PUBLISHED BY THE BELGIAN SOCIETY OF RADIOLOGY (BSR)

DIAGNOSTIC AND INTERVENTIONAL IMAGING, RELATED IMAGING SCIENCES,AND CONTINUING EDUCATION

WETTEREN 1

2 Volume 98 Page 63-103

March-April

Bimonthly – 2015

P 702083

The Belgian radiology journal wishes to thank Olea medical for their continuous support

The Belgian radiology journal wishes to thank Toshiba for their continuous support

The Belgian radiology journal wishes to thank Bracco for their continuous support

The Belgian radiology journal wishes to thank Guerbet for their continuous support

Contrast for life

PLATINUM SPONSOR

BRONZE SPONSORBRONZE SPONSOR

GOLD SPONSOR

Sections of the Belgian Radiological Society (BSR):

Abdominal and digestive imaging B. Op de Beeck, E. DanseBone and joints J.F. Nisolle, M. ShahabpourBreast imaging M. Mortier, S. MurgoCardiac imaging N. Mollet, A. NchimiCardiovascular and interventional radiology S. Heye, D. HenroteauxChest radiology B. Ghaye, W. De WeverHead and neck radiology J. Widelec, R. HermansNeuroradiology M. Lemmerling, L. TshibandaPediatric radiology B. Desprechins, L. Breysem

For addresses and particulars, see website at http://www.bsr-web.be

Instructions to authors

The purpose of The Belgian Journal of Radio­logy is the publication of articles dealing with diagnostic radiology and related imag­ing techniques, therapeutic radiology, allied sciences and continuing education. All — new and revised — manuscripts and correspond­ence should be addressed to JBR­BTR Edi to­rial Office, Avenue W. Churchill 11/30, B­1180 Bruxelles, tel.: 02­374 25 55, fax: 32­2­374 96 28.Please note that the following instructions are based on the “Uniform Requirements for manuscripts Submitted to Biomedical Jour­nals” adopted by the International Committee of Medical Journal Editors (Radiology, 1980,135: 239­243). It should however be noted that presentation modifications may be introduced by the Editorial Office in order to conform with the JBR­BTR personal style.Authors should specify to which of the fol­lowing headings their manuscript is intended:Original Article, Review Article, Case Report, Pictorial Essay, Continuing Education, Technical Note, Book Review, Opinion, Letter to the Editor, Comment, Meeting News, in Memoriam, News.Authors should consider the following remarks and submit their manuscripts accord­ingly.All articles must contain substantive and speci fic scientific material.– Original articles are articles dealing with

one specific area of Radiology or allied science related through the personal expe­rience of the author.

– Review articles are special articles reporting the experience of the author considered in

the general perspective of the literature over the topic.

– Case reports are short descriptions of a particular case providing a message directly linked to an individual patient investigated.

– No more than one case should be described in detail and clinical description should be kept to a minimum. Case reports should invest the usual headings of articles but should focus on the particular radiologic procedure that contributed to the diagno­sis. References should be present, though limited in number. Tables and acknowl­edgements are usually omitted.

– Pictorial essays are articles presenting information through illustrations and leg­ends. The presentation remarks stated in the paragraph dealing with case reports apply to pictorial essays.

– Continuing education articles are designed in accordance with the general guidelines for articles published in the JBR­BTR in particular they are divided into introduc­tion, material and methods, results, discus­sion, references, and are provided with an abstract.

However, papers addressing the continuing education may have only additionnally to their contents an introduction (stating the aim of the article and providing any back­ground information useful to understand why the topic is relevant, and describing the subtopics covered by the study), refer­ences, and an abstract.

Tables should be limited to a maximum of one table per 6 pages of manuscript. Illustrations should also be limited to a maximum of one illustration (1010 cm)

(possibly made up of different parts) per 3 pages of manuscript.

All the material should be made available to the JBR­ BTR editorial office (2 copies of the manuscript with 2 sets of illustrations) with the corresponding diskette though there will not be peer review.

– Images in Clinical Radiology are short (max. 1 typed page) case reports designed to illustrate with max. 3 figures a specific enti­ty. The report should not include abstract nor discussion but consist of a synthetic description of the clinical and radiological features as well as the final diagnosis and one major reference. Technical notes are short descriptions of a specific technique, procedure or equipment of interest to radi­ologists. Technical notes may originate from radiologists having experience of the item presented or from commercial firms (these should contact the Editorial Office to obtain specific guidelines for publication). The manuscript length should be inferior to 1 typed page, original language should be English, the manuscript may be accompa­nied by maxi mum 1 b/w figure, and include one major reference.

– Book reviews should be limited to one typed page, mention full references of the book, including number of pages, of illus­trations (when available), and price. The author should specify to whom the book is intended and give a personal apprecia­tion. They will be publish ed with the initial letters of the signature.

Editor-in-Chief: J. PringotConsulting Co-Editor: Ph. Grenier (Paris)Managing Editor: P. SeynaeveWebmaster: P. Vanhoenacker

Editorial Board: F. Avni, L. Breysem, N. Buls, B. Coulier, B. Daenen, E. Danse, H. Degryse, P. Demaerel, B. Ghaye, J. Gielen, P. Habibollahi, N. Hottat, M. Laureys, F. Lecouvet, M. Lemmerling, B. Lubicz, J.F. Monville, T. Mulkens, A. Nchimi, J.F. Nisolle, B. Op de Beeck, R. Oyen, S. Pans, V.P. Parashar (USA), P. Parizel, P. Peene, H. Rigauts, N. Sadeghi, P. Simoni, S. Sintzoff Jr, A. Snoeckx, J. Struyven, H. Thierens, P. Van Dyck, F. Vanhoenacker, Ph. Van Hover, J. Verschakelen, K. Verstraete.

Board of the Belgian Society of Radiology (BSR):

President: G. Villeirs

Vice-President: D. Henroteaux

Secretary: Ch. Delcour

Treasurer: P. Vanhoenacker

BVAS/ABSYM-Representative: O. Ghekiere

Board Members: P. Aerts, B. De Foer, J.­F. De Wispelaere, M. Grieten, H. Jaspers, J.­P. Joris, R. Oyen, P. Seynaeve, G. Souverijns, D. Tack, B. Vande Berg, G. Vandenbosch, Ch. Van de Velde

Scientific Committee

President: R. Oyen

Secretaries: J. de Mey, B. Vande Berg

Members: R. Achten, D. Bielen, B. De Foer, Y. Lefebvre, M. Lemort, J. Pringot, R. Salgado, D. Tack, J. Ver­schakelen

– Opinion articles are special articles dealing with controversial topics of specific concern to radiologists. They may include tables and figu res, and must provide a references list.

– Letters to the Editor and their replies present objective and useful criticism over an article published in one of the lest four issues of the JBR-BTR. They will be published with the name and address of the author. References are necessary, tables and figures are accepted but acknowledgements are not appropriate.

– Meeting news are reports of national or international congresses, symposia and meetings of radiology. Full references of the meeting, including date, place and summary of the main topics should be mentioned. Text should be kept to major facts. Figures, tables, refe rences and ac knowledgements should not be included.

– In memoriams and News are essentially dealt with in the Editorial Office. Con-tributions may however be submitted under the form of letters addressed to the Editorial Office which will check the adequacy of the information.

General Guidelinesfor Papers

Manuscript Requirements

Send 3 copies of the manuscript, including tables and figures (1 original set + 2 copies of the text and 2 original sets + 1 copy of the illus-trations) and the corresponding diskette (see below Instructions for Electronic Manuscript Submission). In keeping with sound environ-mental and economic principles, the JBR-BTR encourages all authors to submit manuscripts printed on both sides of the page. The practice not only will save paper but also will reduce the price of postage required to mail the manu-script. Note that failure to provide an electronic version of manuscript will result in costs to be charged to the author. The original set should mention the personal references of the author. The copies should be nameless (including the figures). Each section of the manuscript begins on a new page in the following order: titre page running title page + key-words, abstract, text, acknowledgements, references, tables and captions for illustrations. Use English or one of the national languages. In the latter case, an English version of the titre, abstract, key-words, legends must neces sarily be provided. Note that the author will be charged the costs of translation.Submitted manuscripts may not be covered by a previous copyright. The author will be held responsible for any litigation that might pos-sebly ensue. Manuscripts will be submitted to a review Committee whose decision is final. Authors are usually notified within eight weeks as to the acceptability of their paper.

