UvA-DARE (Digital Academic Repository) MRI for Crohn's ...Eur Radiol. 2015 Nov;25(11):3295-313 17 Chapter 3 Comparison of MRI activity scoring systems and features for the terminal

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

https://dare.uva.nl/personal/pure/en/publications/mri-for-crohns-disease(25781b84-5e6d-47ce-a91c-93136ae72a95).html

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

[email protected]

06-15215881

Paranimfen:

Catherine [email protected]

06-11070412

Olvert [email protected]

06-50640649

Uitnodiging promotie.indd 1 12/9/18 21:31


Carl A.J. Puylaert

Proefschrift2018-new2.indb 1 13/9/18 10:06


Carl A.J. Puylaert


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).


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).



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



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


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


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


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


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 6 13/9/18 10:06 Proefschrift2018-new2.indb 6 13/9/18 10:06

CHAPTER 1

Introduction


CHAPTER 1

Introduction


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


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


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


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


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


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


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.


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.


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.


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.


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.


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.


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


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.




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,


79.

9. Panes J, Bouhnik Y, Reinisch W, et al. Imaging techniques for assessment of inflammatory


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.







14. McDonald RJ, McDonald JS, Kallmes DF, et al. Intracranial Gadolinium Deposition after

Contrast-enhanced MR Imaging. Radiology 2015;0:150025.








18. Buisson A, Joubert A, Montoriol P-F, et al. Diffusion-weighted magnetic resonance imaging


2013;37:537–45.




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.


Introduction

15

Bowel Dis 2011;17:1759–68.






Crohn disease endoscopic index of severity. Am J Roentgenol 2013;201:1220–8.




measurements on MR enterography in patients with Crohn’s disease. Br J Radiol 2017;90.



2015;62:1215–25.



Dis 2015;21:101–9.






Eur Radiol 2012;22:2494–501.

28. Cullmann JL, Bickelhaupt S, Froehlich JM, et al. MR imaging in Crohn’s disease: Correlation


Motil 2013;25.



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


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


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.


18

Chapter 2

ABSTRACT

Objectives

To assess the grading of Crohn’s disease activity using CT, MRI, US and scintigraphy.


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.


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.


19


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.


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)


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.


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)


21


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


21


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


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


22

Chapter 2



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.


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 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.


We recorded the following information regarding interpretation of imaging and

reference tests: the interval in days between index and reference tests, bowel


23


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


23


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


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.


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.


25


Figure 1. Flow diagram showing study selection.



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).


25


Figure 1. Flow diagram showing study selection.



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).


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 %


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 %


27


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)


27


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


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)


28

Chapter 2


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 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


28

Chapter 2


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 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


29


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]


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


3.0 T Phased array body coil


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


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]


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


2D T1-FLASH, 2D & 3D T1-FLASH with fat suppres-sion (post-con-trast), bSSFP, RARE

Schreyer 2005 [23]


Macro-gol 3350

1.5 L Gd (5 mmol/L) mixture with water rectally



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


29


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


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


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


Interpolated 3D T1 with fat suppression (post-contrast), SSFP, T2 with fat suppression




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]


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




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


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]


12 hrs fasting

2 L Mannitol solution with carob seed 1 hr prior orally and 0.4–1.0 L 0.9% NaCl rectally



2D T1-FLASH, 2D & 3D T1-FLASH with fat suppres-sion (post-con-trast), bSSFP, RARE

Schreyer 2005 [23]


Macro-gol 3350

1.5 L Gd (5 mmol/L) mixture with water rectally



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


30

Chapter 2


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)




370–555 MBq 3 scans (at 30 min, 2–2.5 hrs and 24 hrs)


MBq = Megabecquerel, NS = Not Specified, Tc-99m HMPAO = Technetium hexamethylpropyleneamine oxime


30

Chapter 2


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)


Biancone 1997 [34]


185 MBq 2 scans (at 30 min and 3 hrs)




370–555 MBq 3 scans (at 30 min, 2–2.5 hrs and 24 hrs)


MBq = Megabecquerel, NS = Not Specified, Tc-99m HMPAO = Technetium hexamethylpropyleneamine oxime


31


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.


31


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.


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

.


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

.


33


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

.


33


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

.


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

).


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

).


35


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

).


35


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

).


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)

.


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)

.


37


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)

.


37


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)

.


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


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 %


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)


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


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 %


39


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


39


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


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


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


41


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.


41


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.


