% (95% CI: 62%–81%) 0.66
Understaging (95% CI) 9% (95% CI: 4%–21%) 14% (95% CI: 7%–26%) <0.01
Overstaging (95% CI) 19% (95% CI: 12%–27%) 13% (95% CI: 8%–21%) 0.03
Grading accuracy per category:
None (95% CI) 75% (95% CI: 59%–86%) 82% (95% CI: 70%–90%) 0.03
Mild (95% CI) 61% (95% CI: 38%–80%) 62% (95% CI: 41%–80%) 0.87
Moderate (95% CI) 60% (95% CI: 50%–69%) 61% (95% CI: 46%–75%) 0.82
Severe (95% CI) 88% (95% CI: 61%–97%) 84% (95% CI: 58%–95%) 0.15
CI, confidence interval. Bold values indicate statistical significant differences (P < .05).
Proefschrift2018-new2.indb 157 13/9/18 10:06
Long-term performance of trained readers
157
of 72% (95% CI: 62%–81%) and 72% (95% CI: 63%–80%), respectively (P = .89).
The participating readers included twelve radiologists and two residents from four
academic hospitals, four teaching hospitals, and five general hospitals. Between
reader training and follow-up evaluation, readers had examined an average of
6.2 (range 1–20) magnetic resonance enterography/enteroclysis and 30.9 (range
5–100) abdominal MRI examinations per month.
Grading Accuracy, Understaging, and Overstaging
Grading accuracy, understaging, and overstaging for reader training and follow-
up evaluation with corresponding P values and grading accuracies per category
of severity are presented in Table 4. Grading accuracies at reader training and the
follow-up evaluation were comparable (P = .66), whereas overstaging decreased
significantly (P = .03) and understaging increased significantly (P < .01). Accuracies
for grading of individual features during follow-up evaluation are presented in Table
5.
Intra-Rater Reliability
Weighted kappa values for agreement between reader training and follow-up
evaluation for each reader ranged from 0.42 to 0.91, with a mean of 0.68.
Table 4. Grading accuracy, understaging and overstaging for reader training and follow-up
evaluation with corresponding P values and grading accuracies per category of severity.
Reader Training Follow-Up P Value
Grading accuracy (95% CI)
72% (95% CI: 61%–80%) 73% (95% CI: 62%–81%) 0.66
Understaging (95% CI) 9% (95% CI: 4%–21%) 14% (95% CI: 7%–26%) <0.01
Overstaging (95% CI) 19% (95% CI: 12%–27%) 13% (95% CI: 8%–21%) 0.03
Grading accuracy per category:
None (95% CI) 75% (95% CI: 59%–86%) 82% (95% CI: 70%–90%) 0.03
Mild (95% CI) 61% (95% CI: 38%–80%) 62% (95% CI: 41%–80%) 0.87
Moderate (95% CI) 60% (95% CI: 50%–69%) 61% (95% CI: 46%–75%) 0.82
Severe (95% CI) 88% (95% CI: 61%–97%) 84% (95% CI: 58%–95%) 0.15
CI, confidence interval. Bold values indicate statistical significant differences (P < .05).
Proefschrift2018-new2.indb 157 13/9/18 10:06
Chapter 7
158
Table 5. Individual MRI features
MRI feature (numbers in parentheses reported by reference standard)
Grading accuracy (95% CI)
Mural thickness 69% (57%–78%)
Mural T2 signal 59% (50%–68%)
T1 enhancement 49% (39%–59%)
Enhancement pattern 65% (54%–75%)
Total length of disease 66% (55%–75%)
Comb sign (n=8) 74% (61%–84%)
Infiltrate (n= 4) 93% (88%–96%)
Abscess (n=1) 97% (92%–99%)
Fistula (n=5) 90% (81%–95%)
Severe stenosis (n=6) 83% (74%–90%)
CI, confidence interval; MRI, magnetic resonance imaging
Table 6. Comparison of grading accuracies between reader training and follow-up evaluation
with readers divided based on experience with MRE and abdominal MRI.
MREReader training, grading accuracy (95% CI)
Follow-up, grading accuracy* (95% CI)
P Value
Low experience 69% (95% CI: 56–79%) 75% (95% CI: 64–83%) 0.1
High experience 75% (95% CI: 63–84%) 71% (95% CI: 59–80%) 0.23
Abdominal MRI
Low experience 70% (95% CI: 58–79%) 69% (95% CI: 57–79%) 0.88
High experience 74% (95% CI: 62–83%) 77% (95% CI: 66–85%) 0.36
CI, confidence interval; MRE, magnetic resonance enterography or enteroclysis; MRI, mag-netic resonance imaging.* Follow-up grading accuracy, understaging, and overstaging did not differ between groups with low and high interim experience in MRE (P = .39, P = .68, and P = .61, respec-tively) and between groups with low and high interim experience with abdominal MRI (P = .06, P = .61, and P = .06).
Interim Experience
When readers were ranked based on interim experience with magnetic resonance
enterography/enteroclysis, the lower half (low interim experience) had evaluated
a mean of 2.7 (range 1–4) examinations per month, whereas the upper half (high
interim experience) had evaluated a mean of 9.7 (range 5–20) examinations per
month. Readers categorized as having low interim experience with abdominal
MRI had evaluated a mean of 9.7 (range 5–16) examinations per month, whereas
Proefschrift2018-new2.indb 158 13/9/18 10:06
Chapter 7
158
Table 5. Individual MRI features
MRI feature (numbers in parentheses reported by reference standard)
Grading accuracy (95% CI)
Mural thickness 69% (57%–78%)
Mural T2 signal 59% (50%–68%)
T1 enhancement 49% (39%–59%)
Enhancement pattern 65% (54%–75%)
Total length of disease 66% (55%–75%)
Comb sign (n=8) 74% (61%–84%)
Infiltrate (n= 4) 93% (88%–96%)
Abscess (n=1) 97% (92%–99%)
Fistula (n=5) 90% (81%–95%)
Severe stenosis (n=6) 83% (74%–90%)
CI, confidence interval; MRI, magnetic resonance imaging
Table 6. Comparison of grading accuracies between reader training and follow-up evaluation
with readers divided based on experience with MRE and abdominal MRI.
MREReader training, grading accuracy (95% CI)
Follow-up, grading accuracy* (95% CI)
P Value
Low experience 69% (95% CI: 56–79%) 75% (95% CI: 64–83%) 0.1
High experience 75% (95% CI: 63–84%) 71% (95% CI: 59–80%) 0.23
Abdominal MRI
Low experience 70% (95% CI: 58–79%) 69% (95% CI: 57–79%) 0.88
High experience 74% (95% CI: 62–83%) 77% (95% CI: 66–85%) 0.36
CI, confidence interval; MRE, magnetic resonance enterography or enteroclysis; MRI, mag-netic resonance imaging.* Follow-up grading accuracy, understaging, and overstaging did not differ between groups with low and high interim experience in MRE (P = .39, P = .68, and P = .61, respec-tively) and between groups with low and high interim experience with abdominal MRI (P = .06, P = .61, and P = .06).
Interim Experience
When readers were ranked based on interim experience with magnetic resonance
enterography/enteroclysis, the lower half (low interim experience) had evaluated
a mean of 2.7 (range 1–4) examinations per month, whereas the upper half (high
interim experience) had evaluated a mean of 9.7 (range 5–20) examinations per
month. Readers categorized as having low interim experience with abdominal
MRI had evaluated a mean of 9.7 (range 5–16) examinations per month, whereas
Proefschrift2018-new2.indb 158 13/9/18 10:06
Long-term performance of trained readers
159
readers with high interim experience had evaluated a mean of 52.1 (range 20–100)
examinations per month. Grading accuracies and P values between groups with low
and high interim experience with magnetic resonance enterography/enteroclysis
and abdominal MRI are presented in Table 6. No significant changes were seen
between reader training and follow-up evaluation for groups with low or high
experience in magnetic resonance enterography/enteroclysis (P = .1 and P = .23,
respectively) and low or high experience in abdominal MRI (P = .88 and P = .36,
respectively) (Table 6). Follow-up evaluation grading accuracy for the group with
low abdominal MRI experience was 69% (95% CI: 57%–79%) compared to 77% (95%
CI: 66%–85%) in the high experience group; this difference was not significant (P =
.06).
Grading Confidence
Grading confidence per case increased from a mean of 7.8 (95% CI: 7.5–8.2) at reader
training to 8.0 (95% CI: 7.7–8.3) at follow-up evaluation, although this increase was
not significant (P = .15).
DISCUSSION
In this study we found that radiologists and residents previously trained at grading
activity in Crohn disease patients (accuracy 72%) showed similar grading accuracy
(73%) at the follow-up evaluation. A decrease in overstaging was seen (19%–13%),
whereas understaging increased (9%–14%). We found a moderate to strong intra-
rater agreement for grading disease activity.
From our findings, we can conclude that readers maintain a consistent accuracy
for grading Crohn disease activity after an extended time period. No change in
grading accuracy was seen irrespective of the extent of interim experience,
including readers with high interim experience in abdominal MRI (although this
was close to the significance level of P = .06). This result might be related to the
small sample size and quite small mean difference between groups for experience
with magnetic resonance enterography/enteroclysis (seven magnetic resonance
enterography/enteroclysis examinations per month). It could be that in daily
practice there is only scarce feedback on clinical outcome of MRI cases of Crohn
disease (the pivotal feature of the initial reader training). This will depend on the
Proefschrift2018-new2.indb 159 13/9/18 10:06
Long-term performance of trained readers
159
readers with high interim experience had evaluated a mean of 52.1 (range 20–100)
examinations per month. Grading accuracies and P values between groups with low
and high interim experience with magnetic resonance enterography/enteroclysis
and abdominal MRI are presented in Table 6. No significant changes were seen
between reader training and follow-up evaluation for groups with low or high
experience in magnetic resonance enterography/enteroclysis (P = .1 and P = .23,
respectively) and low or high experience in abdominal MRI (P = .88 and P = .36,
respectively) (Table 6). Follow-up evaluation grading accuracy for the group with
low abdominal MRI experience was 69% (95% CI: 57%–79%) compared to 77% (95%
CI: 66%–85%) in the high experience group; this difference was not significant (P =
.06).
Grading Confidence
Grading confidence per case increased from a mean of 7.8 (95% CI: 7.5–8.2) at reader
training to 8.0 (95% CI: 7.7–8.3) at follow-up evaluation, although this increase was
not significant (P = .15).
DISCUSSION
In this study we found that radiologists and residents previously trained at grading
activity in Crohn disease patients (accuracy 72%) showed similar grading accuracy
(73%) at the follow-up evaluation. A decrease in overstaging was seen (19%–13%),
whereas understaging increased (9%–14%). We found a moderate to strong intra-
rater agreement for grading disease activity.
From our findings, we can conclude that readers maintain a consistent accuracy
for grading Crohn disease activity after an extended time period. No change in
grading accuracy was seen irrespective of the extent of interim experience,
including readers with high interim experience in abdominal MRI (although this
was close to the significance level of P = .06). This result might be related to the
small sample size and quite small mean difference between groups for experience
with magnetic resonance enterography/enteroclysis (seven magnetic resonance
enterography/enteroclysis examinations per month). It could be that in daily
practice there is only scarce feedback on clinical outcome of MRI cases of Crohn
disease (the pivotal feature of the initial reader training). This will depend on the
Proefschrift2018-new2.indb 159 13/9/18 10:06
Chapter 7
160
local setting including multidisciplinary team meetings and the perseverance of a
reader to obtain feedback. This might indicate that more dedicated training (either
a larger number of cases or another form of training) or consistent clinical feedback
is required for further improvement. We found significant changes in understaging
and overstaging, although these did not affect overall grading accuracy. One could
consider that these changes are related to differences in readers’ interim experience,
but we did not find such an association. Another explanation could be differences
in patient spectra based on the type of center where the readers were working
(academic, teaching, or general hospital), but we found these groups too small
for subanalysis. Furthermore, these data would be unreliable, as several readers
had shifted between hospitals between reader training and follow-up evaluation.
Our findings on intra-rater agreement are similar to results seen for interobserver
agreement in another study, which examined the same mural features that were
used in our study (3). Changes in understaging and overstaging between reader
training and follow-up evaluation certainly lowered intra-rater agreement, although
grading accuracy was not affected.
We found that during follow-up evaluation, readers achieved the highest accuracy
when grading no disease or severe disease (82% and 84%, respectively), whereas
mild and moderate disease were graded less accurately (62% and 61%, respectively).
Severe disease was graded with a high accuracy (84%), which is consistent with the
high accuracies (range: 83%–97%) seen for presence of complications, the most
common criterion for severe disease. These findings are important, as stenotic
and penetrating lesions have high predictive value for abdominal surgery (4).
We observed that readers were equally confident in their grading during reader
training and follow-up evaluation, leading us to conclude that grading confidence is
consistent after reader training.
Other studies have investigated reader training in different fields of gastrointestinal
imaging (5–7). However, to our knowledge, no studies have examined long-term
performance of trained readers. Although our results cannot be directly extrapolated
to other fields of training, they do underline the positive effects of reader training
and its importance in clinical practice.
Our study had several limitations. First, not all readers from initial reader training
participated in the follow-up evaluation. Although this reduced the amount of
Proefschrift2018-new2.indb 160 13/9/18 10:06
Chapter 7
160
local setting including multidisciplinary team meetings and the perseverance of a
reader to obtain feedback. This might indicate that more dedicated training (either
a larger number of cases or another form of training) or consistent clinical feedback
is required for further improvement. We found significant changes in understaging
and overstaging, although these did not affect overall grading accuracy. One could
consider that these changes are related to differences in readers’ interim experience,
but we did not find such an association. Another explanation could be differences
in patient spectra based on the type of center where the readers were working
(academic, teaching, or general hospital), but we found these groups too small
for subanalysis. Furthermore, these data would be unreliable, as several readers
had shifted between hospitals between reader training and follow-up evaluation.
Our findings on intra-rater agreement are similar to results seen for interobserver
agreement in another study, which examined the same mural features that were
used in our study (3). Changes in understaging and overstaging between reader
training and follow-up evaluation certainly lowered intra-rater agreement, although
grading accuracy was not affected.
We found that during follow-up evaluation, readers achieved the highest accuracy
when grading no disease or severe disease (82% and 84%, respectively), whereas
mild and moderate disease were graded less accurately (62% and 61%, respectively).
Severe disease was graded with a high accuracy (84%), which is consistent with the
high accuracies (range: 83%–97%) seen for presence of complications, the most
common criterion for severe disease. These findings are important, as stenotic
and penetrating lesions have high predictive value for abdominal surgery (4).
We observed that readers were equally confident in their grading during reader
training and follow-up evaluation, leading us to conclude that grading confidence is
consistent after reader training.
Other studies have investigated reader training in different fields of gastrointestinal
imaging (5–7). However, to our knowledge, no studies have examined long-term
performance of trained readers. Although our results cannot be directly extrapolated
to other fields of training, they do underline the positive effects of reader training
and its importance in clinical practice.
Our study had several limitations. First, not all readers from initial reader training
participated in the follow-up evaluation. Although this reduced the amount of
Proefschrift2018-new2.indb 160 13/9/18 10:06
Long-term performance of trained readers
161
gathered data, we do not assume this to affect the validity of our results. Biased
selection of readers was unlikely as most readers (24 out of 31) had enlisted to
participate in the follow-up evaluation, but some were unavailable at the time of
follow-up sessions. In addition, both participants and nonparticipants of the follow-
up evaluation showed equal grading accuracy at reader training. Another limitation
of our study was that only 25 MRI cases were used for evaluation, which presents a
limited patient population. Although patients were distributed fairly equally over the
four categories of disease activity, only five patients had one or more complications
on MRI, which could partly explain the high accuracies found for these features.