Instructions for Electronic Manuscript Submission

Please send an electronic version of your manuscript either a floppy disk or a CD-rom in conjunction with the traditional paper version or separately as an e-mail with attachments to JBR-BTR@skynet.be. Please follow the general instructions on style/ arrangements and, in particular, the reference style as given in the present “Instructions to Authors”.Note, however, that while the paper version of the manuscript must be presented in the traditional double spaced format, the electronic version will be typeset and should not contain any extraneous instructions.For exemple: use hard carriage returns only at the end of paragraphs and display lines (e.g. titles, subheadings); do not use an extra hard return between paragraphs; do not use tabs or extra space at the start of a paragraph or for list entries; do not indent runover lines in refer ences; turn off line spacing; turn off

hyphenation and justification; do not specify pages breaks, page numbers, or headers; do not specify typeface. Care should be taken to correctly enter “one“ (1) and longer case “el“ (l), as well as “zero” (O) and capital “o“ (O).Illustrations and tables will be handled conven-tionally. However, figures and table legends should be included at the end of the electronic file. Nonstandard characters (Greek letters, mathematical symbols, etc.) should be coded consistently throughout the text. Please make a list of such characters and provide a listing of the codes used.Note that disks and CD-roms will not be returned to authors.

Title

– Keep it short and relevant.– Title must be followed by the surname(s)

and first name(s) (for computer processing purposes, 2 initial letters only will be admit-ted) of all authors.

– The position held by the authors, their aca-demic degrees, the name of the institution to which they belong and/or from which the article originates and the name of the depart-ment Head (if required) must be indicated at the bottom of the first page.

– The titIe in the national language of the text should be noted after the key-words.

– A running title in English should be provided on a separate page.

– Two copies of a blind titre page are included, giving only the titre (without the authors names) for use in the review process.

Abstract and Key-words

Written in English exclusively, the abstract should head the manuscript and summarize the aim, the methods, results and conclusions. It should not exceed 200 words for major papers and 100 words for the other studies.No abbreviation or references are used in the abstract.Three to six key-words from the terms used in the JBR-BTR Subject Index (and/or the most recent three-year cumulative index of Radiology) should be listed.

Text

The text should be clearly divided in the following sections: introduction, material and methods, results, discussion and conclusion.Abbreviations should be defined in an explanato ry note before being used as such. The definitive text should be typed on one side only of a standard size (A4) typewriting paper, in doublespacing throughout and have at least 3 cm margins.The manuscript should not be longer than 16 typewritten pages, including references and summary for a major paper unless otherwise agreed by the Editor (one typewritten page is equivalent to approximately 250 words) and no longer than 6 typewritten pages for the other types of work.

Specific guest editorials

Specific guest editorials are invited papers writ-ten by selected distinguished specialists. They should summarize in concise the stase of the art in one specific field of medical imaging or related sciences in no more than 8 typewritten pages, including either 1 table or illustration (drawing or graph). The bibliography should not exceed 12-15 recent and/or fundamental references.

References

References should be numbered consecutively in the order in which they appear in the text. Their number should be kept down to 20 for major papers and 8 for case reports and other papers.

They should be given as follows:a) abridged titles of periodicals should conform

to those in the Index Medicus. All authors are listed when six or fewer; when seven or more authors, the first three are listed, fol-lowed by “et al.”.

Ex.: Bomsel F., Couchard M., Henry E.: Respiratory distress in the newborn. J Belge Radiol, 1980, 63: 89-107.

b) in the case of books, references should indicate: the authors of the chapter, the title of the chapter, the title of the book, the editor(s), publisher, edition, city, year and specific pages.

– Ex.: Isengrin P.: Radiologie stomacale. 3e édition, Arscia, Bruxelies, 1974, p. 22.

– Ex.: Weinstein L., Swartz M.N.: Patho-genic properties of invading micro-orga nisms. In: Pathologic physiol-ogy: mechanisms of disease. Edited by Sodeman W.A. Jr, Sodeman W.A., Cds. Printed by Saunders, Philadelphia, 1974, pp. 457-472.

Quote the name and address of the author to whom the reprints will be sent, at the end of the references.

Corresponding author and Reprints

The name and address of the correspond-ing author to should be mentioned affer the referen ces.25 reprints, are offered free by the JBR-BTR.

Tables

Tables should be presented on a separate page and numbered in Roman numerals in the order in which they are cited in the text. They should have an English title and legend. Abbreviations should be defined in a foot note.Only commonly admitted measurement stand-ards should be used.

Figures and Legends

Illustrations should be restricted to the mini-mum required to show the essentiel features describ ed in the paper. They must be men-tioned in the text.Two complete unmounted sets of original fig-ures in labeled envelopes should be provided. All figu re parts relating to one patient should have the same figure number. Use capital let-ters A, B, C, in the ieft longer corner to distin-guish figures from one set.Figures should be marked on the back with an arabic numeral indicating the sequence in which they are to be referred to, with a lightly pencilled “top“ indicating their topside and the name of the first author. Never use ink on front or back of any figure. For uniformity purposes, points of interest should be showed on the figures with removable (Letraset) arrows or/and letters, or should be indicated on an accompanying photocopy of the figures, in order to enable our services to use their own characters. Images should be uniform in size and magnification.1. RadiographsCost and number: depending on the length of the manuscript (a total of 2 to 6 times 14 3 15 cm is availabie free of charge).Presentation: glossy prints, no larger than 18/24 cm.It is advisable for films to be centered on the zone of major interest and they should be grouped. Arrows should indicate the important points.2. Photographs and drawingsFour-colour illustrations will be printed at the expense of the authors.Drawing and graphs should be of professional quality. They should illustrate — not duplicate — data given in the text.Legends are typed separately and preceded by the number of the corresponding illustration.Note that illustrations will not be returned to authors.

(*) Pr J. PRINGOT, Avenue W. Churchill 11/30, B-1180 Bruxelles, Belgique (tél.: 02-374.25.55, fax: 02-374.96.28, e-mail: JBR-BTR@skynet.be).

JBR-BTR 98/2 2015

Journal Belge de Belgisch Tijdschrift voor RADIOLOGIE

Founded in 1907

A bimonthly journal devoted to diagnostic and interventional imaging,related imaging sciences, and continuing education

published by the Belgian Society of Radiology

Contents

Can diffusion weighted MRI differentiate between inflammatory-infectious and malignant pleural effusions?O. Karatag, T. Alar, S. Kosar, G. Ocakoglu, Y. Yildiz, E. Gedik, U. Gonlugur, H. Ozdemir . . . . . . . . . . . . . . . . . . 63

Assessment of normal values of fractional anisotropy and mean diffusivity of mobile lumbar spine roots by diffuson tensor MRI: comparison between 1.5 and 3T.