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 disease: 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 resonance 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 studies. 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 resection: complementary role to endoscopy. Inflamm

Bowel Dis 19: 977–982

18. Kolkman JJ, Falke TH, Roos JC et al (1996) Computed tomography and granulocyte

scintigraphy in active inflammatory bowel disease. Comparison with endoscopy and

operative findings. Dig Dis Sci 41:641–650


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 disease: 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 resonance 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 studies. 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 resection: complementary role to endoscopy. Inflamm

Bowel Dis 19: 977–982

18. Kolkman JJ, Falke TH, Roos JC et al (1996) Computed tomography and granulocyte

scintigraphy in active inflammatory bowel disease. Comparison with endoscopy and

operative findings. Dig Dis Sci 41:641–650


43


19. Shoenut JP, Semelka RC, Silverman R, Yaffe CS, Micflikier AB (1993) Magnetic resonance

imaging in inflammatory bowel disease. 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

inflammatory bowel disease. J Clin Gastroenterol 19:31–35

21. Florie J, Horsthuis K, Hommes DW et al (2005) Magnetic resonance imaging compared

with ileocolonoscopy in evaluating disease severity in Crohn’s disease. Clin Gastroenterol

Hepatol 3: 1221–1228

22. Schreyer AG, Golder S, Scheibl K et al (2005) Dark lumen magnetic 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 colonoscopy 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 resonance 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


43


19. Shoenut JP, Semelka RC, Silverman R, Yaffe CS, Micflikier AB (1993) Magnetic resonance

imaging in inflammatory bowel disease. 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

inflammatory bowel disease. J Clin Gastroenterol 19:31–35

21. Florie J, Horsthuis K, Hommes DW et al (2005) Magnetic resonance imaging compared

with ileocolonoscopy in evaluating disease severity in Crohn’s disease. Clin Gastroenterol

Hepatol 3: 1221–1228

22. Schreyer AG, Golder S, Scheibl K et al (2005) Dark lumen magnetic 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 colonoscopy 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 resonance 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


44

Chapter 2

36. Masselli G, Gualdi G (2013) CT and MR enterography in evaluating 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 comparison 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 symptomatic 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 ionizing 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: derivation and histopathological validation of an MR-

based activity index. Eur J Radiol 81:2080–2088


44

Chapter 2

36. Masselli G, Gualdi G (2013) CT and MR enterography in evaluating 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 comparison 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 symptomatic 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 ionizing 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: derivation and histopathological validation of an MR-

based activity index. Eur J Radiol 81:2080–2088


45


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


45



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


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


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


47


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


47


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


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


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


49


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


49


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


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


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


51


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


51


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


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)


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


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


52

Chapter 2

Table 2. Magnetic resonance imaging (MRI)


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)


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


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




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


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



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.


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.


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.


57


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.


58

Chapter 3


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


58

Chapter 3


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


59


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


59


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


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,


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,


61


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).


61


London, MaRIA and Clermont score) were calculated using the formulas described

in the Supplementary Data 2 (3,5,8).


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


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).


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.


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.


63


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).


63



Variable N [%]


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).


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


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.


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)


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


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.


65


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).


65


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).


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%


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%


67


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.


67


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.


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.


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.


69


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


69


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


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.


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






71


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.



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,


71


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.


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.


Disease Using MRI (the VIGOR++ Project). Acad Radiol. 2018.




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.


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,


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.


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


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.


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


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.


AJR Am J Roentgenol. 2016 Mar;1–9.

30. Pendsé DA, Makanyanga JC, Plumb AA, Bhatnagar G, Atkinson D, Rodriguez-Justo M, et


73


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


Crohn’s disease assessment with histopathology of surgical specimens. Eur Radiol.

2014;24(3):619–29.


73


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


Crohn’s disease assessment with histopathology of surgical specimens. Eur Radiol.

2014;24(3):619–29.


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


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


75


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


75


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


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


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.


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.


77


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.


77


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.



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



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



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.


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.


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.


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



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


81


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


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.


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


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.


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.


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


83


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).


83


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).


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


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


85


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)


85


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)


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


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


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


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


87


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


87


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


88

Chapter 4

Table 1. Clinical Characteristics of the Prospective Cohort


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.


88

Chapter 4

Table 1. Clinical Characteristics of the Prospective Cohort


Female, n (%) 59 (56)

Age at MRI (y), median (IQR) 33 (26–44)




Steroids, n (%) 18 (17)


5-ASA, n (%) 19 (18)



HBI value, median (IQR) 5 (2–8)

CDEIS score, median (IQR) 3.2 (0.5–6.4)


Age at diagnosis (y), median (IQR) 22 (17–28)

Disease location

L1 ileal, n (%) 43 (41)

L2 colonic, n (%) 15 (14)



Disease behavior





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.


89


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).


89


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).


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.


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.


91


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.


91


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.


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.


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.


93


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


93


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


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.


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.


95


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






95


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






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.


disease activity and severity in ileocolonic Crohn’s disease. Gut 2009; 58:1113–1120.





disease: validation of parameters of severity and quantitative index of activity. Inflamm

Bowel Dis 2011; 17:1759–1768.