Furthermore, only four patients had colonic involvement of Crohn disease. In this
study, we used the same qualitative scoring system as was used in reader training
and which is partially based on a validated scoring system. Other scoring systems
have been reported in the literature, such as the quantitative Magnetic Resonance
Index of Activity (MaRIA) (8). Although the MaRIA score has been validated
in several studies, it requires a time-effort not suitable for clinical practice and
additionally does not require readers to fully evaluate all disease features such as
fistulas and abscesses.
Limitations for both reader training and follow-up evaluation include the use of an
expert panel as reference standard, and a potential limitation is not using a rectal
enema for colonic distension. We found an expert panel to be the most suitable
reference standard, recognizing the limitations of other techniques. Endoscopy will
only provide luminal evaluation of the colon and terminal ileum, whereas pathologic
assessment of surgical specimens can only be acquired in a subset of patients with
severe or fibrotic disease activity. A limitation is the lack of independent evaluation
by the expert panel, before reaching consensus, as this would have allowed us to
assess the robustness of our reference standard (9). Lastly, no rectal enema was
used, as this is not a standard technique for magnetic resonance enterography/
enteroclysis examinations.
In conclusion, we found that readers have a consistent longterm accuracy for
grading activity in Crohn disease patients after case-based reader training with
direct feedback. Substantial interim experience with MR enterography/enteroclysis
or abdominal MRI did not improve overall reading results.
Proefschrift2018-new2.indb 161 13/9/18 10:06
Long-term performance of trained readers
161
gathered data, we do not assume this to affect the validity of our results. Biased
selection of readers was unlikely as most readers (24 out of 31) had enlisted to
participate in the follow-up evaluation, but some were unavailable at the time of
follow-up sessions. In addition, both participants and nonparticipants of the follow-
up evaluation showed equal grading accuracy at reader training. Another limitation
of our study was that only 25 MRI cases were used for evaluation, which presents a
limited patient population. Although patients were distributed fairly equally over the
four categories of disease activity, only five patients had one or more complications
on MRI, which could partly explain the high accuracies found for these features.
Furthermore, only four patients had colonic involvement of Crohn disease. In this
study, we used the same qualitative scoring system as was used in reader training
and which is partially based on a validated scoring system. Other scoring systems
have been reported in the literature, such as the quantitative Magnetic Resonance
Index of Activity (MaRIA) (8). Although the MaRIA score has been validated
in several studies, it requires a time-effort not suitable for clinical practice and
additionally does not require readers to fully evaluate all disease features such as
fistulas and abscesses.
Limitations for both reader training and follow-up evaluation include the use of an
expert panel as reference standard, and a potential limitation is not using a rectal
enema for colonic distension. We found an expert panel to be the most suitable
reference standard, recognizing the limitations of other techniques. Endoscopy will
only provide luminal evaluation of the colon and terminal ileum, whereas pathologic
assessment of surgical specimens can only be acquired in a subset of patients with
severe or fibrotic disease activity. A limitation is the lack of independent evaluation
by the expert panel, before reaching consensus, as this would have allowed us to
assess the robustness of our reference standard (9). Lastly, no rectal enema was
used, as this is not a standard technique for magnetic resonance enterography/
enteroclysis examinations.
In conclusion, we found that readers have a consistent longterm accuracy for
grading activity in Crohn disease patients after case-based reader training with
direct feedback. Substantial interim experience with MR enterography/enteroclysis
or abdominal MRI did not improve overall reading results.
Proefschrift2018-new2.indb 161 13/9/18 10:06
Chapter 7
162
REFERENCES
1. Tielbeek JA, Bipat S, Boellaard TN, et al. Training readers to improve their
accuracy in grading Crohn’s disease activity on MRI. Eur Radiol 2014; 24:1059–
1067.
2. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s
disease assessed by MRI enterography: derivation and histopathological
validation of an MR-based activity index. Eur J Radiol 2012; 81:2080–2088.
3. Tielbeek JA, Makanyanga JC, Bipat S, et al. Grading Crohn disease activity
with MRI: interobserver variability of MRI features, MRI scoring of severity,
and correlation with Crohn disease endoscopic index of severity. AJR Am J
Roentgenol 2013; 201:1220–1228.
4. Jauregui-Amezaga A, Rimola J, Ordas I, et al. Value of endoscopy and MRI
for predicting intestinal surgery in patients with Crohn’s disease in the era of
biologics. Gut 2015; 64:1397–1402.
5. Boellaard TN, Nio CY, Bossuyt PM, et al. Can radiographers be trained to triage
CT colonography for extracolonic findings? Eur Radiol 2012; 22:2780–2789.
6. Leeuwenburgh MM, Wiarda BM, Bipat S, et al. Acute appendicitis on abdominal
MR images: training readers to improve diagnostic accuracy. Radiology 2012;
264:455–463.
7. Liedenbaum MH, Bipat S, Bossuyt PM, et al. Evaluation of a standardized CT
colonography training program for novice readers. Radiology 2011; 258:477–
487.
8. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment
of disease activity and severity in ileocolonic Crohn’s disease. Gut 2009; 58:1113–
1120.
9. Bankier AA, Levine D, Halpern EF, et al. Consensus interpretation in imaging
research: is there a better way? Radiology 2010; 257:14–17.
Proefschrift2018-new2.indb 162 13/9/18 10:06
Chapter 7
162
REFERENCES
1. Tielbeek JA, Bipat S, Boellaard TN, et al. Training readers to improve their
accuracy in grading Crohn’s disease activity on MRI. Eur Radiol 2014; 24:1059–
1067.
2. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s
disease assessed by MRI enterography: derivation and histopathological
validation of an MR-based activity index. Eur J Radiol 2012; 81:2080–2088.
3. Tielbeek JA, Makanyanga JC, Bipat S, et al. Grading Crohn disease activity
with MRI: interobserver variability of MRI features, MRI scoring of severity,
and correlation with Crohn disease endoscopic index of severity. AJR Am J
Roentgenol 2013; 201:1220–1228.
4. Jauregui-Amezaga A, Rimola J, Ordas I, et al. Value of endoscopy and MRI
for predicting intestinal surgery in patients with Crohn’s disease in the era of
biologics. Gut 2015; 64:1397–1402.
5. Boellaard TN, Nio CY, Bossuyt PM, et al. Can radiographers be trained to triage
CT colonography for extracolonic findings? Eur Radiol 2012; 22:2780–2789.
6. Leeuwenburgh MM, Wiarda BM, Bipat S, et al. Acute appendicitis on abdominal
MR images: training readers to improve diagnostic accuracy. Radiology 2012;
264:455–463.
7. Liedenbaum MH, Bipat S, Bossuyt PM, et al. Evaluation of a standardized CT
colonography training program for novice readers. Radiology 2011; 258:477–
487.
8. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment
of disease activity and severity in ileocolonic Crohn’s disease. Gut 2009; 58:1113–
1120.
9. Bankier AA, Levine D, Halpern EF, et al. Consensus interpretation in imaging
research: is there a better way? Radiology 2010; 257:14–17.
Proefschrift2018-new2.indb 162 13/9/18 10:06
Proefschrift2018-new2.indb 163 13/9/18 10:06 Proefschrift2018-new2.indb 163 13/9/18 10:06
Proefschrift2018-new2.indb 164 13/9/18 10:06 Proefschrift2018-new2.indb 164 13/9/18 10:06
CHAPTER 8
General discussion
Proefschrift2018-new2.indb 165 13/9/18 10:06
CHAPTER 8
General discussion
Proefschrift2018-new2.indb 165 13/9/18 10:06
Chapter 8
166
GENERAL DISCUSSION
The aim of this thesis was to examine both currently available and newly developed
methods for the evaluation of Crohn’s disease using MRI. We reviewed the place
of MRI within the scope of alternative imaging techniques, the available methods
for disease activity grading, and technological developments such as new MRI
sequences and the use of semiautomatic measurements. A number of these studies
(chapters 3–6) were performed under the VIGOR++ project, an international
collaboration aimed at developing new techniques of image analysis for Crohn’s
disease.
Imaging techniques
A number of imaging techniques have been studied for Crohn’s disease, such as
computed tomography (CT), magnetic resonance imaging (MRI), ultrasonography
(US) and scintigraphy. While for acute imaging, the important role of CT and US
is well established, the role of imaging techniques in the elective setting, such as
detection and grading of disease activity, have been less clearly defined. Although
MRI has shown high accuracy in these types of examinations, few studies have
been performed comparing MRI to other imaging techniques [1,2]. In chapter 2,
we compared a number of imaging techniques (CT, MRI, US and scintigraphy) in
grading of Crohn’s disease by performing a meta-analysis. We reported both per-
patient and per-segment data, as each plays an important role in disease evaluation.
Per-patient data for CT, MRI, US and scintigraphy showed overall grading accuracies
of 86%, 84%, 44% and 40%, respectively, while per-segment data showed overall
grading accuracies of 87%, 75%, 66% and 86%, respectively. Our results led us to
conclude that MRI should be the preferred modality for grading of Crohn’s disease
activity, due to its high grading accuracy and absence of ionizing radiation. These
are valued strong points in the often countless imaging examinations that Crohn’s
disease patients have to undergo during their lifespan. Based on more recent studies,
a new ECCO-ESGAR guideline for diagnostic and monitoring methods for IBD has
extended the role of US in diagnosis, monitoring and post-operative imaging of
small bowel disease (unpublished guideline). The authors suggest both modalities
are adequate and, in general, the choice for MRI or US can be made based on
local availability and expertise. However, individual patient characteristics should be
considered as MRI allows for better visualization of obese patients and deep-seated
Proefschrift2018-new2.indb 166 13/9/18 10:06
Chapter 8
166
GENERAL DISCUSSION
The aim of this thesis was to examine both currently available and newly developed
methods for the evaluation of Crohn’s disease using MRI. We reviewed the place
of MRI within the scope of alternative imaging techniques, the available methods
for disease activity grading, and technological developments such as new MRI
sequences and the use of semiautomatic measurements. A number of these studies
(chapters 3–6) were performed under the VIGOR++ project, an international
collaboration aimed at developing new techniques of image analysis for Crohn’s
disease.
Imaging techniques
A number of imaging techniques have been studied for Crohn’s disease, such as
computed tomography (CT), magnetic resonance imaging (MRI), ultrasonography
(US) and scintigraphy. While for acute imaging, the important role of CT and US
is well established, the role of imaging techniques in the elective setting, such as
detection and grading of disease activity, have been less clearly defined. Although
MRI has shown high accuracy in these types of examinations, few studies have
been performed comparing MRI to other imaging techniques [1,2]. In chapter 2,
we compared a number of imaging techniques (CT, MRI, US and scintigraphy) in
grading of Crohn’s disease by performing a meta-analysis. We reported both per-
patient and per-segment data, as each plays an important role in disease evaluation.
Per-patient data for CT, MRI, US and scintigraphy showed overall grading accuracies
of 86%, 84%, 44% and 40%, respectively, while per-segment data showed overall
grading accuracies of 87%, 75%, 66% and 86%, respectively. Our results led us to
conclude that MRI should be the preferred modality for grading of Crohn’s disease
activity, due to its high grading accuracy and absence of ionizing radiation. These
are valued strong points in the often countless imaging examinations that Crohn’s
disease patients have to undergo during their lifespan. Based on more recent studies,
a new ECCO-ESGAR guideline for diagnostic and monitoring methods for IBD has
extended the role of US in diagnosis, monitoring and post-operative imaging of
small bowel disease (unpublished guideline). The authors suggest both modalities
are adequate and, in general, the choice for MRI or US can be made based on
local availability and expertise. However, individual patient characteristics should be
considered as MRI allows for better visualization of obese patients and deep-seated
Proefschrift2018-new2.indb 166 13/9/18 10:06
General discussion
167
disease in the pelvic area. A recent prospective multi-center trial found that MRI
achieves superior sensitivity and specificity for the detection of activity and extent
of small bowel Crohn’s disease in newly diagnosed patients and those suffering
relapse, when compared to intestinal ultrasound [3]. However, the authors note that
future research should investigate the role of imaging in targeted follow-up or in
patients without an established diagnosis of Crohn’s disease. As mentioned before,
the advantages of MRI are limited in emergency situations or immobile patients. In
these situations, use of CT or US is appropriate.
MRI activity scores
In recent years, the use of MRI activity scores to define and grade Crohn’s
disease activity has become increasingly popular. Several MRI activity scores
have been developed with the goal of providing more accurate evaluation with
high reproducibility [4–6]. Although these scores have been externally validated,
a comprehensive comparison had not been yet performed [7–9]. In chapter
3, these four MRI activity scores – the Magnetic Resonance Index of Activity
(MaRIA), Clermont score, London score and Crohn’s disease MRI index (CDMI)
– were compared using 114 patients of the VIGOR++ database [4–6]. All scores
showed high diagnostic accuracy for active and severe disease, moderate-to-
strong correlation to the endoscopic/histopathologic reference standard and good
reproducibility (intraclass correlation coefficient (ICC): 0.72–0.78). The MaRIA,
Clermont and London scores have the benefit of having validated cut-off values for
severe disease activity, an important characteristic for patient selection in clinical
drug trials and therapeutic guidance in general. Remaining choices can be made on
personal preference. Based on our experience, we suspect there to be differences
in time efficiency due to the different uses of quantitative measurements, although
this was not examined in our study. The similarity in the performance of the scores
may derive from comparable underlying features: bowel wall thickness, contrast
enhancement and mural T2 signal (or edema) are each incorporated in two or
more scores. The importance of these features is reflected by a high correlation to
endoscopic disease activity seen in our study.
Proefschrift2018-new2.indb 167 13/9/18 10:06
General discussion
167
disease in the pelvic area. A recent prospective multi-center trial found that MRI
achieves superior sensitivity and specificity for the detection of activity and extent
of small bowel Crohn’s disease in newly diagnosed patients and those suffering
relapse, when compared to intestinal ultrasound [3]. However, the authors note that
future research should investigate the role of imaging in targeted follow-up or in
patients without an established diagnosis of Crohn’s disease. As mentioned before,
the advantages of MRI are limited in emergency situations or immobile patients. In
these situations, use of CT or US is appropriate.
MRI activity scores
In recent years, the use of MRI activity scores to define and grade Crohn’s
disease activity has become increasingly popular. Several MRI activity scores
have been developed with the goal of providing more accurate evaluation with
high reproducibility [4–6]. Although these scores have been externally validated,
a comprehensive comparison had not been yet performed [7–9]. In chapter
3, these four MRI activity scores – the Magnetic Resonance Index of Activity
(MaRIA), Clermont score, London score and Crohn’s disease MRI index (CDMI)
– were compared using 114 patients of the VIGOR++ database [4–6]. All scores
showed high diagnostic accuracy for active and severe disease, moderate-to-
strong correlation to the endoscopic/histopathologic reference standard and good
reproducibility (intraclass correlation coefficient (ICC): 0.72–0.78). The MaRIA,
Clermont and London scores have the benefit of having validated cut-off values for
severe disease activity, an important characteristic for patient selection in clinical
drug trials and therapeutic guidance in general. Remaining choices can be made on
personal preference. Based on our experience, we suspect there to be differences
in time efficiency due to the different uses of quantitative measurements, although
this was not examined in our study. The similarity in the performance of the scores
may derive from comparable underlying features: bowel wall thickness, contrast
enhancement and mural T2 signal (or edema) are each incorporated in two or
more scores. The importance of these features is reflected by a high correlation to
endoscopic disease activity seen in our study.
Proefschrift2018-new2.indb 167 13/9/18 10:06
Chapter 8
168
Semiautomatic measurements
Despite the promising results of MRI activity scores, varying reproducibility of
included MRI features is reported in literature [9,10]. Improvements of individual
features could lead to a further increase in reproducibility of MRI scores. One
proposed method to achieve this is the use of semiautomatic measurements [11],
which have shown promising results on an individual basis in previous studies
[12,13]. In chapter 4, we have developed and validated a new MRI activity score
(“the VIGOR score”) using both semiautomatic MRI measurements and subjective
radiologist observation of mural T2 signal. The VIGOR score was validated in 106
patients from the VIGOR database and compared to two existing activity scores
– the MaRIA and London scores – and a newly developed subjective score. All
scores showed similar correlation to CDEIS (r=0.34–0.59), although the VIGOR
score showed increased reproducibility compared to other activity scores (ICC:
0.82 vs 0.44–0.59, respectively). These characteristics make the VIGOR score a
suitable candidate for use in therapy evaluation and monitoring of disease activity,
which require robust scoring for repeated examinations using various readers.