J. Lincot, J.P. Laissy, A. Hess, V. Balbi, E. Schouman-Claeys, A. Cotton, B. Dallaudière . . . . . . . . . . . . . . . . . . . 68Lunotriquetral coalition, a normal variant that may rarely cause ulnar sided wrist pain

M. Mespreuves, F. Vanhoenacker, K. Verstraete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72A rare malignant hepatic tumor of childhood: transitional liver cell tumor revisited

O. H. Nursun, O. Berna, S. Tutku, T. Beril, H. Mithat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79Erdheim Chester disease presented isolated breast and axillary involvement

I. Basara, E. Yavuz, P. Balci, E. B. Tuna, I. Sari . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82Cysts in a Brunner’s gland hamartoma: a clue to diagnosis

Y-K. Fan, Y-P. Liu, Y-L. Lin, W-K. Su . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85Bronchial lipoma : an unusual cause of pleural empyema

S. Lanotte, P. Mailleux, R. Frognier, O. Van Cutsem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

IMAGES IN CLINICAL RADIOLOGY

‘Backfill’ of the sacroiliac joint space in spondlyloarthritisF. Laloo, N. Herregods, H. Cypers, K. Verstraete, L. Jans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

Rare presentation of Langerhans cell histiocytosisA. Gieraerts, P. Vandaele, R. Schildermans, L. Daveloose, K. Ramboer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

A glomus tumorP. Kulczycka, B. Dallaudière, O. Barbier, B. Vande Berg, A. Larbi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

Dirty shadowing in emphysematous pyelonephritisA. Eeckhoudt, F.M. Vanhoenacker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

MDCT findings of polymicrobial descending necrotizing mediastinitisC. Karanikas, P. Lampropoulou, D. Karakiklas, C.S. Baltas, A. Demertzis, C. Drosos . . . . . . . . . . . . . . . . . . . . . . 95

Tentorium hypoplasia with partial occipital lobe herniation E. Thomaere, S. Schepers, B. Termote, R. Vanwyck, G. Souverijns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

320-row-detector CT angiography findings of a case with myocardial bridging in the three main coronary arteries.

S. Akay, U. Bozlar, Dermikol Sait, M. Tasar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97The single right coronary artery.

E. Christiaanse, D. Verdries, K. Tanaka, S. Carlier, J. de Mey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98Symptomatic coracoclavicular joint

S. Lanotte, S. Van Den Broeck . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

LETTER TO THE EDITOR

A case of symptomatic mesenteric panniculitis presenting with unusual positive FDG PET/CT nodular components: an atypical imaging strategy with histopathologic correlation.

B. Coulier, I. Bueres, F.C Deprez, F. Richelle, R. Rubay, I. Gielen, C. Fervaille . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

Charter Young Radiologists section (YRS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

u The terms used for indexation of subjects were developed by the Radiological Society of North America (RSNA) over a period of years. Their use here is by permission of the RSNA. The terms may not be used in any other index, print or electronic, except by specific permission of RSNA.

uu Indexed in Index Medicus and in Zentralblatt Radiologie. Evaluated for Medline User, EMBASE and CANCERNET. Abstracted in Excerpta Medica Journals.

Pleural effusions are usually diag-nosed on the basis of clinical, radio-logical and pathological findings. The main types of pleural effusions are defined as transudate and exu-date according to their biochemical features, mostly separated by Light’s criteria (measurement of lactate de-hydrogenase (LDH) and protein con-centrations in both pleural fluid and serum) (1). However, transudate and exudate classification of pleural effu-sions with Light’s criteria can occa-sionally fail, especially in cases that have undergone diuresis (2, 3). If the pleural effusion is definitely identi-fied as a transudate, no further diag-nostic procedures are needed (1, 4). If it is found to be an exudate, addi-tional diagnostic procedures such as pleural fluid cytopathology, Gram staining, culture, etc. are neces-sary (2, 5). Differentiation of transu-date and exudate before thoracente-sis therefore becomes mandatory in order to avoid unnecessary interven-tional procedures and their potential complications.

Exudative effusions can have benign (bacterial pneumonia, viral infection, pulmonary embolism, etc.) and malignant (lung cancer, breast cancer, lymphoma, etc.) causes. We believe that there is a need for an

sion coefficient (ADC) without be-ing affected by the partial volume effect and motion artifact. The thickness of the pleural effusions detected on chest x-ray was eval-uated by US. The thickness of the pleural effusions detected on CT examination was measured. Pleural effusions less than 1 cm in thickness were excluded from the study.

2. Cases where pleural effusion as-piration and analysis were planned were included in the study.

3. Cases who were severely dys-pneic, and would not be able to lie in the supine position during the entire diffusion-weighted im-aging (DWI) examination or had claustrophobia were excluded from the study.

All patients were examined with a 1.5 T MRI scanner (GE Signa, HDxt). The maximum gradient power of this system was 30 mT/m and the slew rate was 150 mT/m/msec. T1- and T2-weighted conventional im-ages on the axial plane were ob-tained to evaluate the anatomical detail better and to determine the pleural effusion localization and sig-nal characteristics. Subsequently, 4 series of single-shot spin-echo echo-planar (SS-SE-EPI) diffusion-weighted (DW) images (TR 4050 ms, TE 78.2 ms, field of view 38 × 30.4 cm2, matrix size 128 × 160, slice thickness 8 mm, interslice gap 2 mm) were obtained using 4 different b values (10, 500, 750 and 1000 s/mm2) on the axial plane.

alternative non-invasive diagnostic method for inflammatory-infectious and malignant differentiation of exu-dative pleural effusions.

In this study, we aimed to assess exudative pleural effusions with dif-fusion-weighted magnetic reso-nance imaging (DW-MRI) in order to determine differentiation criteria for inflammatory-infectious and malig-nant processes.

Materials and methods

This prospective study was ap-proved by the institutional review board and written informed consent was provided by all subjects.

A total of 43 patients who present-ed to Canakkale Onsekiz Mart Uni-versity Research and Application Hospital between January 2011 and December 2012 with various symp-toms and in whom the chest x-ray, ultrasound (US) and computed to-mography (CT) examinations re-vealed pleural effusion were includ-ed in the study. The study inclusion and exclusion criteria were as fol-lows:

1. The amount of pleural effusions had to be adequate for accurate measurement of apparent diffu-

JBR–BTR, 2015, 98: 63-67.

CAN DIFFUSION WEIGHTED MAGNETIC RESONANCE IMAGING DIFFERENTIATE BETWEEN INFLAMMATORY-INFECTIOUS AND MALIGNANT PLEURAL EFFUSIONS?O. Karatag1, T. Alar2, S. Kosar1, G. Ocakoglu4, Y. Yildiz1, E. Gedik2, U. Gonlugur3, H. Ozdemir1

Aim: To assess exudative pleural effusions with diffusion-weighted magnetic resonance imaging (DW-MRI) in order to determine non-invasive differentiation criteria for inflammatory-infectious and malignant effusions.Materials and methods: Thirty-two patients with pleural effusions underwent DW-MRI with 4 different b values (10, 500, 750 and 1000 s/mm2). ADC maps were generated automatically. Signal intensity and ADC values were mea-sured. Following MRI, pleural fluid of 10-15 ml was obtained and analyzed. AUC values were compared for different diffusion levels of ADC and SI measurements. The relationship between ADC values and pleural effusion LDH and total protein levels was examined. Results: The cut-off values obtained from signal intensity and ADC measurements to differentiate exudates with malignant pathology were not found to be statistically significant. In the inflammatory-infectious group, a significant negative correlation was observed between ADC values and pleural fluid LDH measurements in all b values. In the malignant group, a significant positive correlation was observed between ADC values and pleural fluid total protein measurements in b values of 500 and 1000.Conclusion: Infectious/inflammatory and malignant effusions overlap strongly and cannot therefore be differentiated using DW MRI.

Key-words: Lung, effusion – Lung, MR – Magnetic resonance (MR), diffusion study.

From: Department of 1. Radiology, 2. Thoracic Surgery, 3. Chest Diseases, Canakkale Onsekiz Mart University, School of Medicine, 4. Dpt of Biostatistics, Uludag University, Shool of Medicine, Turkey.Address for correspondence: Dr O. Karatag, MD, Canakkale Onsekiz Mart University School of Medicine, 17100 Canakkale, Turkey. E-mail: okaratag17@gmail.com

64 JBR–BTR, 2015, 98 (2)

statistical power: 0.80). Statistical significance was set at p < 0.05. Med-Calc v.12.7.5.0 were used for statisti-cal analyses.

Results

Seven patients could not keep still during MR imaging and the obtained images were of poor quality, so they were excluded from the study. Of the remaining 36 patients, 4 patients who were found to have transuda-tive pleural effusions were also ex-cluded. Finally we included 6 females and 26 males for a total of 32 pa-tients. The median age was 63 (min: 26, max: 88) years. The causes for exudative pleural effusion are given in table I.