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.


Crohn’s disease activity: Future or fiction? Abdom Imaging 2012; 37:967–973.


measurements on MR enterography in patients with Crohn’s disease. Br J Radiol 2017;

20160654.



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.



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.


96

Chapter 4


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.



comparison with surgical pathologic analysis. Inflamm Bowel Dis 2011; 17:984–993.


disease activity and severity in ileocolonic Crohn’s disease. Gut 2009; 58:1113–1120.






Bowel Dis 2011; 17:1759–1768.



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.


Crohn’s disease activity: Future or fiction? Abdom Imaging 2012; 37:967–973.


measurements on MR enterography in patients with Crohn’s disease. Br J Radiol 2017;

20160654.



2015; 62:1215–1225.


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.



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.


97


18. Steiger JH. Tests for comparing elements of a correlation matrix. Psychol Bull 1980;

87:245–251.


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.


97


18. Steiger JH. Tests for comparing elements of a correlation matrix. Psychol Bull 1980;

87:245–251.


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.



Gastroenterology 2014; 146:374–382, e1.


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].


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].


99


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


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


99


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


Coronal 2 380x380-459

2.2-2.4 1.0-1.1 10


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


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


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%).


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





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%).


101


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


101


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


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.


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.


103


Clinical characteristics of the subset group from the prospective cohort


Female, n (%) 29 (58)

Age at MRI (years), median (IQR) 32 (27–47)




Steroids, n. (%) of patients 8 (16)

Thiopurines, no. (%) 8 (16)

5-ASA, no. (%) of patients 10 (20)

Methotrexate, no. (%) 3 (6)


HBI, median (IQR) 5 (2–8)

CDEIS, median (IQR) 4.0 (0.1–7.7)


Age at diagnosis (years), median (IQR) 22 (19–28)

Disease location

L1 ileal, n (%) 21 (42)

L2 colonic, n (%) 9 (18)



Disease behaviour





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.


103


Clinical characteristics of the subset group from the prospective cohort


Female, n (%) 29 (58)

Age at MRI (years), median (IQR) 32 (27–47)




Steroids, n. (%) of patients 8 (16)

Thiopurines, no. (%) 8 (16)

5-ASA, no. (%) of patients 10 (20)

Methotrexate, no. (%) 3 (6)


HBI, median (IQR) 5 (2–8)

CDEIS, median (IQR) 4.0 (0.1–7.7)


Age at diagnosis (years), median (IQR) 22 (19–28)

Disease location

L1 ileal, n (%) 21 (42)

L2 colonic, n (%) 9 (18)



Disease behaviour





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.


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.


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.


105



105



106

Chapter 4

REFERENCES1. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading crohn disease activity with MRI:


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/.


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.


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.



based activity index. Eur. J. Radiol. 2012;81(9):2080–2088.





106

Chapter 4

REFERENCES1. Tielbeek JAW, Makanyanga JC, Bipat S, et al. Grading crohn disease activity with MRI:


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/.


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.


disease activity and severity in ileocolonic Crohn’s disease. Gut. 2009;58:1113–1120.


Platform-Dependent Image Scaling. Transl. Oncol. 2014;7(1):65–71.



intestinal healing in Inflammatory Bowel Disease. J. Crohn’s Colitis. 2011;5(5):484–498.



based activity index. Eur. J. Radiol. 2012;81(9):2080–2088.







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


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


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.


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.


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].


111


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].


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.


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.


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.


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.


113


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.


113


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


(model CF-160L, Olympus) within two weeks of the MRI examination by either a

gastroenterologist or senior resident under direct supervision of a gastroenterologist.


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.


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.


115


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




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


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


115


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




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


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


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


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


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


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


117


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).


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)


Therapy at time of MRI

5-ASA, n (%) 6 (10)

Corticosteroids, n (%) 15 (25)




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).


117


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).


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)


Therapy at time of MRI

5-ASA, n (%) 6 (10)

Corticosteroids, n (%) 15 (25)




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).


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


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


119


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


119


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


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


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


121


(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.


121


(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.


122

Chapter 5



Colitis 2013;7:556–85.




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.



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.


AJR Am J Roentgenol 2016;:1–9.



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.




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.


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.







122

Chapter 5



Colitis 2013;7:556–85.




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.



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.





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.




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.


15. 15 McDonald RJ, McDonald JS, Kallmes DF, et al. Intracranial Gadolinium Deposition

after Contrast-enhanced MR Imaging. Radiology 2015;0:150025.



Radiol Published Online First: 2018.







123


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.


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.



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.


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.


123


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.


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.



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.


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.



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


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


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.


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.


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.


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.


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.


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



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.


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


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,


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,



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


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



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


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


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


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



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.


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



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.


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


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)


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)



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)



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)


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


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



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


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).