Currently, therapeutic monitoring and in turn clinical trials attach major value to
endoscopic endpoints, such as the well-established mucosal healing [14]. Recent
studies have shown promising results for the use MRI scores, such as the MaRIA
and Clermont score, as outcome measures for therapy [15,16]. The VIGOR score
shows promise as an objective measure of Crohn’s disease activity, but this should
be confirmed by performing longitudinal studies to evaluate its responsiveness and
to determine which quantitative change predicts therapeutic improvement. Further,
the semiautomatic method should be applied to different sequences, for example
the measurement of DWI and T2 mural signal intensities. Although this would be a
technically challenging process, the reward could be substantial, as reproducibility is
currently considered one of the limiting factors in the use of these sequences [9,17].
Furthermore, automation of such measurements would facilitate investigations into
distinguishing inflammation and fibrosis, a key issue in Crohn’s disease assessment,
albeit not addressed in this thesis.
The subject of automation in radiology is not confined to MRI or Crohn’s disease.
Many discussions around this subject question whether automation will at some point
replace the radiologist. For the near future, as it would be unfounded to speculate
further beyond, the role of automation should be considered as complementary to
Proefschrift2018-new2.indb 168 13/9/18 10:06
Chapter 8
168
Semiautomatic measurements
Despite the promising results of MRI activity scores, varying reproducibility of
included MRI features is reported in literature [9,10]. Improvements of individual
features could lead to a further increase in reproducibility of MRI scores. One
proposed method to achieve this is the use of semiautomatic measurements [11],
which have shown promising results on an individual basis in previous studies
[12,13]. In chapter 4, we have developed and validated a new MRI activity score
(“the VIGOR score”) using both semiautomatic MRI measurements and subjective
radiologist observation of mural T2 signal. The VIGOR score was validated in 106
patients from the VIGOR database and compared to two existing activity scores
– the MaRIA and London scores – and a newly developed subjective score. All
scores showed similar correlation to CDEIS (r=0.34–0.59), although the VIGOR
score showed increased reproducibility compared to other activity scores (ICC:
0.82 vs 0.44–0.59, respectively). These characteristics make the VIGOR score a
suitable candidate for use in therapy evaluation and monitoring of disease activity,
which require robust scoring for repeated examinations using various readers.
Currently, therapeutic monitoring and in turn clinical trials attach major value to
endoscopic endpoints, such as the well-established mucosal healing [14]. Recent
studies have shown promising results for the use MRI scores, such as the MaRIA
and Clermont score, as outcome measures for therapy [15,16]. The VIGOR score
shows promise as an objective measure of Crohn’s disease activity, but this should
be confirmed by performing longitudinal studies to evaluate its responsiveness and
to determine which quantitative change predicts therapeutic improvement. Further,
the semiautomatic method should be applied to different sequences, for example
the measurement of DWI and T2 mural signal intensities. Although this would be a
technically challenging process, the reward could be substantial, as reproducibility is
currently considered one of the limiting factors in the use of these sequences [9,17].
Furthermore, automation of such measurements would facilitate investigations into
distinguishing inflammation and fibrosis, a key issue in Crohn’s disease assessment,
albeit not addressed in this thesis.
The subject of automation in radiology is not confined to MRI or Crohn’s disease.
Many discussions around this subject question whether automation will at some point
replace the radiologist. For the near future, as it would be unfounded to speculate
further beyond, the role of automation should be considered as complementary to
Proefschrift2018-new2.indb 168 13/9/18 10:06
General discussion
169
the work of the radiologist, rather than as a replacement. Aside from optimizing
radiological assessment, automation can be used to relieve the radiologist from
repetitive and time-consuming measurements.
New MRI sequences
Diffusion-weighted imaging (DWI) has been investigated for evaluation of Crohn’s
disease, due to its interesting contrast properties for detection of inflammation.
Previously, the addition of DWI to conventional MRI has not resulted in increased
diagnostic benefit [18]. However, a recent study showed that DW-MRI was non-
inferior to contrast-enhanced (CE)-MRI in detection of small bowel disease in
Crohn’s disease patients, although a considerable discrepancy was found in
diagnosis of penetrating complications [19]. In chapter 5, contrast-enhanced (CE)-
MRI and DW-MRI were compared for diagnosis and grading of bowel inflammation.
Although similar levels of diagnostic and grading accuracy were found, higher
levels of confidence were seen for CE-MRI. Additionally, discrepancies were seen for
diagnosis of penetrating complications in favor of CE-MRI. Considering our results,
we would advise the use of CE-MRI for routine examinations, although DW-MRI
can be seen as a good alternative for cases with contraindications for intravenous
contrast use. Future studies and perhaps a meta-analysis of the previous studies will
be necessary to finally determine the usefulness of DW-MRI, especially regarding
the diagnosis of penetrating disease. At the same time, we recognize that technical
improvements in DWI sequences will continue in the coming years; e.g. use of intra-
voxel incoherent motion (IVIM) model parameters might provide measurements
superior to the currently used apparent diffusion coefficient (ADC), which shows
considerable variation due to differences between machines and scanning protocols
[20].
Assessment of Crohn’s disease activity has largely relied on the use of static images.
Improved fast acquisition of MR images has now facilitated examination of bowel
motility. A previous study showed a negative correlation between terminal ileal
motility and the eAIS (r=-0.52), and a significant difference in motility between
inflamed and non-inflamed segments [21]. In chapter 6, bowel motility of the
terminal ileum is evaluated in 82 patients using both histopathologic and endoscopic
reference standards (eAIS and CDEIS). Motility showed a sensitivity and specificity
of 92% and 74% against eAIS, and 93% and 67% against CDEIS, respectively.
Proefschrift2018-new2.indb 169 13/9/18 10:06
General discussion
169
the work of the radiologist, rather than as a replacement. Aside from optimizing
radiological assessment, automation can be used to relieve the radiologist from
repetitive and time-consuming measurements.
New MRI sequences
Diffusion-weighted imaging (DWI) has been investigated for evaluation of Crohn’s
disease, due to its interesting contrast properties for detection of inflammation.
Previously, the addition of DWI to conventional MRI has not resulted in increased
diagnostic benefit [18]. However, a recent study showed that DW-MRI was non-
inferior to contrast-enhanced (CE)-MRI in detection of small bowel disease in
Crohn’s disease patients, although a considerable discrepancy was found in
diagnosis of penetrating complications [19]. In chapter 5, contrast-enhanced (CE)-
MRI and DW-MRI were compared for diagnosis and grading of bowel inflammation.
Although similar levels of diagnostic and grading accuracy were found, higher
levels of confidence were seen for CE-MRI. Additionally, discrepancies were seen for
diagnosis of penetrating complications in favor of CE-MRI. Considering our results,
we would advise the use of CE-MRI for routine examinations, although DW-MRI
can be seen as a good alternative for cases with contraindications for intravenous
contrast use. Future studies and perhaps a meta-analysis of the previous studies will
be necessary to finally determine the usefulness of DW-MRI, especially regarding
the diagnosis of penetrating disease. At the same time, we recognize that technical
improvements in DWI sequences will continue in the coming years; e.g. use of intra-
voxel incoherent motion (IVIM) model parameters might provide measurements
superior to the currently used apparent diffusion coefficient (ADC), which shows
considerable variation due to differences between machines and scanning protocols
[20].
Assessment of Crohn’s disease activity has largely relied on the use of static images.
Improved fast acquisition of MR images has now facilitated examination of bowel
motility. A previous study showed a negative correlation between terminal ileal
motility and the eAIS (r=-0.52), and a significant difference in motility between
inflamed and non-inflamed segments [21]. In chapter 6, bowel motility of the
terminal ileum is evaluated in 82 patients using both histopathologic and endoscopic
reference standards (eAIS and CDEIS). Motility showed a sensitivity and specificity
of 92% and 74% against eAIS, and 93% and 67% against CDEIS, respectively.
Proefschrift2018-new2.indb 169 13/9/18 10:06
Chapter 8
170
Motility showed negative correlation to eAIS (r=-0.61) and CDEIS (r=-0.59). These
results indicate that quantified motility is a potential biomarker for endoscopic and
histopathologic disease activity on MRI. Motility imaging is a relatively new field of
MRI imaging, for which research has concentrated on improving acquisition speed,
while sustaining spatial resolution. Recently, the use of motility imaging has been
explored for grading of Crohn’s disease inflammation. The current study provides a
validation of quantified motility assessment against endoscopic and histopathologic
reference standards. The results support the use of motility features as imaging
biomarkers of Crohn’s disease inflammation. Additionally, disease evaluation might
be further improved by including quantified motility into MRI activity scores, which
should be investigated in future studies.
MRI training
Previous research has shown that inexperienced readers can be trained in grading
Crohn’s disease on MRI using a standardized scoring system [22,23]. In chapter 7,
we evaluated the long-term performance of readers who previously participated
in MRI reader training. Twenty-five MRI cases from the initial reader training were
presented to 14 readers (out of 31 original participants of reader training). Results
showed an overall grading accuracy of 73% at the follow-up evaluation, which is
similar to the 72% accuracy at the initial reader training. A decrease in overstaging
was seen (19% to 13%), while an increase in understaging occurred (9% to 14%).
Our study has shown that long-term performance is consistent with accuracy
directly after reader training and is not influenced by interim experience with MR
enterography and abdominal MRI. To our opinion this underlines the importance of
the direct feedback provided during MRI training, in contrast to clinical practice, in
which clinical, endoscopic or surgical feedback is often limited or delayed.
Reference standards
A number of reference standards are available for small bowel and colonic Crohn’s
disease imaging, namely ileocolonoscopy, histopathologic biopsies or surgical
specimens. Luminal evaluation by ileocolonoscopy has considerable limitations as
a reference standard, when we attempt to compare it to transmural and extramural
disease parameters on MRI. The use of surgical specimens would be preferable,
but this carries its own bias in terms of a selective disease spectrum. Biopsies
carry an interest as a reference standard for their specific and objective measure
Proefschrift2018-new2.indb 170 13/9/18 10:06
Chapter 8
170
Motility showed negative correlation to eAIS (r=-0.61) and CDEIS (r=-0.59). These
results indicate that quantified motility is a potential biomarker for endoscopic and
histopathologic disease activity on MRI. Motility imaging is a relatively new field of
MRI imaging, for which research has concentrated on improving acquisition speed,
while sustaining spatial resolution. Recently, the use of motility imaging has been
explored for grading of Crohn’s disease inflammation. The current study provides a
validation of quantified motility assessment against endoscopic and histopathologic
reference standards. The results support the use of motility features as imaging
biomarkers of Crohn’s disease inflammation. Additionally, disease evaluation might
be further improved by including quantified motility into MRI activity scores, which
should be investigated in future studies.
MRI training
Previous research has shown that inexperienced readers can be trained in grading
Crohn’s disease on MRI using a standardized scoring system [22,23]. In chapter 7,
we evaluated the long-term performance of readers who previously participated
in MRI reader training. Twenty-five MRI cases from the initial reader training were
presented to 14 readers (out of 31 original participants of reader training). Results
showed an overall grading accuracy of 73% at the follow-up evaluation, which is
similar to the 72% accuracy at the initial reader training. A decrease in overstaging
was seen (19% to 13%), while an increase in understaging occurred (9% to 14%).
Our study has shown that long-term performance is consistent with accuracy
directly after reader training and is not influenced by interim experience with MR
enterography and abdominal MRI. To our opinion this underlines the importance of
the direct feedback provided during MRI training, in contrast to clinical practice, in
which clinical, endoscopic or surgical feedback is often limited or delayed.
Reference standards
A number of reference standards are available for small bowel and colonic Crohn’s
disease imaging, namely ileocolonoscopy, histopathologic biopsies or surgical
specimens. Luminal evaluation by ileocolonoscopy has considerable limitations as
a reference standard, when we attempt to compare it to transmural and extramural
disease parameters on MRI. The use of surgical specimens would be preferable,
but this carries its own bias in terms of a selective disease spectrum. Biopsies
carry an interest as a reference standard for their specific and objective measure
Proefschrift2018-new2.indb 170 13/9/18 10:06
General discussion
171
of inflammatory activity, but cannot substitute the macro-visualization offered by
other techniques. No simple solution is at hand, but it is important to remain aware
of the suboptimal character of currently used reference standards for MRI.
Future research
Areas of potential future research can be divided in a number of areas. The use of
MRI activity scores as imaging biomarkers for therapeutic monitoring should be
further investigated. There is a scarcity of high-quality longitudinal MRI studies in
Crohn’s disease, due to the large variation in therapeutic methods and individual
strategies, which limit patient numbers. The use of MRI scores as a therapeutic
marker in clinical drug trials could help to establish a foundation for longitudinal
data acquisition and to stimulate its use in clinical practice. The concept of MRI
automation can be expanded to other measurements on T2-weighted and DWI
sequences to facilitate investigations into distinguishing inflammation and fibrosis.
Future research should not only be aimed at developing new technologies, but also
at integrating currently available methods. A large array of sequences and disease
features is available, of which the strength lies in a combined approach.
Conclusion
In recent years, MRI has been the most actively investigated radiologic technique for
evaluation of Crohn’s disease of the small bowel and colon. MRI activity scores can
be used to accurately diagnose and grade disease activity, while new semiautomatic
measurements can further improve reproducibility. New MRI sequences can help
improve evaluation and provide safer alternatives to contrast use in vulnerable
patient groups. MRI evaluation of Crohn’s disease has evolved rapidly over the last
decade and we should continue this path of technological and clinical advancement.
Proefschrift2018-new2.indb 171 13/9/18 10:06
General discussion
171
of inflammatory activity, but cannot substitute the macro-visualization offered by
other techniques. No simple solution is at hand, but it is important to remain aware
of the suboptimal character of currently used reference standards for MRI.
Future research
Areas of potential future research can be divided in a number of areas. The use of
MRI activity scores as imaging biomarkers for therapeutic monitoring should be
further investigated. There is a scarcity of high-quality longitudinal MRI studies in
Crohn’s disease, due to the large variation in therapeutic methods and individual
strategies, which limit patient numbers. The use of MRI scores as a therapeutic
marker in clinical drug trials could help to establish a foundation for longitudinal
data acquisition and to stimulate its use in clinical practice. The concept of MRI
automation can be expanded to other measurements on T2-weighted and DWI
sequences to facilitate investigations into distinguishing inflammation and fibrosis.
Future research should not only be aimed at developing new technologies, but also
at integrating currently available methods. A large array of sequences and disease
features is available, of which the strength lies in a combined approach.
Conclusion
In recent years, MRI has been the most actively investigated radiologic technique for
evaluation of Crohn’s disease of the small bowel and colon. MRI activity scores can
be used to accurately diagnose and grade disease activity, while new semiautomatic
measurements can further improve reproducibility. New MRI sequences can help
improve evaluation and provide safer alternatives to contrast use in vulnerable
patient groups. MRI evaluation of Crohn’s disease has evolved rapidly over the last
decade and we should continue this path of technological and clinical advancement.
Proefschrift2018-new2.indb 171 13/9/18 10:06
Chapter 8
172
REFERENCES 1. Horsthuis K, Bipat S, Bennink RJ, et al. Inflammatory bowel disease diagnosed with US,
MR, scintigraphy, and CT: meta-analysis of prospective studies. Radiology 2008;247:64–
79.
2. Horsthuis K, Bipat S, Stokkers PCF, et al. Magnetic resonance imaging for evaluation of
disease activity in Crohn’s disease: a systematic review. Eur Radiol 2009;19:1450–60.