We evaluated DWI in terms of inflammatory-infectious/malignant differentiation of exudative pleural effusions and found that the SI and ADC values measured with different b values were not appropriate for in-flammatory-infectious/malignant dif-ferentiation of the exudates. After ROC Analysis, the cut-off values ob-tained from SI and ADC measure-ments to differentiate exudates with malignant pathology were not found to be statistically significant. No difference was found between ROC curves obtained with different b val-ues for SI and ADC. There was no significant difference between the AUC in the analyses performed for each b value. The p value was > 0.05 for all analyses in both tables (Table II).

When inflammatory-infectious and malignant groups are analysed separately, in the former group, a significant negative correlation was observed between ADC values and pleural fluid LDH measurements in all b values of 10 (r = –0.46, p = 0.021), 500 (r = –0.76, p < 0.001), 750 (r = –0.70, p < 0.001), and 1000 (r = –0.59, p = 0.002) (Fig. 2) and no statistically significant correlation was found between ADC values and pleural fluid total protein measure-ments.

In the malignant group, a signifi-cant positive correlation was ob-served between ADC values and pleural fluid total protein measure-ments in b values of 500 (r = 0.67, p = 0.012) and 750 (r = 0.73, p = 0.005) (Fig. 3) and no statistically significant correlation was found between ADC values and pleural fluid LDH mea-surements. The correlation analysis results of ADC values with LDH and total protein levels of pleural effu-sion in both groups are given in table III.

Thoracentesis was performed by a thoracic surgeon after MRI. Pleural fluid of 10-15 ml was obtained and analyzed. The pleural fluid was pri-marily classified as transudate or exudate according to Light’s crite-ria (6). The pleural fluid sample was then microbiologically and patholog-ically analyzed. Whether the fluid was of benign or malignant origin was determined.

Statistics

Data was presented as median (minimum-maximum) values. ROC (receiver operating characteristic) analysis was used for obtaining cut-off values for inflammatory-infec-tious/malignant differentiation and comparing the AUC (Area under the curve) values for different diffusion levels of ADC and SI measurements. The relationship between ADC val-ues and pleural effusion LDH and to-tal protein levels was examined by performing correlation analysis and Spearman correlation coefficient was computed. The minimum sam-ple size required for each group was determined at least n = 30 (desired

Unidirectional diffusion gradients (readout direction; right-left, R-L) were applied. ADC maps were gen-erated automatically from the DW images obtained.

Quantitative analyses were per-formed on a dedicated workstation (General Electric, Advantage work-station, 4.4 edition). An average of 203 mm2 elliptical or spherical re-gions of interest (ROI) were placed in 3 different locations of the pleural effusion in order to measure signal intensity values (SI) on DW images and the ADC values on ADC maps (Fig. 1). Both values were averaged separately and the resulting values were considered the final quantita-tive values. ROI of the same dimen-sions was used for each b factor. In order not to be affected from possi-ble magnetic susceptibility and motion artifacts, we stayed away from the lung-pleural fluid sections, diaphragmatic areas and the regions where the heart and major vascular structures were present during ROI placement and measurement, placing the ROI’s at the most homo-geneous segments possible of the pleural effusions.

Fig. 1. — A fifty-six year old male diagnosed with para-pneumonic effusion. Axial T2W image shows a 53 mm wide left pleural effusion with high signal intensity (A). Three different ROI placement for SI measurement on DW images obtained with a b-value of 10 (B), 500 (C), 750 (D) and 1000 (E) and ADC measurement on corresponding ADC map (F) are shown.

A

C

E F

D

B

DIFFUSION WEIGHTED MRI OF PLEURAL EFFUSIONS — KARATAG et al 65

Fig. 2. — Fifty-five (A, B and C) and seventy-four (D) year old male patients diagnosed with empyema. The former patient who has higher pleural effusion LDH level (2388 U/L), demon-strates lower ADC measurement (0.002757) compared with the latter patient who has a LDH level of 151 U/L and mean ADC measurement of 0.004057.

Table I. — Causes of pleural effusions.

EtiologyBenign (n = 19)Parapneumonic effusion 12Empyema 4Tuberculosis 1Sarcoidosis 1Ruptured hydatid cyst 1Malignant (n = 13)Non-small cell lung carcinoma 3Small cell lung carcinoma 2Adenocarcinoma 2Malignant mesothelioma 2Deciduoid mesothelioma 1Breast carcinoma metastasis 1 Non-specific primary 2

Total n = 32

A

C

B

D

Table II. — ROC analysis results of SI and ADC values calculated for each b factor for differentiation of inflammatory-infectious / malignant exudative pleural effusions.

SI AUC SE p-valueb = 10 0.60 0.11 0.319b = 500 0.55 0.10 0.625b = 750 0.54 0.10 0.710b = 1000 0.54 0.10 0.712Comparison of ROC curves between b factor levels

DBA SE p-valueb = 10-b = 500 0.06 0.10 0.580b = 10-b = 750 0.07 0.11 0.530b = 10-b = 1000 0.07 0.11 0.534b = 500-b = 750 0.01 0.05 0.812b = 500-b = 1000 0.01 0.04 0.800b = 750-b = 1000 0 0.02 1.00

ADC AUC SE p-valueb = 10 0.65 0.10 0.131b = 500 0.55 0.11 0.667b = 750 0.55 0.11 0.675Comparison of ROC curves between b factor levels

DBA SE p-valueb = 10-b = 500 0.10 0.13 0.425b = 10-b = 750 0.10 0.13 0.411b = 10-b = 1000 0.05 0.15 0.729b = 500-b = 750 0.01 0.07 0.977b = 500-b = 1000 0.05 0.11 0.662b = 750-b = 1000 0.05 0.09 0.577

AUC: Area under ROC curve, DBA: Difference between ROC areas.

66 JBR–BTR, 2015, 98 (2)

the transudative ones. The reason was thought to be the high protein-aceous and cellular content of pleu-ral fluid in exudative effusions lead-ing to a decrease in ADC values. DWI was performed with b = 0 s/mm2 and b = 1000 s/mm2 values and the mean ADC values were calculated as 3.42 ± 0.76 × 10-3 mm2/s and 3.18 ± 1.82 × 10-3 mm2/s in transudative and exudative effusions by Baysal et al. (1). DWI was performed with b = 0 s/mm2, b = 500 s/mm2 and b = 1000 s/mm2 values and the ADC cut-off value was found to be 3.6 × 10-3 mm2/s for the transudative/exudative pleural effu-sion differentiation with DWI by Inan et al. (2).

We aimed to perform the inflam-matory-infectious/malignant differ-entiation of exudative pleural effu-sion with DWI in our study. Our main objective was therefore different from the two previous studies. The SI and ADC measurements we con-ducted by using 4 different b values (b = 10, 500, 750 and 1000) did not lead to statistically significant results for inflammatory-infectious/malig-nant differentiation in exudates.

In the inflammatory-infectious group, a significant negative correla-tion was observed between ADC val-ues and pleural fluid LDH measure-ments in all b values. This result was in accordance with Baysal et al’s study supporting the higher the vis-cosity of the pleural fluid, the smaller is the diffusion (1). On the other hand, in the malignant group, a sig-nificant positive correlation was ob-served between ADC values and pleural fluid total protein measure-ments in b values of 500 and 1000. This result can be explained by de-crease in serum protein levels rela-tive to pleural fluid (hypoalbumin-emia). Due to this relative decrease of proteins, an increase in diffusion of water molecules may develop leading to increase in ADC values (1).

In another study, Coolen et al. in-vestigated combination of DWI and dynamic contrast enhanced MRI

and phased-array coils (9, 10). How-ever, application to thoracic imaging is still difficult since DWI is very sen-sitive to artifacts (7).

Stimulated-echo and SE pulse sequences that require a very long acquisition time and are therefore very sensitive to motion artifacts were used in the first DWI applica-tions (11). Today, SS-SE-EPI sequence that is not markedly affected by mo-tion artifacts is used (7).