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



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).







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).







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


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).


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).



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)



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)


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


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



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



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


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


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



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.



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.


Chapter 6

142


bowel disease: joint ECCO and ESGAR evidence-based consensus guidelines. J Crohns

Colitis 2013;7(7):556–85.


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.



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.



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.


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


Chapter 6

142


bowel disease: joint ECCO and ESGAR evidence-based consensus guidelines. J Crohns

Colitis 2013;7(7):556–85.


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.



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.



Pharmacol Ther 2015;42(3):343-55.



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.


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



143


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.



2015;62(4):1215–25.


measurements on MR enterography in patients with Crohn’s disease. Br J Radiol [Internet]

2017;90(1074):20160654.



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.


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.



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.



143


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.



2015;62(4):1215–25.


measurements on MR enterography in patients with Crohn’s disease. Br J Radiol [Internet]

2017;90(1074):20160654.



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.


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.



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.


Chapter 6

144


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


Chapter 6

144


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



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



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


Chapter 6

146

Appendix 5. Per-hospital analysis

CDEIS Standard at AMC


95% Con-fi dence interval

Signifi cance (P) against motility

Specifi city n=49(TN+FP)

95% Con-fi dence Interval


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







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







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







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


Chapter 6

146

Appendix 5. Per-hospital analysis

CDEIS Standard at AMC







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







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







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







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



147

Appendix 6. Scatter plot of CDEIS & eAIS.



147

Appendix 6. Scatter plot of CDEIS & eAIS.



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


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


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.


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.


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.


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.


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 successfully 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.


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.



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 successfully 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.


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.


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.


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.



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



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


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.


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.



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.



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.


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


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



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).



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


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).


Chapter 7

158

Table 5. Individual MRI features

MRI feature (numbers in parentheses reported by reference standard)


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



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


Chapter 7

158

Table 5. Individual MRI features

MRI feature (numbers in parentheses reported by reference standard)


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



Abdominal MRI



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



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



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


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


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



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.



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.


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.


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.




CHAPTER 8

General discussion


CHAPTER 8

General discussion


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


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


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.


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.


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


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


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.


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].


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.


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


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


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


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.


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.


Chapter 8

172

REFERENCES 1. Horsthuis K, Bipat S, Bennink RJ, et al. Inflammatory bowel disease diagnosed with US,


79.



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.






6. Buisson A, Joubert A, Montoriol PF, et al. Diffusion-weighted magnetic resonance imaging


2013;37:537–45.



Bowel Dis 2011;17:1759–68.






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.




measurements on MR enterography in patients with Crohn’s disease. Br J Radiol

2017;:20160654.



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.


Chapter 8

172

REFERENCES 1. Horsthuis K, Bipat S, Bennink RJ, et al. Inflammatory bowel disease diagnosed with US,


79.



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.






6. Buisson A, Joubert A, Montoriol PF, et al. Diffusion-weighted magnetic resonance imaging


2013;37:537–45.



Bowel Dis 2011;17:1759–68.






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.




measurements on MR enterography in patients with Crohn’s disease. Br J Radiol

2017;:20160654.



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.


General discussion

173





Dis 2015;21:101–9.



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.



Eur Radiol 2012;22:2494–501.



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.


General discussion

173





Dis 2015;21:101–9.



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.



Eur Radiol 2012;22:2494–501.



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.



CHAPTER 9

Summary


CHAPTER 9

Summary


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.


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.


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


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.


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.


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


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.


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.


Chapter 9

178


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.


Chapter 9

178


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.




CHAPTER 10

Nederlandse samenvatting


CHAPTER 10



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


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



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



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


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.


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.




APPENDIX

List of contributing authorsPortfolioList of publicationsCurriculum vitaeDankwoord


APPENDIX

List of contributing authorsPortfolioList of publicationsCurriculum vitaeDankwoord


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


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


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


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


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


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





Athens

2017 0.5

Conferences



Salzburg

2014 1.0


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



VIGOR++ workshop, UCL London 2014 0.2

Presentations



Salzburg

2014 0.5



Young Initiative on Crohn's and colitis, Amsterdam 2015 0.5






Athens

2017 0.5

Conferences



Salzburg

2014 1.0


191

Portfolio

IBD Today & Tomorrow, Amsterdam 2014 0.5






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


191

Portfolio

IBD Today & Tomorrow, Amsterdam 2014 0.5






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


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.


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.


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.


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.


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.


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.


Appendix

194


Appendix

194


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.


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.


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.


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.


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!


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!


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!


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!


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!


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!



Documents

UvA-DARE (Digital Academic Repository) MRI for Crohn's ...Eur Radiol. 2015 Nov;25(11):3295-313 17 Chapter 3 Comparison of MRI activity scoring systems and features for the terminal