3. Taylor S, Mallett S, Bhatnagar G, et al. Diagnostic accuracy for the extent and activity of
newly diagnosed and relapsed Crohn’s disease: a multicentre prospective comparison of
magnetic resonance enterography and small bowel ultrasound – the METRIC trial. Lancet
Gastroenterol Hepatol 2018; accepted.
4. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment of
disease activity and severity in ileocolonic Crohn’s disease. Gut 2009;58:1113–20.
5. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: Derivation and histopathological validation of an MR-
based activity index. Eur J Radiol 2012;81:2080–8.
6. Buisson A, Joubert A, Montoriol PF, et al. Diffusion-weighted magnetic resonance imaging
for detecting and assessing ileal inflammation in Crohn’s disease. Aliment Pharmacol Ther
2013;37:537–45.
7. Rimola J, Ordás I, Rodriguez S, et al. Magnetic resonance imaging for evaluation of Crohn’s
disease: Validation of parameters of severity and quantitative index of activity. Inflamm
Bowel Dis 2011;17:1759–68.
8. Hordonneau C, Buisson A, Scanzi J, et al. Diffusion-weighted magnetic resonance
imaging in ileocolonic Crohn’s disease: validation of quantitative index of activity. Am J
Gastroenterol 2014;109:89–98.
9. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading crohn disease activity with MRI:
Interobserver variability of MRI features, MRI scoring of severity, and correlation with
crohn disease endoscopic index of severity. Am J Roentgenol 2013;201:1220–8.
10. Ziech MLW, Bipat S, Roelofs JJTH, et al. Retrospective comparison of magnetic
resonance imaging features and histopathology in Crohn’s disease patients. Eur J Radiol
2011;80:e299–305.
11. Tielbeek JAW, Vos FM, Stoker J. A computer-assisted model for detection of MRI signs of
Crohn’s disease activity: Future or fiction? Abdom Imaging 2012;37:967–73.
12. Naziroglu RE, Puylaert CAJ, Tielbeek JAW, et al. Semi-automatic bowel wall thickness
measurements on MR enterography in patients with Crohn’s disease. Br J Radiol
2017;:20160654.
13. Li Z, Tielbeek JAW, Caan MWA, et al. Expiration-Phase Template-Based Motion Correction
of Free-Breathing Abdominal Dynamic Contrast Enhanced MRI. IEEE Trans Biomed Eng
2015;62:1215–25.
14. Dave M, Loftus E V. Mucosal healing in inflammatory bowel disease-a true paradigm of
success? Gastroenterol Hepatol (N Y) 2012;8:29–38.
15. Buisson A, Hordonneau C, Goutte M, et al. Diffusion-weighted magnetic resonance
enterocolonography in predicting remission after anti-TNF induction therapy in Crohn’s
disease. Dig Liver Dis 2016;48:260–6.
16. Stoppino LP, Della Valle N, Rizzi S, et al. Magnetic resonance enterography changes after
antibody to tumor necrosis factor (anti-TNF) alpha therapy in Crohn’s disease: Correlation
with SES-CD and clinical-biological markers. BMC Med Imaging 2016;16.
Proefschrift2018-new2.indb 172 13/9/18 10:06
Chapter 8
172
REFERENCES 1. Horsthuis K, Bipat S, Bennink RJ, et al. Inflammatory bowel disease diagnosed with US,
MR, scintigraphy, and CT: meta-analysis of prospective studies. Radiology 2008;247:64–
79.
2. Horsthuis K, Bipat S, Stokkers PCF, et al. Magnetic resonance imaging for evaluation of
disease activity in Crohn’s disease: a systematic review. Eur Radiol 2009;19:1450–60.
3. Taylor S, Mallett S, Bhatnagar G, et al. Diagnostic accuracy for the extent and activity of
newly diagnosed and relapsed Crohn’s disease: a multicentre prospective comparison of
magnetic resonance enterography and small bowel ultrasound – the METRIC trial. Lancet
Gastroenterol Hepatol 2018; accepted.
4. Rimola J, Rodriguez S, Garcia-Bosch O, et al. Magnetic resonance for assessment of
disease activity and severity in ileocolonic Crohn’s disease. Gut 2009;58:1113–20.
5. Steward MJ, Punwani S, Proctor I, et al. Non-perforating small bowel Crohn’s disease
assessed by MRI enterography: Derivation and histopathological validation of an MR-
based activity index. Eur J Radiol 2012;81:2080–8.
6. Buisson A, Joubert A, Montoriol PF, et al. Diffusion-weighted magnetic resonance imaging
for detecting and assessing ileal inflammation in Crohn’s disease. Aliment Pharmacol Ther
2013;37:537–45.
7. Rimola J, Ordás I, Rodriguez S, et al. Magnetic resonance imaging for evaluation of Crohn’s
disease: Validation of parameters of severity and quantitative index of activity. Inflamm
Bowel Dis 2011;17:1759–68.
8. Hordonneau C, Buisson A, Scanzi J, et al. Diffusion-weighted magnetic resonance
imaging in ileocolonic Crohn’s disease: validation of quantitative index of activity. Am J
Gastroenterol 2014;109:89–98.
9. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading crohn disease activity with MRI:
Interobserver variability of MRI features, MRI scoring of severity, and correlation with
crohn disease endoscopic index of severity. Am J Roentgenol 2013;201:1220–8.
10. Ziech MLW, Bipat S, Roelofs JJTH, et al. Retrospective comparison of magnetic
resonance imaging features and histopathology in Crohn’s disease patients. Eur J Radiol
2011;80:e299–305.
11. Tielbeek JAW, Vos FM, Stoker J. A computer-assisted model for detection of MRI signs of
Crohn’s disease activity: Future or fiction? Abdom Imaging 2012;37:967–73.
12. Naziroglu RE, Puylaert CAJ, Tielbeek JAW, et al. Semi-automatic bowel wall thickness
measurements on MR enterography in patients with Crohn’s disease. Br J Radiol
2017;:20160654.
13. Li Z, Tielbeek JAW, Caan MWA, et al. Expiration-Phase Template-Based Motion Correction
of Free-Breathing Abdominal Dynamic Contrast Enhanced MRI. IEEE Trans Biomed Eng
2015;62:1215–25.
14. Dave M, Loftus E V. Mucosal healing in inflammatory bowel disease-a true paradigm of
success? Gastroenterol Hepatol (N Y) 2012;8:29–38.
15. Buisson A, Hordonneau C, Goutte M, et al. Diffusion-weighted magnetic resonance
enterocolonography in predicting remission after anti-TNF induction therapy in Crohn’s
disease. Dig Liver Dis 2016;48:260–6.
16. Stoppino LP, Della Valle N, Rizzi S, et al. Magnetic resonance enterography changes after
antibody to tumor necrosis factor (anti-TNF) alpha therapy in Crohn’s disease: Correlation
with SES-CD and clinical-biological markers. BMC Med Imaging 2016;16.
Proefschrift2018-new2.indb 172 13/9/18 10:06
General discussion
173
17. Park SH. DWI at MR Enterography for Evaluating Bowel Inflammation in Crohn Disease.
AJR Am J Roentgenol 2016;:1–9.
18. Kim K-J, Lee Y, Park SH, et al. Diffusion-weighted MR enterography for evaluating Crohn’s
disease: how does it add diagnostically to conventional MR enterography? Inflamm Bowel
Dis 2015;21:101–9.
19. Seo N, Park SH, Kim K-J, et al. MR Enterography for the Evaluation of Small-Bowel
Inflammation in Crohn Disease by Using Diffusion-weighted Imaging without Intravenous
Contrast Material: A Prospective Noninferiority Study. Radiology 2015;0:150809.
20. Freiman M, Perez-Rossello JM, Callahan MJ, et al. Characterization of fast and slow
diffusion from diffusion-weighted MRI of pediatric Crohn’s disease. J Magn Reson Imaging
2013;37:156–63.
21. Menys A, Atkinson D, Odille F, et al. Quantified terminal ileal motility during MR
enterography as a potential biomarker of Crohn’s disease activity: A preliminary study.
Eur Radiol 2012;22:2494–501.
22. Tielbeek JAW, Bipat S, Boellaard TN, et al. Training readers to improve their accuracy in
grading Crohn’s disease activity on MRI. Eur Radiol 2014;:1059–67.
23. Negaard A, Mulahasanovic A, Reisaeter LA, et al. Crohn’s disease evaluated with magnetic
resonance enteroclysis: diagnostic performance of experienced and inexperienced
readers before and after training. Acta Radiol 2008;49:967–74.
Proefschrift2018-new2.indb 173 13/9/18 10:06
General discussion
173
17. Park SH. DWI at MR Enterography for Evaluating Bowel Inflammation in Crohn Disease.
AJR Am J Roentgenol 2016;:1–9.
18. Kim K-J, Lee Y, Park SH, et al. Diffusion-weighted MR enterography for evaluating Crohn’s
disease: how does it add diagnostically to conventional MR enterography? Inflamm Bowel
Dis 2015;21:101–9.
19. Seo N, Park SH, Kim K-J, et al. MR Enterography for the Evaluation of Small-Bowel
Inflammation in Crohn Disease by Using Diffusion-weighted Imaging without Intravenous
Contrast Material: A Prospective Noninferiority Study. Radiology 2015;0:150809.
20. Freiman M, Perez-Rossello JM, Callahan MJ, et al. Characterization of fast and slow
diffusion from diffusion-weighted MRI of pediatric Crohn’s disease. J Magn Reson Imaging
2013;37:156–63.
21. Menys A, Atkinson D, Odille F, et al. Quantified terminal ileal motility during MR
enterography as a potential biomarker of Crohn’s disease activity: A preliminary study.
Eur Radiol 2012;22:2494–501.
22. Tielbeek JAW, Bipat S, Boellaard TN, et al. Training readers to improve their accuracy in
grading Crohn’s disease activity on MRI. Eur Radiol 2014;:1059–67.
23. Negaard A, Mulahasanovic A, Reisaeter LA, et al. Crohn’s disease evaluated with magnetic
resonance enteroclysis: diagnostic performance of experienced and inexperienced
readers before and after training. Acta Radiol 2008;49:967–74.
Proefschrift2018-new2.indb 173 13/9/18 10:06
Proefschrift2018-new2.indb 174 13/9/18 10:06 Proefschrift2018-new2.indb 174 13/9/18 10:06
CHAPTER 9
Summary
Proefschrift2018-new2.indb 175 13/9/18 10:06
CHAPTER 9
Summary
Proefschrift2018-new2.indb 175 13/9/18 10:06
Chapter 9
176
SUMMARY
Crohn’s disease is an inflammatory bowel disease (IBD) characterized by episodes
of relapse and remission. Over the last decade, MRI has become an important
tool for the diagnosis and evaluation of Crohn’s disease, and is a promising, non-
invasive alternative to endoscopy. In this thesis, we investigate and compare
currently available MRI methods for evaluation of Crohn’s disease, while exploring
new technological advancements, such as new MRI sequences and the use of
semiautomatic measurements.
In chapter 1, a general introduction is provided with an outline of the different
chapters. Most of the studies included in this thesis (chapters 3-6) were derived
from a large prospective patient cohort included under the VIGOR++ project,
an international collaboration under the European Union’s Seventh Framework
Programme. The VIGOR++ project was aimed at developing and applying new
techniques in the field of quantitative image analysis to improve the MRI evaluation
of Crohn’s disease patients.
A number of frequently used imaging techniques for evaluation of Crohn’s are
compared in chapter 2 by performing a meta-analysis of available literature. This
study shows that computed tomography (CT) and magnetic resonance imaging
(MRI) have comparable high grading accuracies, while results for ultrasonography
(US) and scintigraphy were inconsistent and with limited availability of studies. Given
the available data, we advise the use of MRI for grading of Crohn’s disease activity,
due to the high grading accuracy, while avoiding the use of ionizing radiation.
Several MRI activity scores have been developed over the last years, which have led
to improvements in the uniformity and accuracy of Crohn’s disease evaluation using
MRI. In chapter 3, we evaluate a number of commonly used MRI scoring systems for
Crohn’s disease are compare these in terms of diagnostic and grading accuracy, and
reproducibility. The Magnetic Resonance Index of Activity (MaRIA), Clermont score,
London score and Crohn’s disease MRI index (CDMI) showed equal performance
characteristics: high sensitivity and specificity for active and severe disease, high
grading accuracy and good reproducibility. From this we conclude that the choice
for MRI score should depend on personal preference.
Proefschrift2018-new2.indb 176 13/9/18 10:06
Chapter 9
176
SUMMARY
Crohn’s disease is an inflammatory bowel disease (IBD) characterized by episodes
of relapse and remission. Over the last decade, MRI has become an important
tool for the diagnosis and evaluation of Crohn’s disease, and is a promising, non-
invasive alternative to endoscopy. In this thesis, we investigate and compare
currently available MRI methods for evaluation of Crohn’s disease, while exploring
new technological advancements, such as new MRI sequences and the use of
semiautomatic measurements.
In chapter 1, a general introduction is provided with an outline of the different
chapters. Most of the studies included in this thesis (chapters 3-6) were derived
from a large prospective patient cohort included under the VIGOR++ project,
an international collaboration under the European Union’s Seventh Framework
Programme. The VIGOR++ project was aimed at developing and applying new
techniques in the field of quantitative image analysis to improve the MRI evaluation
of Crohn’s disease patients.
A number of frequently used imaging techniques for evaluation of Crohn’s are
compared in chapter 2 by performing a meta-analysis of available literature. This
study shows that computed tomography (CT) and magnetic resonance imaging
(MRI) have comparable high grading accuracies, while results for ultrasonography
(US) and scintigraphy were inconsistent and with limited availability of studies. Given
the available data, we advise the use of MRI for grading of Crohn’s disease activity,
due to the high grading accuracy, while avoiding the use of ionizing radiation.
Several MRI activity scores have been developed over the last years, which have led
to improvements in the uniformity and accuracy of Crohn’s disease evaluation using
MRI. In chapter 3, we evaluate a number of commonly used MRI scoring systems for
Crohn’s disease are compare these in terms of diagnostic and grading accuracy, and
reproducibility. The Magnetic Resonance Index of Activity (MaRIA), Clermont score,
London score and Crohn’s disease MRI index (CDMI) showed equal performance
characteristics: high sensitivity and specificity for active and severe disease, high
grading accuracy and good reproducibility. From this we conclude that the choice
for MRI score should depend on personal preference.
Proefschrift2018-new2.indb 176 13/9/18 10:06
Summary
177
MRI activity scores have been investigated for their potential role in therapeutic
monitoring of Crohn’s disease and use in clinical trials. Despite promising results,
varying reproducibility of included MRI features is reported in the literature.
This suggests that there is ample room for improvement of these MRI scores. In
chapter 4, we have developed and validated a new MRI activity score (“the VIGOR
score”) using a combination of semiautomatic MRI features and conventional
subjective features. The VIGOR score was compared to two existing activity scores
– the MaRIA and London score – and a newly developed subjective score. The
validation was performed on a prospective cohort of 106 Crohn’s disease patients
using an endoscopic reference standard. The VIGOR score showed comparable
accuracy to existing MRI activity scores, while showing improved interobserver
agreement (intraclass correlation coefficient = 0.81 vs 0.44–0.59, respectively).
These characteristics make the VIGOR score the preferred candidate for disease
monitoring and therapy evaluation.
The use of diffusion-weighted (DW) imaging in evaluation of Crohn’s disease
inflammation has been investigated in numerous studies, although the additional
value to conventional MRI remains unclear. In chapter 5, contrast-enhanced (CE)-
MRI and DW-MRI were compared for diagnosis and grading of disease activity in
the terminal ileum. Although similar levels of diagnostic and grading accuracy were
found, higher levels of confidence were seen for CE-MRI. Additionally, discrepancies
were seen for diagnosis of penetrating complications in favor of CE-MRI. DW-MRI
could be an adequate substitute for CE-MRI in cases where contrast should be
avoided. CE-MRI should remain the first choice examination in healthy, adult, and
non-pregnant patients.