The differentiation of whether the effusion is a transudate or exudate has been mainly attempted in stud-ies on the detection of the nature of the pleural effusion by DWI (1, 2). When DWI was applied with the ap-propriate b values, the ADC values were found to be significantly lower in exudative pleural effusions than

Discussion

DWI is an MR imaging method providing tissue analysis based on the diffusion of water molecules in-side the tissue (7). Carr and Purcell were the first to report that the MRI signal is affected by the diffusion of water molecules in 1954 (8). DWI was first used to determine acute ce-rebral ischemia and then in other body regions in recent years thanks to developing technologies (7, 9, 10). A high image quality could not be obtained in areas outside the brain in the first studies conducted with DWI and the ADC quantification was not very successful. These problems have later been overcome by the introduction of stronger gradient fields, parallel imaging techniques

Table III. — The correlation analysis between ADC values and pleural fluid LDH and total protein levels in all b values.

ADC value Benign Malignantb = 10 b = 500 b = 750 b = 1000 b = 10 b = 500 b = 750 b = 1000

LDH U/Lr -0,46 -0,76 -0,70 -0,59 r -0,02 -0,21 -0,25 -0,34p 0,021* <0,001* <0,001* 0,002* p 0,943 0,494 0,409 0,263

TOTAL PROTEIN r -0,05 -0,09 -0,13 0,13 r 0,13 0,67 0,73 0,28p 0,817 0,678 0,532 0,551 p 0,674 0,012* 0,005* 0,353

*Significant correlation.

A

C

B

DFig. 3. — Fifty-nine (A, B and C) and seventy-nine (D) year-old

male patients diagnosed with non-small cell lung carcinoma. The former patient who has higher pleural effusion total protein level (7 gr/dl), demonstrates higher ADC measurement (0.004913) compared with the latter patient who has a total pro-tein level of 2.5 gr/dl and mean ADC measurement of 0.00197.

DIFFUSION WEIGHTED MRI OF PLEURAL EFFUSIONS — KARATAG et al 67

3. Romero-Candeira S., Fernández C., Martín C., Sánchez-Paya J., Hernández L.: Influence of diuretics on the concentration of proteins and other components of pleural transudates in patients with heart failure. Am J Med, 2001, 110: 681-686.

4. Romero-Candeira S., Hernández L., Romero-Brufao S., Orts D., Fernández C., Martín C.: Is it mean-ingful to use biochemical parameters to discriminate between transudative and exudative pleural effusions? Chest, 2002, 122: 1524-1529.

5. Tarn A.C., Lapworth R.: BTS guidelines for investigation of unil-ateral pleural effusion in adults. Thorax, 2004, 59: 358-359.

6. Light R.W.: Management of pleural effusions. J Formos Med Assoc, 2000, 99: 523-531.

7. Luna A., Sánchez-Gonzalez J., Caro P.: Diffusion-weighted imaging of the chest. Magn Reson Imaging Clin N Am, 2011, 19: 69-94.

8. Carr H.Y., Purcell E.M.: Effects of diffusion on free precession in nuclear magnetic resonance. Phys Rev, 1954, 94: 630-638.

9. Guo Y., Cai Y.Q., Cai Z.L., et al.: Dif-ferentiation of clinically benign and malignant breast lesions using diffu-sion-weighted imaging. J Magn Re-son Imaging, 2002 , 16: 172-178.

10. Wang J., Takashima S., Takayama F., et al.: Head and neck lesions: characterization with diffusion-weighted echo-planar MR imaging. Radiology, 2001, 220: 621-630.

11. Le Bihan D., Breton E., Lallemand D., Grenier P., Cabanis E., Laval-Jeantet M.: MR imaging of intravoxel in-coherent motions: application to diffusion and perfusion in neurologic disorders. Radiology, 1986, 161: 401-407.

12. Coolen J., De Keyzer F., Nafteux P., et al.: Malignant pleural disease: diag-nosis by using diffusion-weighted and dynamic contrast-enhanced MR imaging – initial experience. Radiolo-gy, 2012, 263: 884-892.

13. Ozsunar Y., Sorensen A.G.: Diffusion- and perfusion-weighted magnetic resonance imaging in human acute ischemic stroke: technical con-siderations. Top Magn Reson Imaging, 2000, 11: 259-272.

14. Naganawa S., Kawai H., Fukatsu H., et al.: Diffusion-weighted imaging of the liver: technical challenges and prospects for the future. Magn Reson Med Sci, 2005, 4: 175-186.

15. Mürtz P., Flacke S., Träber F., van den Brink J.S., Gieseke J., Schild H.H.: Abdomen: diffusion-weighted MR imaging with pulse-triggered single-shot sequences. Radiology, 2002, 224: 258-264.

dient in order to avoid respiratory distress and to ensure that patients could stay motionless during the ex-amination so that we could obtain better image quality. However, at least three orthogonal directions should be used in DWI imaging. Mo-lecular diffusion is a tridirectional process and diffusion can be aniso-tropic as in the cerebral white matter (13). We believe that data obtained by applying tridirectional diffusion gradient will provide statis-tically healthier results. Another limi-tation was that the DW images we obtained with the EPI sequence had low signal-to-noise ratio especially when a high b value was used and this caused image distortion. In addi-tion, the EPI sequence can also cause anatomic distortion with its suscepti-bility effect (14). We did not perform pulse-triggered DWI in our study. Mürtz et al used the SS-SE-EPI se-quence with ECG triggering in their study to avoid the negative effect of cardiac pulsation and emphasized that the accuracy of ADC measure-ments was reduced in DWI without pulse-triggering (15). We believe that the ADC values obtained by DWI using the pulse-triggering technique will provide more accurate results.

Conclusion

DWI did not provide sufficiently accurate results in inflammatory-in-fectious/malignant differentiation of the exudative effusions in our study. Prospective studies with larger se-ries are required to support the effi-cacy of DWI which can be performed within a short time and also can eas-ily be added to routine thoracic MRI examination for transudate/exudate differentiation of pleural fluids and inflammatory-infectious/malignant differentiation of exudates.

References

1. Baysal T., Bulut T., Gökirmak M., Kal-kan S., Dusak A., Dogan M.: Diffusion-weighted MR imaging of pleural fluid: differentiation of transudative vs exu-dative pleural effusions. Eur Radiol, 2004, 14: 890-896.

2. Inan N., Arslan A., Akansel G., Arslan Z., Eleman L., Demirci A.: Diffusion-weighted MRI in the charac-terization of pleural effusions. Diagn Interv Radiol, 2009, 15: 13-18.

(DCE-MRI) for differentiation of be-nign and malignant lesions of pleu-ra. They made the ADC measure-ments of the pleural lesions taking care to exclude necrotic areas and considering the most solid parts for ROI placement. Similar to Baysal et al’s and Inan et al’s study, they calcu-lated ADC values of malignant pleu-ral diseases significantly lower than benign lesions (1.40 × 10-3 mm2/s ± 0.33 and 2.49 × 10-3 mm2/s ± 0.81, re-spectively). They stated that this re-sult was most likely because of hy-percellularity and hypervascularity of malignant lesions causing diffu-sion restriction. They concluded that DWI can differentiate malignant pleural diseases from benign lesions with high accuracy especially with the combination of DCE-MRI. They also concluded that by using a pure-ly ADC-based diagnosis, false-nega-tive diagnoses may occur mostly due to necrotic and inflammatory area within the tumor (12).

In our study, we aimed to assess only the pleural effusions with DWI and we disregarded the solid parts of the malignant pleural lesions in order to make a truer comparison between the ADC values of inflam-matory-infectious and malignant ef-fusions. Because we were not be able to assess the cellularity of the effusions due to technical limita-tions, we could not give support to Coolen et al. (12) regarding their as-sumption about the higher the cellu-larity of the lesion, the smaller is the diffusion.