Assessment of Crohn’s disease activity has largely relied on the use of static images.
Improved fast acquisition of MR images facilitated examination of bowel motility.
In chapter 6, we found that bowel motility of the terminal ileum could predict
the presence of endoscopic disease (93% sensitivity and 67% specificity) and
histopathologic disease (92% sensitivity and 74% specificity). Motility was found
to have an inverse correlation with the severity of endoscopic and histopathologic
disease activity. These results indicate that bowel motility can be used as a potential
marker for endoscopic and histopathologic disease activity on MRI.
Proefschrift2018-new2.indb 177 13/9/18 10:06
Summary
177
MRI activity scores have been investigated for their potential role in therapeutic
monitoring of Crohn’s disease and use in clinical trials. Despite promising results,
varying reproducibility of included MRI features is reported in the literature.
This suggests that there is ample room for improvement of these MRI scores. In
chapter 4, we have developed and validated a new MRI activity score (“the VIGOR
score”) using a combination of semiautomatic MRI features and conventional
subjective features. The VIGOR score was compared to two existing activity scores
– the MaRIA and London score – and a newly developed subjective score. The
validation was performed on a prospective cohort of 106 Crohn’s disease patients
using an endoscopic reference standard. The VIGOR score showed comparable
accuracy to existing MRI activity scores, while showing improved interobserver
agreement (intraclass correlation coefficient = 0.81 vs 0.44–0.59, respectively).
These characteristics make the VIGOR score the preferred candidate for disease
monitoring and therapy evaluation.
The use of diffusion-weighted (DW) imaging in evaluation of Crohn’s disease
inflammation has been investigated in numerous studies, although the additional
value to conventional MRI remains unclear. In chapter 5, contrast-enhanced (CE)-
MRI and DW-MRI were compared for diagnosis and grading of disease activity in
the terminal ileum. Although similar levels of diagnostic and grading accuracy were
found, higher levels of confidence were seen for CE-MRI. Additionally, discrepancies
were seen for diagnosis of penetrating complications in favor of CE-MRI. DW-MRI
could be an adequate substitute for CE-MRI in cases where contrast should be
avoided. CE-MRI should remain the first choice examination in healthy, adult, and
non-pregnant patients.
Assessment of Crohn’s disease activity has largely relied on the use of static images.
Improved fast acquisition of MR images facilitated examination of bowel motility.
In chapter 6, we found that bowel motility of the terminal ileum could predict
the presence of endoscopic disease (93% sensitivity and 67% specificity) and
histopathologic disease (92% sensitivity and 74% specificity). Motility was found
to have an inverse correlation with the severity of endoscopic and histopathologic
disease activity. These results indicate that bowel motility can be used as a potential
marker for endoscopic and histopathologic disease activity on MRI.
Proefschrift2018-new2.indb 177 13/9/18 10:06
Chapter 9
178
Previous research has shown that inexperienced readers can be trained in grading
Crohn’s disease on MRI using a standardized scoring system. In chapter 7, we
evaluated the long-term performance of readers who previously participated in MRI
reader training. Results showed an overall grading accuracy of 73% at the follow-up
evaluation, which was similar to the 72% accuracy at the initial reader training. A
decrease in overstaging was seen (19% to 13%), while an increase in understaging
occurred (9% to 14%). Based on the overall grading accuracy, these results indicate
that readers have consistent long-term performance for grading Crohn’s disease
activity after reader training.
In recent years, MRI has been the most actively investigated radiologic technique
for evaluation of Crohn’s disease. In this thesis, we have put forward a number of
new methods to improve MRI evaluation of Crohn’s disease. To continue this effort,
new studies should explore the use of new MRI sequences and semiautomatic
measurements in other areas, such as therapeutic monitoring and the characterization
of Crohn’s disease lesions as inflammatory or fibrotic. Overall, we should continue
expanding the use of automation, not to replace radiologists, but to relieve them
from arduous measurements, thereby allowing more attentive care for the patient.
Proefschrift2018-new2.indb 178 13/9/18 10:06
Chapter 9
178
Previous research has shown that inexperienced readers can be trained in grading
Crohn’s disease on MRI using a standardized scoring system. In chapter 7, we
evaluated the long-term performance of readers who previously participated in MRI
reader training. Results showed an overall grading accuracy of 73% at the follow-up
evaluation, which was similar to the 72% accuracy at the initial reader training. A
decrease in overstaging was seen (19% to 13%), while an increase in understaging
occurred (9% to 14%). Based on the overall grading accuracy, these results indicate
that readers have consistent long-term performance for grading Crohn’s disease
activity after reader training.
In recent years, MRI has been the most actively investigated radiologic technique
for evaluation of Crohn’s disease. In this thesis, we have put forward a number of
new methods to improve MRI evaluation of Crohn’s disease. To continue this effort,
new studies should explore the use of new MRI sequences and semiautomatic
measurements in other areas, such as therapeutic monitoring and the characterization
of Crohn’s disease lesions as inflammatory or fibrotic. Overall, we should continue
expanding the use of automation, not to replace radiologists, but to relieve them
from arduous measurements, thereby allowing more attentive care for the patient.
Proefschrift2018-new2.indb 178 13/9/18 10:06
Proefschrift2018-new2.indb 179 13/9/18 10:06 Proefschrift2018-new2.indb 179 13/9/18 10:06
Proefschrift2018-new2.indb 180 13/9/18 10:06 Proefschrift2018-new2.indb 180 13/9/18 10:06
CHAPTER 10
Nederlandse samenvatting
Proefschrift2018-new2.indb 181 13/9/18 10:06
CHAPTER 10
Nederlandse samenvatting
Proefschrift2018-new2.indb 181 13/9/18 10:06
Chapter 10
182
SAMENVATTING
De ziekte van Crohn is een inflammatoire darmziekte, die wordt gekenmerkt door
wisselend episodes van ziekteactiviteit en remissie. In de laatste tien jaar is MRI een
belangrijk hulpmiddel geworden voor de diagnose en beoordeling van de ziekte
van Crohn en lijkt een veelbelovend, non-invasief alternatief voor endoscopie. In dit
proefschrift onderzoeken en vergelijken we verschillende bestaande technieken voor
de beoordeling van de ziekte van Crohn met MRI, en technologische ontwikkelingen,
zoals nieuwe MRI sequenties en het gebruik van semiautomatische metingen.
In hoofdstuk 1 van dit proefschrift wordt een algemene introductie gegeven en een
overzicht van de verschillende hoofdstukken.
Een aantal veel gebruikte beeldvormingstechnieken worden in hoofdstuk 2 met
elkaar vergeleken door middel van een meta-analyse van de beschikbare literatuur.
Deze studie liet zien dat CT en MRI vergelijkbare, hoge accuratesse hebben voor het
graderen van de ziekte van Crohn (75–87%), terwijl de resultaten voor echografie
en scintigrafie inconsistent waren en beperkt werden door de weinig beschikbare
studies. Het gebruik van MRI heeft de voorkeur voor het graderen van de ziekte van
Crohn, vanwege de hoge accuratesse en daarnaast omdat het geen gebruik maakt
van schadelijke röntgenstraling.
Hoofdstukken 3 t/m 6 komen voort uit het VIGOR++ project, een internationale
samenwerking onder het Seventh Framework Programme van de Europese Unie.
Het VIGOR++ project heeft zich gericht op de ontwikkeling en toepassing van
nieuwe beeldbewerkingstechnieken voor MRI om daarmee de beoordeling van de
ziekte van Crohn te kunnen verbeteren.
In de afgelopen jaren zijn er meerdere MRI scores ontwikkeld voor het graderen van
de ziekte van Crohn, welke verbeteringen hebben opgeleverd in de uniformiteit en
accuratesse van de beoordeling. In hoofdstuk 3 worden een aantal vaak gebruikte
MRI scores vergeleken, waarmee men de ziekte van Crohn kan graderen. We
vergeleken de scores op verschillende kenmerken, zoals de diagnose en gradering
van ziekteactiviteit, en de reproduceerbaarheid van de score door verschillende
radiologen. De Magnetic Resonance Index of Activity (MaRIA), Clermont score,
London score en Crohn’s disease MRI index (CDMI) werden in deze studie vergeleken
Proefschrift2018-new2.indb 182 13/9/18 10:06
Chapter 10
182
SAMENVATTING
De ziekte van Crohn is een inflammatoire darmziekte, die wordt gekenmerkt door
wisselend episodes van ziekteactiviteit en remissie. In de laatste tien jaar is MRI een
belangrijk hulpmiddel geworden voor de diagnose en beoordeling van de ziekte
van Crohn en lijkt een veelbelovend, non-invasief alternatief voor endoscopie. In dit
proefschrift onderzoeken en vergelijken we verschillende bestaande technieken voor
de beoordeling van de ziekte van Crohn met MRI, en technologische ontwikkelingen,
zoals nieuwe MRI sequenties en het gebruik van semiautomatische metingen.
In hoofdstuk 1 van dit proefschrift wordt een algemene introductie gegeven en een
overzicht van de verschillende hoofdstukken.
Een aantal veel gebruikte beeldvormingstechnieken worden in hoofdstuk 2 met
elkaar vergeleken door middel van een meta-analyse van de beschikbare literatuur.
Deze studie liet zien dat CT en MRI vergelijkbare, hoge accuratesse hebben voor het
graderen van de ziekte van Crohn (75–87%), terwijl de resultaten voor echografie
en scintigrafie inconsistent waren en beperkt werden door de weinig beschikbare
studies. Het gebruik van MRI heeft de voorkeur voor het graderen van de ziekte van
Crohn, vanwege de hoge accuratesse en daarnaast omdat het geen gebruik maakt
van schadelijke röntgenstraling.
Hoofdstukken 3 t/m 6 komen voort uit het VIGOR++ project, een internationale
samenwerking onder het Seventh Framework Programme van de Europese Unie.
Het VIGOR++ project heeft zich gericht op de ontwikkeling en toepassing van
nieuwe beeldbewerkingstechnieken voor MRI om daarmee de beoordeling van de
ziekte van Crohn te kunnen verbeteren.
In de afgelopen jaren zijn er meerdere MRI scores ontwikkeld voor het graderen van
de ziekte van Crohn, welke verbeteringen hebben opgeleverd in de uniformiteit en
accuratesse van de beoordeling. In hoofdstuk 3 worden een aantal vaak gebruikte
MRI scores vergeleken, waarmee men de ziekte van Crohn kan graderen. We
vergeleken de scores op verschillende kenmerken, zoals de diagnose en gradering
van ziekteactiviteit, en de reproduceerbaarheid van de score door verschillende
radiologen. De Magnetic Resonance Index of Activity (MaRIA), Clermont score,
London score en Crohn’s disease MRI index (CDMI) werden in deze studie vergeleken
Proefschrift2018-new2.indb 182 13/9/18 10:06
Nederlandse samenvatting
183
en lieten gelijke diagnostische accuratesse zien voor actieve en ernstige ziekte en
een hoge accuratesse voor het graderen de ziekte van Crohn met daarnaast een
goede reproduceerbaarheid. Hieruit volgt dat de keuze voor een van de scores mag
berusten op persoonlijke voorkeur.
MRI scores zijn tevens onderzocht voor hun potentiele rol in het bepalen van
het effect van behandeling en het gebruik in therapeutische trials. Ondanks
veelbelovende resultaten, blijft er wisselende reproduceerbaarheid gezien worden
voor de MRI kenmerken in deze scores. Dit geeft aan dat er nog ruimte is voor
mogelijke verbetering van deze MRI scores. In hoofdstuk 4 werd een nieuwe MRI
score (“de VIGOR score”) ontwikkeld en gevalideerd, die gebruik maakte van
zowel conventionele subjectieve MRI kenmerken, als van nieuwe semiautomatische
metingen. De VIGOR score werd vergeleken met twee bestaande scores – de MaRIA
en de London score – en een nieuw ontwikkelde subjectieve score. Hiervoor werd
een prospectief cohort gebruikt van 106 patiënten met de ziekte van Crohn en
gebruik makende van een endoscopische referentiestandaard. De VIGOR score liet
vergelijkbare accuratesse zien met andere MRI scores, maar een sterk verbeterde
reproduceerbaarheid (intraclass correlatie coefficient = 0.81 vs. 0.44–0.59). Deze
kenmerken maken de VIGOR score goed toepasbaar voor het vervolgen van
ziekteactiviteit en het bepalen het effect van behandeling.
Het gebruik van diffusion-gewogen beelden (DWI) voor de beoordeling van
activiteit bij de ziekte van Crohn is in vele studies onderzocht. Desondanks is de
toegevoegde waarde van DWI tot nu toe onduidelijk. In hoofdstuk 5 vergeleken
we contrast-enhanced (CE)-MRI en DW-MRI op de accuratesse voor de diagnose
en gradering van ziekteactiviteit in het terminale ileum. Alhoewel er vergelijkbare
accuratesse werd gevonden, rapporteerden de radiologen meer zekerheid te
hebben bij de beoordeling met CE-MRI. Daarnaast werden er verschillen gevonden
in de diagnose van fistels en abcessen, in het voordeel van CE-MRI. Gezien deze
resultaten kan DW-MRI een goed alternatief zijn voor CE-MRI in gevallen waar
intraveneus contrast gecontra-indiceerd is. In gezonde, niet-zwangere volwassenen
moet CE-MRI het MRI onderzoek van eerste keus blijven.
De beoordeling van de ziekte van Crohn berust grotendeels op het gebruik van
statische beelden. Met nieuwe technieken waarmee sneller beelden kunnen worden
gemaakt, is het mogelijk om de darmmotiliteit te beoordelen op bewegende
Proefschrift2018-new2.indb 183 13/9/18 10:06
Nederlandse samenvatting
183
en lieten gelijke diagnostische accuratesse zien voor actieve en ernstige ziekte en
een hoge accuratesse voor het graderen de ziekte van Crohn met daarnaast een
goede reproduceerbaarheid. Hieruit volgt dat de keuze voor een van de scores mag
berusten op persoonlijke voorkeur.
MRI scores zijn tevens onderzocht voor hun potentiele rol in het bepalen van
het effect van behandeling en het gebruik in therapeutische trials. Ondanks
veelbelovende resultaten, blijft er wisselende reproduceerbaarheid gezien worden
voor de MRI kenmerken in deze scores. Dit geeft aan dat er nog ruimte is voor
mogelijke verbetering van deze MRI scores. In hoofdstuk 4 werd een nieuwe MRI
score (“de VIGOR score”) ontwikkeld en gevalideerd, die gebruik maakte van
zowel conventionele subjectieve MRI kenmerken, als van nieuwe semiautomatische
metingen. De VIGOR score werd vergeleken met twee bestaande scores – de MaRIA
en de London score – en een nieuw ontwikkelde subjectieve score. Hiervoor werd
een prospectief cohort gebruikt van 106 patiënten met de ziekte van Crohn en
gebruik makende van een endoscopische referentiestandaard. De VIGOR score liet
vergelijkbare accuratesse zien met andere MRI scores, maar een sterk verbeterde
reproduceerbaarheid (intraclass correlatie coefficient = 0.81 vs. 0.44–0.59). Deze
kenmerken maken de VIGOR score goed toepasbaar voor het vervolgen van
ziekteactiviteit en het bepalen het effect van behandeling.
Het gebruik van diffusion-gewogen beelden (DWI) voor de beoordeling van
activiteit bij de ziekte van Crohn is in vele studies onderzocht. Desondanks is de
toegevoegde waarde van DWI tot nu toe onduidelijk. In hoofdstuk 5 vergeleken
we contrast-enhanced (CE)-MRI en DW-MRI op de accuratesse voor de diagnose
en gradering van ziekteactiviteit in het terminale ileum. Alhoewel er vergelijkbare
accuratesse werd gevonden, rapporteerden de radiologen meer zekerheid te
hebben bij de beoordeling met CE-MRI. Daarnaast werden er verschillen gevonden
in de diagnose van fistels en abcessen, in het voordeel van CE-MRI. Gezien deze
resultaten kan DW-MRI een goed alternatief zijn voor CE-MRI in gevallen waar
intraveneus contrast gecontra-indiceerd is. In gezonde, niet-zwangere volwassenen
moet CE-MRI het MRI onderzoek van eerste keus blijven.
De beoordeling van de ziekte van Crohn berust grotendeels op het gebruik van
statische beelden. Met nieuwe technieken waarmee sneller beelden kunnen worden
gemaakt, is het mogelijk om de darmmotiliteit te beoordelen op bewegende
Proefschrift2018-new2.indb 183 13/9/18 10:06
Chapter 10
184
MRI beelden. In hoofdstuk 6 vonden we dat de darmmotiliteit van het terminal
ileum een voorspeller is voor de aanwezigheid van ziekteactiviteit, gezien door
endoscopie (93% sensitiviteit en 67% specificiteit) en op histopathologische
coupes (92% sensitiviteit en 74% specificiteit). Daarnaast was de mate van
darmmotiliteit omgekeerd evenredig gecorreleerd met de ernst van endoscopische
en histopathologische ziekteactiviteit. Deze resultaten geven aan dat darmmotiliteit
op MRI gebruikt kan worden als een potentiële marker voor endoscopische en
histopathologische ziekteactiviteit.
Eerder onderzoek heeft laten zien dat onervaren radiologen (in opleiding) getraind
kunnen worden in het graderen van activiteit op MRI bij de ziekte van Crohn met
het gebruik van een gestandaardiseerde score. In hoofdstuk 7 bekeken we de
lange-termijn prestaties van de deelnemers van deze training. Bij deze test werd
een accuratesse voor het graderen van ziekteactiviteit gezien van 73%, vrijwel gelijk
aan de 72% bij de initiële training. Daarnaast werd er een afname in overstagiering
(19% naar 13%) gezien en een toename in onderstagiering (9% naar 14%). Hieruit kan
geconcludeerd worden MRI training voor de ziekte van Crohn een consistent lange-
termijn effect heeft op de prestaties van de deelnemers.
In de afgelopen jaren is MRI de meest onderzochte radiologische techniek geweest
voor de beoordeling van de ziekte van Crohn. In dit proefschrift hebben we een
aantal nieuwe methoden voorgesteld om deze beoordeling te kunnen verbeteren.
Om deze vooruitgang voort te blijven zetten zijn nieuwe studies nodig die het
gebruik van nieuwe MRI sequenties en semiautomatische metingen onderzoeken
op andere aspecten van de ziekte van Crohn, zoals het bepaling van het effect van
therapie en het onderscheiden van actieve ontsteking en fibrose. Daarnaast moeten
we het gebruik van automatisering verder blijven ontwikkelen, niet om radiologen
te vervangen, maar om hen tijdrovende metingen te besparen, en zo meer tijd en
aandacht te gunnen voor de patiënt.
Proefschrift2018-new2.indb 184 13/9/18 10:06
Chapter 10
184
MRI beelden. In hoofdstuk 6 vonden we dat de darmmotiliteit van het terminal
ileum een voorspeller is voor de aanwezigheid van ziekteactiviteit, gezien door
endoscopie (93% sensitiviteit en 67% specificiteit) en op histopathologische
coupes (92% sensitiviteit en 74% specificiteit). Daarnaast was de mate van
darmmotiliteit omgekeerd evenredig gecorreleerd met de ernst van endoscopische
en histopathologische ziekteactiviteit. Deze resultaten geven aan dat darmmotiliteit
op MRI gebruikt kan worden als een potentiële marker voor endoscopische en
histopathologische ziekteactiviteit.
Eerder onderzoek heeft laten zien dat onervaren radiologen (in opleiding) getraind
kunnen worden in het graderen van activiteit op MRI bij de ziekte van Crohn met
het gebruik van een gestandaardiseerde score. In hoofdstuk 7 bekeken we de
lange-termijn prestaties van de deelnemers van deze training. Bij deze test werd
een accuratesse voor het graderen van ziekteactiviteit gezien van 73%, vrijwel gelijk
aan de 72% bij de initiële training. Daarnaast werd er een afname in overstagiering
(19% naar 13%) gezien en een toename in onderstagiering (9% naar 14%). Hieruit kan
geconcludeerd worden MRI training voor de ziekte van Crohn een consistent lange-
termijn effect heeft op de prestaties van de deelnemers.
In de afgelopen jaren is MRI de meest onderzochte radiologische techniek geweest
voor de beoordeling van de ziekte van Crohn. In dit proefschrift hebben we een
aantal nieuwe methoden voorgesteld om deze beoordeling te kunnen verbeteren.
Om deze vooruitgang voort te blijven zetten zijn nieuwe studies nodig die het
gebruik van nieuwe MRI sequenties en semiautomatische metingen onderzoeken
op andere aspecten van de ziekte van Crohn, zoals het bepaling van het effect van
therapie en het onderscheiden van actieve ontsteking en fibrose. Daarnaast moeten
we het gebruik van automatisering verder blijven ontwikkelen, niet om radiologen
te vervangen, maar om hen tijdrovende metingen te besparen, en zo meer tijd en
aandacht te gunnen voor de patiënt.
Proefschrift2018-new2.indb 184 13/9/18 10:06
Proefschrift2018-new2.indb 185 13/9/18 10:06 Proefschrift2018-new2.indb 185 13/9/18 10:06
Proefschrift2018-new2.indb 186 13/9/18 10:06 Proefschrift2018-new2.indb 186 13/9/18 10:06
APPENDIX
List of contributing authorsPortfolioList of publicationsCurriculum vitaeDankwoord
Proefschrift2018-new2.indb 187 13/9/18 10:07
APPENDIX
List of contributing authorsPortfolioList of publicationsCurriculum vitaeDankwoord
Proefschrift2018-new2.indb 187 13/9/18 10:07
Appendix
188
LIST OF CONTRIBUTING AUTHORSDavid AtkinsonCentre for Medical Imaging, University College London Hospitals, London, United Kingdom
Shandra BipatDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Thierry N. BoellaardDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Lodewijk A.A. BrosensDepartment of Pathology, University Medical Center Utrecht, Utrecht, the Netherlands
Joachim M. BuhmannDepartment of Computer Science, ETH Zurich, Zurich, Switzerland
Alastair ForbesNorwich Medical School, University of East Anglia, Norwich, United Kingdom
Thomas J. FuchsDepartment of Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, USA
Haralambos HatzakisCEO of Biotronics Ltd, London, United Kingdom
Karin HorsthuisDepartment of Radiology and Nuclear Medicine, VUMC, Amsterdam, the Netherlands
Zhang LiDepartment of Imaging Physics, Technical University Delft, Delft, the Netherlands
Jesica C. MakanyangaCentre for Medical Imaging, University College London Hospitals, London, United Kingdom
Banafsche MearadjiDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Alex MenysCentre for Medical Imaging, University College London Hospitals, London, United Kingdom
Robiel E. NazirogluDepartment of Imaging Physics, Technical University Delft, Delft, the Netherlands
Proefschrift2018-new2.indb 188 13/9/18 10:07
Appendix
188
LIST OF CONTRIBUTING AUTHORSDavid AtkinsonCentre for Medical Imaging, University College London Hospitals, London, United Kingdom
Shandra BipatDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Thierry N. BoellaardDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Lodewijk A.A. BrosensDepartment of Pathology, University Medical Center Utrecht, Utrecht, the Netherlands
Joachim M. BuhmannDepartment of Computer Science, ETH Zurich, Zurich, Switzerland
Alastair ForbesNorwich Medical School, University of East Anglia, Norwich, United Kingdom
Thomas J. FuchsDepartment of Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, USA
Haralambos HatzakisCEO of Biotronics Ltd, London, United Kingdom
Karin HorsthuisDepartment of Radiology and Nuclear Medicine, VUMC, Amsterdam, the Netherlands
Zhang LiDepartment of Imaging Physics, Technical University Delft, Delft, the Netherlands
Jesica C. MakanyangaCentre for Medical Imaging, University College London Hospitals, London, United Kingdom
Banafsche MearadjiDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Alex MenysCentre for Medical Imaging, University College London Hospitals, London, United Kingdom
Robiel E. NazirogluDepartment of Imaging Physics, Technical University Delft, Delft, the Netherlands
Proefschrift2018-new2.indb 188 13/9/18 10:07
189
List of contributing authors
C. Yung NioDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Douglas A. PendséCentre for Medical Imaging, University College London Hospitals, London, United Kingdom
Cyriel Y. PonsioenDepartment of Gastroenterology, Academic Medical Center, Amsterdam, the Netherlands
Manuel Rodriguez-JustoDepartment of Pathology, University College London Hospitals, London, United Kingdom
Peter J. SchüfflerDepartment of Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, USA
Jaap StokerDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Stuart A. TaylorCentre for Medical Imaging, University College London Hospitals, London, United Kingdom
Jeroen A.W. TielbeekDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Charlotte J. Tutein NoltheniusDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Lucas J. van VlietDepartment of Imaging Physics, Technical University Delft, Delft, the Netherlands
Frans M. VosDepartment of Imaging Physics, Technical University Delft, Delft, the Netherlands
Proefschrift2018-new2.indb 189 13/9/18 10:07
189
List of contributing authors
C. Yung NioDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Douglas A. PendséCentre for Medical Imaging, University College London Hospitals, London, United Kingdom
Cyriel Y. PonsioenDepartment of Gastroenterology, Academic Medical Center, Amsterdam, the Netherlands
Manuel Rodriguez-JustoDepartment of Pathology, University College London Hospitals, London, United Kingdom
Peter J. SchüfflerDepartment of Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, USA
Jaap StokerDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Stuart A. TaylorCentre for Medical Imaging, University College London Hospitals, London, United Kingdom
Jeroen A.W. TielbeekDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Charlotte J. Tutein NoltheniusDepartment of Radiology and Nuclear Medicine, Academic Medical Center, Amsterdam, the
Netherlands
Lucas J. van VlietDepartment of Imaging Physics, Technical University Delft, Delft, the Netherlands
Frans M. VosDepartment of Imaging Physics, Technical University Delft, Delft, the Netherlands
Proefschrift2018-new2.indb 189 13/9/18 10:07
Appendix
190
PORTFOLIO
Name PhD student: Carl A.J. Puylaert
PhD period: November 2013 – November 2018
Supervisors: Prof. dr. J. Stoker, Prof. dr. ir. L.J. van Vliet
Co-supervisors: Dr. Frans M. Vos, Dr. C.Y. Ponsioen
General courses
Year Workload
(ECTS)
Good Clinical Practice – BROK 2014 1.0
Practical Biostatistics (e-learning) 2014 1.1
Oral Presentation in English 2014 0.8
Clinical Epidemiology 3: Evaluation of Medical Tests 2015 0.9
Seminars and workshops
Young Initiative on Crohn's and colitis, Utrecht 2013 0.2
Young Initiative on Crohn's and colitis, Utrecht 2014 0.2
VIGOR++ workshop, UCL London 2014 0.2
Presentations
European Congress of Radiology, Vienna 2014 0.5
European Society of Gastrointestinal Abdominal Radiology,
Salzburg
2014 0.5
Radiologic Society of North America, Chicago 2014 0.5
Radiologendagen, Den Bosch 2014 0.5
Young Initiative on Crohn's and colitis, Amsterdam 2015 0.5
IBD Lunch and Learn, Amsterdam 2016 0.5
European Congress of Radiology, Vienna 2016 0.5
IBD Lunch and Learn, Amsterdam 2017 0.5
European Congress of Radiology, Vienna 2017 0.5
European Society of Gastrointestinal Abdominal Radiology,
Athens
2017 0.5
Conferences
European Congress of Radiology, Vienna 2014 1.0
European Society of Gastrointestinal Abdominal Radiology,
Salzburg
2014 1.0
Proefschrift2018-new2.indb 190 13/9/18 10:07
Appendix
190
PORTFOLIO
Name PhD student: Carl A.J. Puylaert
PhD period: November 2013 – November 2018
Supervisors: Prof. dr. J. Stoker, Prof. dr. ir. L.J. van Vliet
Co-supervisors: Dr. Frans M. Vos, Dr. C.Y. Ponsioen
General courses
Year Workload
(ECTS)
Good Clinical Practice – BROK 2014 1.0
Practical Biostatistics (e-learning) 2014 1.1
Oral Presentation in English 2014 0.8
Clinical Epidemiology 3: Evaluation of Medical Tests 2015 0.9
Seminars and workshops
Young Initiative on Crohn's and colitis, Utrecht 2013 0.2
Young Initiative on Crohn's and colitis, Utrecht 2014 0.2
VIGOR++ workshop, UCL London 2014 0.2
Presentations
European Congress of Radiology, Vienna 2014 0.5
European Society of Gastrointestinal Abdominal Radiology,
Salzburg
2014 0.5
Radiologic Society of North America, Chicago 2014 0.5
Radiologendagen, Den Bosch 2014 0.5
Young Initiative on Crohn's and colitis, Amsterdam 2015 0.5
IBD Lunch and Learn, Amsterdam 2016 0.5
European Congress of Radiology, Vienna 2016 0.5
IBD Lunch and Learn, Amsterdam 2017 0.5
European Congress of Radiology, Vienna 2017 0.5
European Society of Gastrointestinal Abdominal Radiology,
Athens
2017 0.5
Conferences
European Congress of Radiology, Vienna 2014 1.0
European Society of Gastrointestinal Abdominal Radiology,
Salzburg
2014 1.0
Proefschrift2018-new2.indb 190 13/9/18 10:07
191
Portfolio
IBD Today & Tomorrow, Amsterdam 2014 0.5
Radiologic Society of North America, Chicago 2014 1.0
Radiologendagen, Den Bosch 2014 0.5
European Congress of Radiology, Vienna 2016 1.0
European Congress of Radiology, Vienna 2017 0.5
European Society of Gastrointestinal Abdominal Radiology,
Athens
2017 0.5
Teaching
MRE in Crohn's disease; grading of disease activity 2015 6.0
MRE in Crohn's disease; cases and research seminar 2015 6.0
Scientific internship of Michael Bekendam, medical student 2015 2.0
Other
ESGAR European guidelines for MR enterography 2015 2.0
NVVR guideline for MR enterography 2014 2.0
MRI Crohn's disease on the Radiology Assistant 2015 3.0
Proefschrift2018-new2.indb 191 13/9/18 10:07
191
Portfolio
IBD Today & Tomorrow, Amsterdam 2014 0.5
Radiologic Society of North America, Chicago 2014 1.0
Radiologendagen, Den Bosch 2014 0.5
European Congress of Radiology, Vienna 2016 1.0
European Congress of Radiology, Vienna 2017 0.5
European Society of Gastrointestinal Abdominal Radiology,
Athens
2017 0.5
Teaching
MRE in Crohn's disease; grading of disease activity 2015 6.0
MRE in Crohn's disease; cases and research seminar 2015 6.0
Scientific internship of Michael Bekendam, medical student 2015 2.0
Other
ESGAR European guidelines for MR enterography 2015 2.0
NVVR guideline for MR enterography 2014 2.0
MRI Crohn's disease on the Radiology Assistant 2015 3.0
Proefschrift2018-new2.indb 191 13/9/18 10:07
Appendix
192
LIST OF PUBLICATIONS
Puylaert CAJ, Tutein Nolthenius CJ, Tielbeek JAW, Makanyanga JC, Rodriguez-
Justo M, Brosens LAA, Nio CY, Pendsé DA, Ponsioen CY, Vos FM, Taylor SA, Stoker J.
Comparison of MRI activity scoring systems and features for the terminal ileum in
Crohn’s disease patients.
AJR Am J Roentgenol – Accepted on July 30th 2018.
Menys A, Puylaert CAJ, Tutein Nolthenius CJ, Plumb AA, Makanyanga JC, Pendsé
DA, Ponsioen CY, Vos FM, Stoker J, Taylor SA.