The main limitation in our study was the inadequate number of cas-es. We believe that much larger number of patients should be includ-ed in the study to perform inflam-matory-infectious/malignant differ-en tiation of the exudative qualified effusions with DWI first and then to perform subgrouping within the ma-lignant processes afterwards, and that perhaps more statistically sig-nificant results can be obtained with a large case series. Another limita-tion was our application of unidirec-tional diffusion gradient in DWI. The patients in our study had marked pleural effusions and therefore many suffered respiratory distress, espe-cially in the supine position. We tried to keep the examination short by ap-plying a unidirectional diffusion gra-

Diffusion tensor imaging (DTI) can provide non-invasive, quantitative data to evaluate neural pathways in the central and peripheral nervous systems in vivo (1-5). This technique explores the anisotropic microscopic Brownian motions of water mole-cules along the preferential direction of fibers (6, 7). In each voxel, the dif-fusion tensor allows the calculation of eigenvalues, which are used to characterize the anisotropy, as re-flected by fractional anisotropy (FA), and the mean diffusivity (MD, aver-age of the 3 eigenvalues) of the mid-dle. The degree of anisotropy and the average diffusion lead to the determination of the main diffusion direction in each voxel of the explored tissue (8), reflective of the orientation of the tissular compo-nents, e.g. axonal fibers.

DTI has been mostly used in neu-roradiology in order to study neural connectivity in white matter. It is a tool to approach microstructural networks, capable to provide a three-dimensional visualization tool of nerves and muscles fibers (4).

The feasibility of DTI and tracto-graphy of human peripheral nerves (9, 10) was recently demon-strated in the carpal and ulnar tunnel syndromes (9-11). In lumbar spine,

undergone by compressed nerve roots. However, only few studies (1, 12, 13) established normal values of FA and MD of lumbar nerve roots at 1.5T MRI and none at 3T MRI, ac-cording to demographic data (14). More, inter individual variations of normal diffusion parameters may exist, as well as physiological varia-tions according to the level and the portion of the considered nerve root.

Because of the growing impor-tance of DTI in lumbar imaging and its clinical implications, we are inter-ested in determing normative diffu-sion tensor parameters and to as-sess whether these normative findings differ according to the mag-netic field strength.

Consequently, the purposes of our study were to confirm the feasi-bility of the DTI technique in the exploration of nerve roots of the mobile lumbar spine and to define normal values of FA and MD in healthy subjects at 1.5T and to compare them with results at 3T.

Material and methods

Patients

Thirty-seven volunteers without previous clinical history of lumbalgia or lumbar radiculalgia (BD, JL) were included prospectively in our single center study from April 2011 to January 2012. Written inform con-sent was obtained from each subject before inclusion.

Exclusion criteria were a previous history of spinal trauma, surgery, or neurological disease and contraindi-cation to MRI (pregnancy, metallic

few studies reported fiber tracking of the nerve roots. Compression of lumbar nerve root consecutive to disc herniation has been associated to modifications of the diffusion pa-rameters, namely FA and MD (12). DTI fiber tracking may reflect histo-logical changes in the nerve root tis-sue secondary to the compression, independently of a patent discora-dicular conflict seen on MRI. It may then be used as an additional diag-nostic tool in clinical routine, particu-larly in case of discordance between anatomical MRI and clinical symp-toms. Indeed, increase in the vascu-lar permeability with disruption of the nerve root barrier, intraneural edema, intra and perineural hyper-aemia have been attributed to chron-ic compression of the nerve roots and may explain modifications of water diffusion along the nerve root (11-13). Moreover, ischemia, de-myelination and Wallerian degener-ation may reduce anisotropy by in-creasing the distance between axons fascicles, thus leading to a decrease in the FA value, as well as an in-crease in that of MD. Thereby, DTI evaluation of lumbar nerve roots may stand as a new imaging ap-proach with more functional assess-ment of the microstructural changes

JBR–BTR, 2015, 98: 68-71.

AssessMent of norMAl vAlues of frActionAl Anisotropy And MeAn diffusivity of Mobile luMbAr spine nerve roots by diffusion tensor Mr iMAging: coMpArison between 1.5 And 3tJ. Lincot1, J.P. Laissy1,2, A. Hess1,2, V. Balbi3,4, E. Schouman-Claeys1,2, A. Cotten3,4, B. Dallaudière1,2,5

Purpose: to assess the normal values of fractional anisotropy (fA) and mean diffusivity (Md) of l4, l5 and s1 nerve roots using diffusion tensor imaging (dti) in healthy volunteers.Materials and methods: 37 subjects without previous history of lumbalgia or radiculalgia were prospectively examined: 27 at 1.5t and 10 at 3t Mri. the protocol included standard anatomical sequences and a dti acquisition. nerve root fibers were semi automatically extracted from dti tractography. fA and Md values were measured at 4 key portions along each l4, l5 and s1 nerve roots. Results: At 1.5t Mri, fA and Md were 0.221 ± 0.011 and 460.9 ± 35.5 mm2.s-1 respectively; at 3t Mri, fA and Md were 0.216 ± 0.01 and 480.1 ± 36.1 mm2.s-1 respectively, which may be considered as normal values for mobile lumbar spine nerve roots, independently of intersomatic space level (p = 0.06) and nerve root portion (p = 0.08) or magnetic field (p = 0.06).Conclusion: normal fA and Md values can be measured along lumbar mobile spine nerve roots in healthy subjects. these values were not dependent on intersomatic space level, side or anatomical portion of the nerve root or magnetic field.

Key words: Magnetic resonance (Mr), diffusion study – spine, Mr.

From: 1. Service de Radiologie, CHU Bichat – Claude Bernard, Paris, France, 2. Faculté de Médecine Xavier Bichat, Université Paris 7, 3. Service de Radiologie Ostéo-Articulaire, CHRU Roger Salengro, Lille, France, 4. Faculté de Médecine, Université Lille 2, 5. Inserm U698, CHU Bichat – Claude Bernard, Paris, France.Address for correspondence: Dr J. Lincot, M.D., Dpt of Radiology, Bichat-Claude Bernard University Hospital, 46, rue Henri Huchard, F-75018, Paris, France.E-mail : julienlincot@gmail.com

NORMAL FA AND MD VALUES OF LUMBAR SPINE NERVE ROOTS — LINCOT et al 69

ramen and in the extra foraminal portion, except for L4 nerve roots in which origin part was not in the ex-ploration field on 1.5T (Fig. 2). No ROI was used.

FA and MD values were measured in both sides of L4, L5 and S1 nerve roots and compared between them.

Statistical analysis

We described FA and MD data generated by FiberViewer software as mean, median, minimal, maximal and standard deviation for continu-ous variables. Association between diffusion parameters and topogra-phy of the measure was attested us-ing non-parametric tests (Wilcoxon test).

Time data extraction was also described as mean, minimal with standard deviation.

FA and MD data were compared and analyzed to clinical data using Medcalc© v11.0 software. Statistical testing was done at the 2-tailed alpha level of 0.05.

results

Subjects

Thirty-seven subjects were in-cluded:

– 27 clinically healthy volunteers (17 men, 10 women) on 1.5T MRI

– 10 clinically healthy volunteers (5 men, 5 women) on 3T MRI

Mean age was 62 years old (range, 43-86; SD, 5.6).

Standard MRI and DTI analysis

The DTI sequence was interpreta-ble in all cases, with a good depic-tion of L4, L5 and S1 nerve roots. Fusion between DTI and axial T2-weighted images permitted a good anatomical correlation in all cases. We insured that the entire path of the root was taken into account from its emergence to its extraforaminal por-tion by MedInria and Fiber Viewer softwares in 22 patients (132 nerve roots). In 26 nerve roots (11.7%) in 13 patients, fiber tracking was dis-continuous in isolation on 1.5 T MRI and on 3 nerve roots (1.4%) in 2 pa-tients on 3T MRI, the largest bundle gap measuring 5 mm.