Terminal ileal motility during MR enterography in Crohn’s disease: A Prospective
Multi-Institution Study.
Radiology. 2018 Aug 21.
Puylaert CAJ, Tielbeek JAW, Schüffler PJ, Nio CY, Horsthuis K, Mearadji B, Ponsioen
CY, Vos FM, Stoker J.
Comparison of contrast-enhanced and diffusion-weighted MRI in assessment of the
terminal ileum in Crohn’s disease patients.
Abdom Radiol (NY). 2018 Aug 14.
Gollifer RM, Menys A, Makanyanga J, Puylaert CA, Vos FM, Stoker J, Atkinson D,
Taylor SA.
Relationship between MRI quantified small bowel motility and abdominal symptoms
in Crohn’s disease patients-a validation study.
Br J Radiol. 2018 Jun 19:20170914.
Puylaert CAJ, Schüffler PJ, Naziroglu RE, Tielbeek JAW, Li Z, Makanyanga JC,
Tutein Nolthenius CJ, Nio CY, Pendsé DA, Menys A, Ponsioen CY, Atkinson D, Forbes
A, Buhmann JM, Fuchs TJ, Hatzakis H, van Vliet LJ, Stoker J, Taylor SA, Vos FM.
Semiautomatic Assessment of the Terminal Ileum and Colon in Patients with Crohn
Disease Using MRI (the VIGOR++ Project).
Acad Radiol. 2018 Aug;25(8):1038-1045.
Samaan MA, Puylaert CAJ, Levesque BG, Zou GY, Stitt L, Taylor SA, Shackelton
LM, Vandervoort MK, Khanna R, Santillan C, Rimola J, Hindryckx P, Nio CY,
Sandborn WJ, D’Haens G, Feagan BG, Jairath V, Stoker J. The development of a
magnetic resonance imaging index for fistulising Crohn’s disease.
Aliment Pharmacol Ther. 2017 Sep;46(5):516-528.
Proefschrift2018-new2.indb 192 13/9/18 10:07
Appendix
192
LIST OF PUBLICATIONS
Puylaert CAJ, Tutein Nolthenius CJ, Tielbeek JAW, Makanyanga JC, Rodriguez-
Justo M, Brosens LAA, Nio CY, Pendsé DA, Ponsioen CY, Vos FM, Taylor SA, Stoker J.
Comparison of MRI activity scoring systems and features for the terminal ileum in
Crohn’s disease patients.
AJR Am J Roentgenol – Accepted on July 30th 2018.
Menys A, Puylaert CAJ, Tutein Nolthenius CJ, Plumb AA, Makanyanga JC, Pendsé
DA, Ponsioen CY, Vos FM, Stoker J, Taylor SA.
Terminal ileal motility during MR enterography in Crohn’s disease: A Prospective
Multi-Institution Study.
Radiology. 2018 Aug 21.
Puylaert CAJ, Tielbeek JAW, Schüffler PJ, Nio CY, Horsthuis K, Mearadji B, Ponsioen
CY, Vos FM, Stoker J.
Comparison of contrast-enhanced and diffusion-weighted MRI in assessment of the
terminal ileum in Crohn’s disease patients.
Abdom Radiol (NY). 2018 Aug 14.
Gollifer RM, Menys A, Makanyanga J, Puylaert CA, Vos FM, Stoker J, Atkinson D,
Taylor SA.
Relationship between MRI quantified small bowel motility and abdominal symptoms
in Crohn’s disease patients-a validation study.
Br J Radiol. 2018 Jun 19:20170914.
Puylaert CAJ, Schüffler PJ, Naziroglu RE, Tielbeek JAW, Li Z, Makanyanga JC,
Tutein Nolthenius CJ, Nio CY, Pendsé DA, Menys A, Ponsioen CY, Atkinson D, Forbes
A, Buhmann JM, Fuchs TJ, Hatzakis H, van Vliet LJ, Stoker J, Taylor SA, Vos FM.
Semiautomatic Assessment of the Terminal Ileum and Colon in Patients with Crohn
Disease Using MRI (the VIGOR++ Project).
Acad Radiol. 2018 Aug;25(8):1038-1045.
Samaan MA, Puylaert CAJ, Levesque BG, Zou GY, Stitt L, Taylor SA, Shackelton
LM, Vandervoort MK, Khanna R, Santillan C, Rimola J, Hindryckx P, Nio CY,
Sandborn WJ, D’Haens G, Feagan BG, Jairath V, Stoker J. The development of a
magnetic resonance imaging index for fistulising Crohn’s disease.
Aliment Pharmacol Ther. 2017 Sep;46(5):516-528.
Proefschrift2018-new2.indb 192 13/9/18 10:07
193
List of publications
Naziroglu RE, Puylaert CAJ, Tielbeek JAW, Makanyanga J, Menys A, Ponsioen CY,
Hatzakis H, Taylor SA, Stoker J, van Vliet LJ, Vos FM.
Semi-automatic bowel wall thickness measurements on MR enterography in patients
with Crohn’s disease.
Br J Radiol. 2017 Jun;90(1074):20160654.
Bekendam MIJ, Puylaert CAJ, Phoa SKSS, Nio CY, Stoker J.
Shortened oral contrast preparation for improved small bowel distension at MR
enterography.
Abdom Radiol (NY). 2017 Sep;42(9):2225-2232.
Taylor SA, Avni F, Cronin CG, Hoeffel C, Kim SH, Laghi A, Napolitano M, Petit
P, Rimola J, Tolan DJ, Torkzad MR, Zappa M, Bhatnagar G, Puylaert CAJ, Stoker J.
The first joint ESGAR/ ESPR consensus statement on the technical performance of
cross-sectional small bowel and colonic imaging.
Eur Radiol. 2017 Jun;27(6):2570-2582.
Puylaert CA, Tielbeek JA, Bipat S, Boellaard TN, Nio CY, Stoker J.
Long-Term Performance of Readers Trained in Grading Crohn Disease Activity
Using MRI.
Acad Radiol. 2016 Dec;23(12):1539-1544.
Puylaert CA, van Thiel PP.
Images in Clinical Medicine. Katayama Fever.
N Engl J Med. 2016 Feb 4;374(5):469.
Puylaert CA, Tielbeek JA, Bipat S, Stoker J.
Grading of Crohn’s disease activity using CT, MRI, US and scintigraphy: a meta-
analysis. Eur Radiol. 2015 Nov;25(11):3295-313.
Li Z, Tielbeek JAW, Caan MWA, Puylaert CAJ, Ziech MLW, Nio CY, Stoker J, van
Vliet LJ, Vos FM.
Expiration-phase template-based motion correction of free-breathing abdominal
dynamic contrast enhanced MRI.
IEEE Trans Biomed Eng. 2015 Apr;62(4):1215-1225.
Proefschrift2018-new2.indb 193 13/9/18 10:07
193
List of publications
Naziroglu RE, Puylaert CAJ, Tielbeek JAW, Makanyanga J, Menys A, Ponsioen CY,
Hatzakis H, Taylor SA, Stoker J, van Vliet LJ, Vos FM.
Semi-automatic bowel wall thickness measurements on MR enterography in patients
with Crohn’s disease.
Br J Radiol. 2017 Jun;90(1074):20160654.
Bekendam MIJ, Puylaert CAJ, Phoa SKSS, Nio CY, Stoker J.
Shortened oral contrast preparation for improved small bowel distension at MR
enterography.
Abdom Radiol (NY). 2017 Sep;42(9):2225-2232.
Taylor SA, Avni F, Cronin CG, Hoeffel C, Kim SH, Laghi A, Napolitano M, Petit
P, Rimola J, Tolan DJ, Torkzad MR, Zappa M, Bhatnagar G, Puylaert CAJ, Stoker J.
The first joint ESGAR/ ESPR consensus statement on the technical performance of
cross-sectional small bowel and colonic imaging.
Eur Radiol. 2017 Jun;27(6):2570-2582.
Puylaert CA, Tielbeek JA, Bipat S, Boellaard TN, Nio CY, Stoker J.
Long-Term Performance of Readers Trained in Grading Crohn Disease Activity
Using MRI.
Acad Radiol. 2016 Dec;23(12):1539-1544.
Puylaert CA, van Thiel PP.
Images in Clinical Medicine. Katayama Fever.
N Engl J Med. 2016 Feb 4;374(5):469.
Puylaert CA, Tielbeek JA, Bipat S, Stoker J.
Grading of Crohn’s disease activity using CT, MRI, US and scintigraphy: a meta-
analysis. Eur Radiol. 2015 Nov;25(11):3295-313.
Li Z, Tielbeek JAW, Caan MWA, Puylaert CAJ, Ziech MLW, Nio CY, Stoker J, van
Vliet LJ, Vos FM.
Expiration-phase template-based motion correction of free-breathing abdominal
dynamic contrast enhanced MRI.
IEEE Trans Biomed Eng. 2015 Apr;62(4):1215-1225.
Proefschrift2018-new2.indb 193 13/9/18 10:07
Appendix
194
Proefschrift2018-new2.indb 194 13/9/18 10:07
Appendix
194
Proefschrift2018-new2.indb 194 13/9/18 10:07
195
Curriculum vitae
CURRICULUM VITAE
Carl Alejandro Julien Puylaert werd op 11 april 1989 in Den Haag geboren als tweede
kind van Julien Puylaert en Gabrielle Saleh. Samen met zijn oudere zus en jongere
tweelingzusjes groeide hij op in Oegstgeest. Na zijn eerste vijf jaar middelbare
school aan het Stedelijk Gymnasium Leiden, slaagde hij in 2007 cum laude voor
het eindexamen aan het Vrijbergen in Oegstgeest. Dezelfde zomer startte hij met
zijn studie Geneeskunde aan de Universiteit van Amsterdam. Na het behalen van
het doctoraalexamen, begon hij in november 2013 met zijn promotieonderzoek op
de afdeling Radiologie in het Academisch Medisch Centrum. Onder begeleiding
van Prof. dr. J. Stoker deed hij daar onderzoek in het Europese VIGOR++ project
naar de ontwikkeling van MRI bij de ziekte van Crohn. Tijdens zijn promotie was hij
tevens betrokken bij de ontwikkeling van de Europese richtlijn voor de technische
uitvoering van beeldvorming van de dunne en dikke darm. Naast het onderzoek is
hij ontwikkelaar van de medische app ‘the Radiology Assistant for Android’. Sinds
november 2016 is hij bezig met zijn coschappen, welke hij verwacht begin 2019 af
te ronden.
Proefschrift2018-new2.indb 195 13/9/18 10:07
195
Curriculum vitae
CURRICULUM VITAE
Carl Alejandro Julien Puylaert werd op 11 april 1989 in Den Haag geboren als tweede
kind van Julien Puylaert en Gabrielle Saleh. Samen met zijn oudere zus en jongere
tweelingzusjes groeide hij op in Oegstgeest. Na zijn eerste vijf jaar middelbare
school aan het Stedelijk Gymnasium Leiden, slaagde hij in 2007 cum laude voor
het eindexamen aan het Vrijbergen in Oegstgeest. Dezelfde zomer startte hij met
zijn studie Geneeskunde aan de Universiteit van Amsterdam. Na het behalen van
het doctoraalexamen, begon hij in november 2013 met zijn promotieonderzoek op
de afdeling Radiologie in het Academisch Medisch Centrum. Onder begeleiding
van Prof. dr. J. Stoker deed hij daar onderzoek in het Europese VIGOR++ project
naar de ontwikkeling van MRI bij de ziekte van Crohn. Tijdens zijn promotie was hij
tevens betrokken bij de ontwikkeling van de Europese richtlijn voor de technische
uitvoering van beeldvorming van de dunne en dikke darm. Naast het onderzoek is
hij ontwikkelaar van de medische app ‘the Radiology Assistant for Android’. Sinds
november 2016 is hij bezig met zijn coschappen, welke hij verwacht begin 2019 af
te ronden.
Proefschrift2018-new2.indb 195 13/9/18 10:07
Appendix
196
DANKWOORD
Al jaren kijk ik uit naar dit moment, maar het moet gezegd worden: de weg er naar
toe was minstens zo mooi. Onderzoek doe je niet alleen en er zijn dan ook veel
mensen die hebben bijgedragen aan de totstandkoming van dit proefschrift.
Allereerst wil ik alle patiënten bedanken voor hun deelname aan de VIGOR++ studie
en daarmee hun bijdrage aan het onderzoek naar de ziekte van Crohn.
Mijn promotoren, Prof. dr. J. Stoker en Prof. dr. ir. L.J. van Vliet.
Beste Jaap, ik mag mezelf gelukkig prijzen met een promotor met zoveel ervaring en
kennis. Het begeleiden van promovendi is voor jou een tweede natuur en daardoor
kon ik mij er op elk moment gerust van zijn dat dit promotietraject tot een mooi
resultaat zou leiden. Zelfs toen je hoofd van de afdeling werd, bleef je altijd nauw
betrokken bij mijn onderzoek.
Beste Lucas, mijn bezoeken aan de TU Delft waren een welkome afwisseling van het
klinisch onderzoek en een prachtige kans om te leren over de nieuwe ontwikkelingen
op het gebied van beeldanalyse. Dank ook voor jouw begeleiding van de Delftse
promovendi binnen ons project, waar ik altijd zo’n goede samenwerking mee heb
gehad.
Mijn co-promotoren, Dr. F.M. Vos en Dr. C.Y. Ponsioen.
Beste Frans, onze dagelijkse gesprekken waren een onmisbaar deel van mijn promotie
en brachten mij altijd nieuwe ideeën en kostbare leermomenten. We konden altijd
snel schakelen, maar tegelijkertijd gaf je de ruimte voor vrije gedachten, wat vaak
tot goede inzichten leidde. Ik ben trots op ons werk samen aan het VIGOR project
en de grote potentie voor vervolgonderzoek die het heeft opgeleverd.
Beste Cyriel, door jouw betrokken houding bij het VIGOR project leken de inclusies
voor onze studie eigenlijk vanzelf te gaan. Ik heb veel geleerd van onze gesprekken,
de vele scopieën waar ik heb kunnen meekijken en natuurlijk de IBD-besprekingen.
Jouw uitgebreide klinische ervaring en kennis, en de sterke band met je onderzoekers
zijn een voorbeeld voor elk begeleider.
Proefschrift2018-new2.indb 196 13/9/18 10:07
Appendix
196
DANKWOORD
Al jaren kijk ik uit naar dit moment, maar het moet gezegd worden: de weg er naar
toe was minstens zo mooi. Onderzoek doe je niet alleen en er zijn dan ook veel
mensen die hebben bijgedragen aan de totstandkoming van dit proefschrift.
Allereerst wil ik alle patiënten bedanken voor hun deelname aan de VIGOR++ studie
en daarmee hun bijdrage aan het onderzoek naar de ziekte van Crohn.
Mijn promotoren, Prof. dr. J. Stoker en Prof. dr. ir. L.J. van Vliet.
Beste Jaap, ik mag mezelf gelukkig prijzen met een promotor met zoveel ervaring en
kennis. Het begeleiden van promovendi is voor jou een tweede natuur en daardoor
kon ik mij er op elk moment gerust van zijn dat dit promotietraject tot een mooi
resultaat zou leiden. Zelfs toen je hoofd van de afdeling werd, bleef je altijd nauw
betrokken bij mijn onderzoek.
Beste Lucas, mijn bezoeken aan de TU Delft waren een welkome afwisseling van het
klinisch onderzoek en een prachtige kans om te leren over de nieuwe ontwikkelingen
op het gebied van beeldanalyse. Dank ook voor jouw begeleiding van de Delftse
promovendi binnen ons project, waar ik altijd zo’n goede samenwerking mee heb
gehad.
Mijn co-promotoren, Dr. F.M. Vos en Dr. C.Y. Ponsioen.