At 1.5T, anatomical disruption were right lateral recess L4 (n = 4), right foraminal L4 (n = 1), left lateral recess L4 (n = 2); right spinal canal L5 (n = 2),right foraminal L5 (n = 1), left lateral recess L5 (n = 4); right spi-nal canal S1 (n = 3), right extra fo-raminal S1 (n = 3), left foraminal S1

0; b value, 900 s.mm-2; motion prob-ing gradients applied in 25 non-col-linear directions; acquisition time, 7 min53.

Data analysis

All MRI scans were reviewed in consensus by 2 readers (JL and BD), with respectively 2 and 4 year experi-ence in spine imaging, blinded to clinical data. Image analysis was per-formed for each subject, immediate-ly after the acquisition for qualitative assessment and secondly for data extraction (24 days later in mean; range, 13-35 days).

A coregistration of DTI and axial T2-weighted images was systemati-cally performed to increase the ana-tomical resolution of DTI images.

A “neurography” was obtained using the diffusion volume (b value, 900 s.mm-2) which was visualized as maximum intensity projection, to evaluate neurograms, before trac-tography color maps, in order not to include obvious artifacts. Indeed, Diffusion-weighted magnetic reso-nance imaging postprocessed by maximum-intensity projection re-portedly demonstrates the nerve roots (15, 16).

Image processing was first per-formed using MedINRIA v1.9.4 soft-ware (©Sofia Antipolis, France).

The following parameters were defined for automatic fiber tracking across the whole study DTI volume: FA threshold, 0.1; minimum fiber length, 10 mm. No ROI was used to initiate the fiber tracking. L4, L5 and S1 fiber bundles were manually seg-mented on each side for all subjects. Anatomical fusion between the axial T2 sequence and the DTI reconstruc-tions was performed to allow better visualization of the different anatom-ic spaces. Once reconstructed, L4, L5 and S1 fiber bundles were manually segmented on each side (Fig. 1). We considered as being significant at least 5 fibers for each nerve root. FA color maps were displayed using the classic three-directional color code: blue for fibers running in the cepha-locaudal direction, green for those running in the anteroposterior direc-tion and red for those running right and left (12). Matching between the encoded color maps and the T2-weighted images was also manually verified. Processing with FiberView-er v1.2.3 (©University of North Caro-lina, http://www.ia.unc.edu/dev/) software permitted automatic FA and MD values measurement for each fiber bundle at the root emer-gence, in the lateral recess, in the fo-

implants, and claustrophobia).We collected clinical data includ-

ing age and gender whereas ethnic group or sports habits were not con-sidered.

MRI

The MRI scans were performed on the day of inclusion on GE systems (GE Healthcare, Milwaukee, WI) in random order: on a 1.5T unit in 27 subjects and on 3T unit in the re-maining 10 volunteers. We used a 6 elements phased array spine coil with the patient in supine position.

The standardized MRI protocol (with non-use of parallel imaging) typically included T1 weighted FSE (for 1.5T: TR, 660ms; TE, 9.5ms; num-ber of averages (NEX), 1; field of view (FOV), 380 × 380 mm; matrix, 512 × 512; slice count, 12; slice thick-ness, 4 mm; slice gap, 0.4 mm; ac-quisition time 2 min53 s; for 3T: TR, 973 ms; TE, 8ms; NEX, 1; FOV, 360 × 360 mm; matrix 512 × 512; slice count, 8, slice thickness, 3 mm; slice gap, 0.3 mm; acquisition time 3min23s) and T2 weighted TSE (for 1.5T: TR, 2960 ms; TE, 70ms; NEX, 2; 380 × 380 mm; matrix, 512 × 512; slice count, 12; slice thickness, 4 mm; slice gap, 0.4 mm; acquisition time, 3 min21 s; for 3T: TR, 3781ms, TE, 57.4ms, NEX, 1.5; FOV, 360 × 360 mm; matrix, 512 × 512; slice count, 8; slice thickness, 3 mm; slice gap, 0.3 mm, acquisition time, 3min06) sequences of the lumbar spine both in the sagit-tal plane and a T2 weighted TSE (for 1.5T: TR, 5680 ms; TE, 123 ms; FOV, 200 × 200 mm; matrix, 512 × 512; NEX, 2; slice count, 30; slice thickness, 3 mm; slice gap, 0; acqui-sition time, 3 min40 s; for 3T: TR, 3769 ms; TE, 116.7 ms; NEX, 1.5; FOV, 200 × 200 mm; matrix, 512 × 512; slice count, 30; slice thickness, 3 mm; slice gap, 0; acquisition time, 3min18) sequence in the axial plane in the last 2 mobile levels L4-L5 to L5-S1 of the lumbar spine. In addi-tion to these previous sequences, single-shot echo-planar spin-echo DTI sequence was performed in axial plane from L4-L5 to L5-S1 interso-matic spaces with the use of the fol-lowing parameters: for 1.5T: TR, 8400ms; TE, 85.1 ms; FOV, 200 × 200 mm; matrix, 256 × 256; NEX, 4; slice count, 30; slice thickness, 3 mm; slice gap, 0; b value, 900 s.mm-2; mo-tion probing gradients applied in 25 non-collinear directions; acquisition time, 9 min12 s; for 3T: TR, 4500 ms; TE, 83.5 ms; FOV, 200 × 200 mm; matrix, 256 × 256; NEX, 4; slice count, 30; slice thickness, 3 mm; slice gap,

70 JBR–BTR, 2015, 98 (2)

quences permitted a good anatomic correlation. Measurements of FA and MD according to anatomical im-aging probably contributed to the precision of our measurements (17). Determination of normal values of diffusion parameters in lumbar nerve roots may have further implications in the comprehension and the man-agement of mechanical nerve root pain due to disc herniation. Some in-teresting reports (1, 12) have already demonstrated modifications of FA and MD values in case of com-pressed nerve root, emphasizing the importance of defining reliable and reproducible normal values in the healthy population.

We acknowledge that our study has several limitations. First, we fo-cused on the last 2 intersomatic space levels since L4, L5 and S1 are the most frequently involved nerve roots in disc herniation or foraminal nerve root entrapment. In fact, ex-ploration of the other intersomatic space levels would have implied an additional DTI sequence on 1.5T with substantial increase in the acquisi-tion time. The exploration of the oth-er intersomatic space levels was lim-ited by the size of the FOV and the number of slices of the DTI sequence, fixed to optimize the spatial resolu-tion and the acquisition time.

Another important limitation concerns the small sample size of

480.1 ± 36.1 mm2.s-1 at 3T MRI, which might be considered as normal val-ues for mobile lumbar spine nerve roots, independently of MRI field. Fiber tracking and measurement of diffusion parameters was success-fully obtained in 86.9 % of the sub-jects, confirming the feasibility of DTI for lumbar nerve roots (12). In-deed, these disruptions couldn’t be corrected because when we modify and particularly when we increase the parameters of automatic fiber tracking across the whole study DTI volume, too much artifacts like para vertebral musculature fibers were identified. Hence, according to our experience, the mentioned param eters seems to be the better compromise and permits reliable measurements, with a negligible gap.

Our values of FA on volunteers are consistent with those reported in the literature of 0.218-0.219 (12). Furthermore, FA and MD were measured and compared according to the same post-processing algo-rithms, software and readout proce-dure at 1.5 and 3T.

However, to our knowledge there are only few reports of normal diffu-sion parameters values for lumbar nerve roots in the literature and these findings have to be confirmed by further studies. The systematic co registration of axial T2 and DTI se-

(n = 2) and left extra foraminal S1 (n = 4) roots.

At 3T, anatomical disruption were right lateral recess L4 (n = 1), left fo-raminal S1 (n = 2).

Fiber tracking can’t be performed after changing parameters (thresh-old and/or minimum length) whose initial tracking was discontinuous. According to our experience, FA threshold, 0.1; minimum fiber length, 10 mm are optimal parameters for lumbar nerve roots fiber tracking.

The mean values of FA and MD for all subjects were respectively: FA, 0.221 ± 0.011; MD, 460.9 ± 35.5 mm2.s-1 at 1.5T MRI; FA, 0.216 ± 0.01; MD, 480.1 ± 36.1 mm2.s-1 at 3T MRI.