Beste Frans, onze dagelijkse gesprekken waren een onmisbaar deel van mijn promotie
en brachten mij altijd nieuwe ideeën en kostbare leermomenten. We konden altijd
snel schakelen, maar tegelijkertijd gaf je de ruimte voor vrije gedachten, wat vaak
tot goede inzichten leidde. Ik ben trots op ons werk samen aan het VIGOR project
en de grote potentie voor vervolgonderzoek die het heeft opgeleverd.
Beste Cyriel, door jouw betrokken houding bij het VIGOR project leken de inclusies
voor onze studie eigenlijk vanzelf te gaan. Ik heb veel geleerd van onze gesprekken,
de vele scopieën waar ik heb kunnen meekijken en natuurlijk de IBD-besprekingen.
Jouw uitgebreide klinische ervaring en kennis, en de sterke band met je onderzoekers
zijn een voorbeeld voor elk begeleider.
Proefschrift2018-new2.indb 196 13/9/18 10:07
197
Dankwoord
Geachte leden van de promotiecommissie, Prof. dr. J.S. Laméris, Prof. dr. G.R.A.M.
D’Haens, Prof. dr. W.A. Bemelman, Prof. dr. W.M. Prokop, Prof. dr. J.P.W. Pluim en Dr.
D.J. de Jong, dank voor de kritische beoordeling van dit manuscript en voor het
zitting nemen in de commissie.
Mede-auteurs van de verschillende studies uit dit proefschrift, veel dank voor jullie
inspanningen bij het vervaardigen van de artikelen in dit manuscript.
All of my colleagues from the VIGOR++ project, thank you so much for your
dedication and incredible work over the last years. Our meetings were highlights
of my time as a PhD student. Special thanks to Peter, Alex, Stuart, Robiel and
Zhang for their integral part in the completion of this thesis. Many thanks to the
radiologists that participated as readers in the included studies: Yung, Banafsche,
Karin, Douglas, and all the participants of the Crohn’s MRI course, thank you for
your time and effort. Alle endoscopisten en andere medewerkers van de afdeling
MDL, die hebben bijgedragen aan de VIGOR++ studie – bij jullie beter bekend als
de ‘Acute Crohn Snelweg” – dank voor jullie inzet en hulp bij de vele succesvolle
inclusies.
Ann-Sophie, dank je voor onze gezellige gesprekken en natuurlijk alle momenten
dat je me last-minute op Jaap zijn agenda wist te zetten.
Mijn collega onderzoekers van G1, Laura, Ruud, Ivo, Kilian, Sjoerd, Moos, Conijn,
Jordi, Henk-Jan, Charlotte, Anouk, Jurgen, Onno, Martine en Kerensa, dank voor de
vele gezellige koffiepauzes, vrijmibo’s, kerstdiners, sinterklaasvieringen en de altijd
goede sfeer binnen onze groep. Mijn kamergenoten, Sjoerd, Jordi en Ruud, ik kan
niet zeggen dat ik door jullie harder ben gaan werken, maar gezellig was het zeker!
Nussie en Floortje, wie had gedacht dat je aan een promotie zulke goede vrienden
kon overhouden. Olvert ook natuurlijk, maar die komt later aan bod. We hebben al
zoveel meegemaakt, dat ik niet weet waar ik moet beginnen: samen feesten in de
Bourbon street, Floortje die (weer eens) moest huilen van het lachen, Stivale d’Oro,
onze vakantie samen en de daarop volgende Malaga diners, en er gaat ongetwijfeld
nog veel moois op volgen!
Proefschrift2018-new2.indb 197 13/9/18 10:07
197
Dankwoord
Geachte leden van de promotiecommissie, Prof. dr. J.S. Laméris, Prof. dr. G.R.A.M.
D’Haens, Prof. dr. W.A. Bemelman, Prof. dr. W.M. Prokop, Prof. dr. J.P.W. Pluim en Dr.
D.J. de Jong, dank voor de kritische beoordeling van dit manuscript en voor het
zitting nemen in de commissie.
Mede-auteurs van de verschillende studies uit dit proefschrift, veel dank voor jullie
inspanningen bij het vervaardigen van de artikelen in dit manuscript.
All of my colleagues from the VIGOR++ project, thank you so much for your
dedication and incredible work over the last years. Our meetings were highlights
of my time as a PhD student. Special thanks to Peter, Alex, Stuart, Robiel and
Zhang for their integral part in the completion of this thesis. Many thanks to the
radiologists that participated as readers in the included studies: Yung, Banafsche,
Karin, Douglas, and all the participants of the Crohn’s MRI course, thank you for
your time and effort. Alle endoscopisten en andere medewerkers van de afdeling
MDL, die hebben bijgedragen aan de VIGOR++ studie – bij jullie beter bekend als
de ‘Acute Crohn Snelweg” – dank voor jullie inzet en hulp bij de vele succesvolle
inclusies.
Ann-Sophie, dank je voor onze gezellige gesprekken en natuurlijk alle momenten
dat je me last-minute op Jaap zijn agenda wist te zetten.
Mijn collega onderzoekers van G1, Laura, Ruud, Ivo, Kilian, Sjoerd, Moos, Conijn,
Jordi, Henk-Jan, Charlotte, Anouk, Jurgen, Onno, Martine en Kerensa, dank voor de
vele gezellige koffiepauzes, vrijmibo’s, kerstdiners, sinterklaasvieringen en de altijd
goede sfeer binnen onze groep. Mijn kamergenoten, Sjoerd, Jordi en Ruud, ik kan
niet zeggen dat ik door jullie harder ben gaan werken, maar gezellig was het zeker!
Nussie en Floortje, wie had gedacht dat je aan een promotie zulke goede vrienden
kon overhouden. Olvert ook natuurlijk, maar die komt later aan bod. We hebben al
zoveel meegemaakt, dat ik niet weet waar ik moet beginnen: samen feesten in de
Bourbon street, Floortje die (weer eens) moest huilen van het lachen, Stivale d’Oro,
onze vakantie samen en de daarop volgende Malaga diners, en er gaat ongetwijfeld
nog veel moois op volgen!
Proefschrift2018-new2.indb 197 13/9/18 10:07
Appendix
198
Jeroen, dank voor het warme onthaal dat ik van jou kreeg op G1 en de geweldige
kickstart die je gaf aan mijn promotie. Je droeg jouw onderzoek aan mij over, maar
bleef altijd nauw betrokken en geïnteresseerd in de voortgang. Ik kijk met plezier
terug naar onze vele gesprekken, trips naar de ECR en Londen, de MRI cursus en
natuurlijk ons mooie artikel op de Radiology Assistant.
Sofieke en Kyra, mijn sparring partners over MRI Crohn. Sofieke, wat was het leuk
om na een tijd als enige Crohn onderzoeker opeens een metgezel te hebben bij
ons onderzoek. Je hebt je eigen draai eraan gegeven en een ontzettend originele
onderzoekslijn opgezet. Kyra, met jouw leergierigheid en enthousiasme ben je een
prachtige aanwinst voor de groep. Ik kijk al uit naar jullie promoties!
Charlotte Tuut, wat gaaf om samen met mijn grote nicht een artikel te hebben
geschreven. Als Boppie dat zou weten! Gelukkig hebben we nog veel andere trotse
familieleden, die we eindeloos kunnen vertellen over onze prestatie.
Mijn leuke co-groep, wat is het een plezier om samen met jullie dokter te worden.
Dank dat jullie altijd zo geïnteresseerd zijn geweest in mijn onderzoek en voor de
ontzettend gezellige tijd samen.
Mijn (oud-)huisgenoten, Remco en Pieter, met jullie was thuiskomen op de
Rustenburgerstraat aka the Rusty altijd een feest. Drie werkende mannen met
de illusie nog student te zijn – wat een zorgeloos bestaan. Dank voor jullie altijd
oprechte interesse in mijn onderzoek, de gezellige tijd en het eindeloze lachen om
goede (of hele slechte) grappen!
Catherine en Olvert, wat ben ik blij en trots om jullie vandaag naast me te hebben
staan als mijn paranimfen. Lieve Caas, sister, ik ken je al mijn hele leven als zus,
maar nu kennen we elkaar inmiddels ook op veel andere manieren – samen naar
Amsterdam, samen geneeskunde studeren, samen bij het Corps – je bent niet weg
te denken uit mijn leven. Wat ben ik blij dat je vandaag mijn paranimf wil zijn!
Ollie, naast onze gezellige tijd samen bij de radiologie, konden wij samen ook lekker
klagen. We hebben allebei een sterke ‘inner nerd’ (jij iets meer dan ik), dus er is
altijd ruimte voor een goed gesprek over harde schijven of andere gadgets. Met
jouw energie ben je altijd een plezier om in de buurt te hebben. Mooi dat jij vandaag
naast me staat als paranimf!
Proefschrift2018-new2.indb 198 13/9/18 10:07
Appendix
198
Jeroen, dank voor het warme onthaal dat ik van jou kreeg op G1 en de geweldige
kickstart die je gaf aan mijn promotie. Je droeg jouw onderzoek aan mij over, maar
bleef altijd nauw betrokken en geïnteresseerd in de voortgang. Ik kijk met plezier
terug naar onze vele gesprekken, trips naar de ECR en Londen, de MRI cursus en
natuurlijk ons mooie artikel op de Radiology Assistant.
Sofieke en Kyra, mijn sparring partners over MRI Crohn. Sofieke, wat was het leuk
om na een tijd als enige Crohn onderzoeker opeens een metgezel te hebben bij
ons onderzoek. Je hebt je eigen draai eraan gegeven en een ontzettend originele
onderzoekslijn opgezet. Kyra, met jouw leergierigheid en enthousiasme ben je een
prachtige aanwinst voor de groep. Ik kijk al uit naar jullie promoties!
Charlotte Tuut, wat gaaf om samen met mijn grote nicht een artikel te hebben
geschreven. Als Boppie dat zou weten! Gelukkig hebben we nog veel andere trotse
familieleden, die we eindeloos kunnen vertellen over onze prestatie.
Mijn leuke co-groep, wat is het een plezier om samen met jullie dokter te worden.
Dank dat jullie altijd zo geïnteresseerd zijn geweest in mijn onderzoek en voor de
ontzettend gezellige tijd samen.
Mijn (oud-)huisgenoten, Remco en Pieter, met jullie was thuiskomen op de
Rustenburgerstraat aka the Rusty altijd een feest. Drie werkende mannen met
de illusie nog student te zijn – wat een zorgeloos bestaan. Dank voor jullie altijd
oprechte interesse in mijn onderzoek, de gezellige tijd en het eindeloze lachen om
goede (of hele slechte) grappen!
Catherine en Olvert, wat ben ik blij en trots om jullie vandaag naast me te hebben
staan als mijn paranimfen. Lieve Caas, sister, ik ken je al mijn hele leven als zus,
maar nu kennen we elkaar inmiddels ook op veel andere manieren – samen naar
Amsterdam, samen geneeskunde studeren, samen bij het Corps – je bent niet weg
te denken uit mijn leven. Wat ben ik blij dat je vandaag mijn paranimf wil zijn!
Ollie, naast onze gezellige tijd samen bij de radiologie, konden wij samen ook lekker
klagen. We hebben allebei een sterke ‘inner nerd’ (jij iets meer dan ik), dus er is
altijd ruimte voor een goed gesprek over harde schijven of andere gadgets. Met
jouw energie ben je altijd een plezier om in de buurt te hebben. Mooi dat jij vandaag
naast me staat als paranimf!
Proefschrift2018-new2.indb 198 13/9/18 10:07
199
Dankwoord
Lieve familie van Floor, wat een geluk dat ik met jullie zo een leuke en gezellige
familie erbij heb gekregen. Altijd enthousiast en geïnteresseerd in wat ik doe. Ik ben
blij met de hechte band die we hebben en dat jullie er vandaag ook bij zullen zijn.
Lieve Boppie, wat had ik dit boek graag aan u willen geven. Met Bommie aan uw
zijde gaf u onze familie een begerenswaardig voorbeeld om naar te streven. Ik
ben trots dat ik nu als derde generatie op rij een radiologisch proefschrift mag
verdedigen.
Lieve opa en oma, op de Snipweg geven jullie ons al familie een tweede thuis op
Curacao. TIjdens de vele avonden met jullie op de porch leer ik elke keer wel iets
nieuws: van muziek, politiek, tot een geneeskundige les. Opa, ik ben blij dat u als
eerste mijn proefschrift hebt kunnen lezen en dat deze de goedkeuring van onze
Prof heeft mogen ontvangen. Ayo!
Mijn lieve familie, wat ben ik gezegend met zo een warm en gezellig nest. Eem en
Gi, wat ben ik blij met jullie als zusjes en onze hechte band. Ik ben ongelofelijk trots
op hoe jullie je hebben ontwikkeld, de een als arts, de ander als zangeres, maar
beide met passie, doorzettingsvermogen en jullie altijd creatieve geest. Lieve pap
en mam, als ik er even tussen uit moest kon ik bij jullie in Oegstgeest altijd écht
thuiskomen. Pap, wat ben ik blij met onze sterke band en dat we elkaar deze jaren
als collega’s zoveel hebben kunnen zien. Dank dat ik op elk moment kon rekenen
op jouw goede advies, waarmee ik elk probleem aan kon. Mam, bedankt voor je
onvoorwaardelijke steun en liefde bij alles wat ik doe. Nu mijn promotie op jouw
verjaardag – I hope I’m not stealing your thunder. Laten we het samen vieren en er
een groot feest van maken!
Allerliefste Floor, dank voor al je liefde en de mooie momenten die ik elke dag met
je beleef. Ik kan niet wachten op wat het leven ons samen nog gaat brengen!
Proefschrift2018-new2.indb 199 13/9/18 10:07
199
Dankwoord
Lieve familie van Floor, wat een geluk dat ik met jullie zo een leuke en gezellige
familie erbij heb gekregen. Altijd enthousiast en geïnteresseerd in wat ik doe. Ik ben
blij met de hechte band die we hebben en dat jullie er vandaag ook bij zullen zijn.
Lieve Boppie, wat had ik dit boek graag aan u willen geven. Met Bommie aan uw
zijde gaf u onze familie een begerenswaardig voorbeeld om naar te streven. Ik
ben trots dat ik nu als derde generatie op rij een radiologisch proefschrift mag
verdedigen.
Lieve opa en oma, op de Snipweg geven jullie ons al familie een tweede thuis op
Curacao. TIjdens de vele avonden met jullie op de porch leer ik elke keer wel iets
nieuws: van muziek, politiek, tot een geneeskundige les. Opa, ik ben blij dat u als
eerste mijn proefschrift hebt kunnen lezen en dat deze de goedkeuring van onze
Prof heeft mogen ontvangen. Ayo!
Mijn lieve familie, wat ben ik gezegend met zo een warm en gezellig nest. Eem en
Gi, wat ben ik blij met jullie als zusjes en onze hechte band. Ik ben ongelofelijk trots
op hoe jullie je hebben ontwikkeld, de een als arts, de ander als zangeres, maar
beide met passie, doorzettingsvermogen en jullie altijd creatieve geest. Lieve pap
en mam, als ik er even tussen uit moest kon ik bij jullie in Oegstgeest altijd écht
thuiskomen. Pap, wat ben ik blij met onze sterke band en dat we elkaar deze jaren
als collega’s zoveel hebben kunnen zien. Dank dat ik op elk moment kon rekenen
op jouw goede advies, waarmee ik elk probleem aan kon. Mam, bedankt voor je
onvoorwaardelijke steun en liefde bij alles wat ik doe. Nu mijn promotie op jouw
verjaardag – I hope I’m not stealing your thunder. Laten we het samen vieren en er
een groot feest van maken!
Allerliefste Floor, dank voor al je liefde en de mooie momenten die ik elke dag met
je beleef. Ik kan niet wachten op wat het leven ons samen nog gaat brengen!
Proefschrift2018-new2.indb 199 13/9/18 10:07
Proefschrift2018-new2.indb 200 13/9/18 10:07 Proefschrift2018-new2.indb 200 13/9/18 10:07