Mean values of FA and MD in the 37 healthy volunteers were not significantly different according to intersomatic space level (p = 0.06), nerve root portion (p = 0.08) and MRI magnetic strength (p = 0.06).

discussion

We aimed to determine normal values of diffusion tensor parame-ters in lumbar nerve roots of asymp-tomatic volunteers without prior his-tory of low back surgery or nerve root pain. In this study, FA was 0.221 ± 0.011, MD was 460.9 ± 35.5mm2.s-1 at 1.5T MRI; and FA was 0.216 ± 0.01, MD was

Fig. 1. — 55 year-old male; image fusion of diffusion tensor tractography and T2-weighted acquisition. A. Unprocessed tracto-graphy across the entire acquisition volume showing the lumbosacral roots as polylines within the cropping box; three-directional color code for fiber direction: blue, cephalocaudal; green, anteroposterior; red, transverse. B. Processed tractography showing individualized radicular fiber bundles as polytubes; color code: red, right S1; green, left S1; blue, right L5; cyan; left L5.

A B

NORMAL FA AND MD VALUES OF LUMBAR SPINE NERVE ROOTS — LINCOT et al 71

7. Tensaouti F., Lahlou I., Clarisse P., Lotterie J.A., Berry I.: Quantitative and reproducibility study of four tractography algorithms used in clini-cal routine. J Magn Reson Imaging, 2011, 34: 165-172.

8. Le Bihan D., Johansen-Berg H.: Diffusion MRI at 25: exploring brain tissue structure and function. Neuro Image, 2012, 61: 324-341.

9. Breitenseher J., Moritz T., Bodner G., Paul A., Prayer D., Kasprian G.: 3 T MR Tractography of Forearm Nerves and Muscles in Cubital Tunnel Syndrome. RSNA. Chicago, , 2011.

10. Lindberg P., Feydy A., Maier M., Drape J.-L.: Relation between Diffu-sion Tensor Imaging, Electromyogra-phy, and Grip Force Control in Patients with Recurrent Carpal Tunnel Syndrome after Surgical Release. RSNA. Chicago, 2011.

11. Hiltunen J., Suortti T., Arvela S., Seppa M., Joensuu R., Hari R.: Diffusion tensor imaging and tracto-graphy of distal peripheral nerves at 3 T. Clin Neurophysiol, 2005, 116: 2315-2323.

12. Balbi V., Budzik J.F., Duhamel A., Bera-Louville A., Le Thuc V., Cotten A.: Tractography of lumbar nerve roots: initial results. Eur Radiol, 2011, 21: 1153-1159.

13. Eguchi Y., Ohtori S., Yamashita M., et al.: Clinical applications of diffusion magnetic resonance imaging of the lumbar foraminal nerve root entrap-ment. Eur Spine J, 2010, 19: 1874-1882.

14. Budzik J.F., et al.: Assessment of reduced field of view in diffusion tensor imaging of the lumbar nerve roots at 3 T. Eur Radiol, 2013, 23: 1361-1366.

15. Tsuchiya K., et al.: Demonstration of spinal cord and nerve root abnormali-ties by diffusion neurography. J Com-put Assist Tomogr, 2008. 32: 286-290.

16. Yamashita T., Kwee T.C., Takahara T.: Whole-body magnetic resonance neuro graphy. N Engl J Med, 2009. 361: 538-539.

17. Arana E., Royuela A., Kovacs F.M., et al.: Lumbar spine: agreement in the interpretation of 1.5-T MR images by using the Nordic Modic Consensus Group classification form. Radiology, 2010, 254: 809-817.

results on intrasubject variability and precision of measurements. AJR Am J Roentgenol, 2010, 194: W65- 72.

3. Filippi C.G., Andrews T., Gonyea J.V., Linnell G., Cauley K.A.: Magnetic resonance diffusion tensor imaging and tractography of the lower spinal cord: application to diastematomyelia and tethered cord. Eur Radiol, 2010, 20: 2194-2199.

4. Rha D.W., Chang W.H., Kim J., Sim E.G., Park E.S.: Comparing quantitative tractography metrics of motor and sensory pathways in children with periventricular leuko-malacia and different levels of gross motor function. Neuroradiology, 2012, 54: 615-621.

5. Budzik J.F., Balbi V., Le Thuc V., Duhamel A., Assaker R., Cotten A.: Diffusion tensor imaging and fibre tracking in cervical spondylotic myelopathy. Eur Radiol, 2011, 21: 426-433.

6. Le Bihan D., Mangin J.F., Poupon C., et al.: Diffusion tensor imaging: concepts and applications. J Magn Reson Imaging, 2001, 13: 534-546.

subjects included in our study, re-quiring confirmation of the results by further larger studies. Neverthe-less, to our best knowledge we re-ported the largest series of lumbar nerve roots DTI and also the only one on 3T MRI and studied FA and MD measurements on 222 nerves roots.

In conclusion, our study shows that FA and MD are not subject to variations according to the magnetic field.

references

1. Eguchi Y., Ohtori S., Orita S., et al.: Quantitative evaluation and visualiza-tion of lumbar foraminal nerve root entrapment by using diffusion tensor imaging: preliminary results. AJNR Am J Neuroradiol, 2011, 32: 1824-1829.

2. Andreisek G., White L.M., Kassner A., Sussman M.S.: Evaluation of diffusion tensor imaging and fiber tractography of the median nerve: preliminary

Fig. 2. — Same patient as in Fig. 1. A. Right S1 root fiber bundle processed in Fiber-Viewer© software shown within the cropping box with crossing analysis plane. B,C: fractional anisotropy (FA) and mean diffusivity (MD) measurements along this fiber bundle with red squares on the graphs marking the level of the crossing analysis plane.

A B

Carpal fusion may occur in two or more adjacent bones in almost any combination (lunotriquetral, capi-tate-hamate, trapezium-trapezoid, scapho-trapezium, scapho-lunate, capitate-lunate, pisiform-hamate, hamate-triquetrum, triquetral-pisi-form and trapezoid-capitate) (Fig. 1). Those on the ulnar side are more commonly involved (1). The bone fu-sion may involve two or more car-pals or even all carpals may appear as a single bony mass. However, most coalitions occur between car-pals within the same carpal row (1). Coalitions between carpals from dif-ferent rows are thought to be quite rare (2) (Fig. 2). Complicated carpal fusions (Fig. 2E and 2F) are likely to be associated with more widespread anomalies. Fusion of carpal bones is hereditary and the trait is transmit-ted as a dominant factor which is not sex linked (3).

The lunotriquetral coalition is caused by a failure of cavitation of the cartilaginous hand bud precursor (during the fourth to eighth week of gestation) or cartilaginous segmen-tation of a common cartilaginous carpal precursor of the lunate and triquetral bone (4). The resulting malsegmentation between two nor-mally distinct carpals results in a car-pal coalition. This coalition may be fibrous (syndesmosis), cartilaginous (synchondrosis), frequently a mix-ture of both or osseous. As their os-sification centers occur between 6 to 15 years the fusion may become vis-ible on plain film.

Burnett (5) proposed a very sim-ple classification with only two forms of coalition, osseous and non osse-ous.

Results

Our series includes 6 females and 3 males, all known with a LTC. Imag-ing findings and clinical manifesta-tions are summarized in Table II. The age of our patients ranges between 11 and 56 years, with a mean of 30 years. The majority of the LTC were asymptomatic coincidental findings (Fig. 4). Two of them with type 1 LTC – of whom also the second youngest patient – had a symptomatic LTC (Fig. 5, 6). The other two with type 1 LTC were asymptomatic. Most of the LTC were of Minnaar type 3. In three patients LTC was bilateral. One of them (Fig. 7A) had a different Min-naar type on either side, the two oth-er persons had the same type of LTC on both sides (Fig. 7B). On plain radi-ography one patient findings were in

de Villiers Minnaar (6) classified the carpal coalitions into four types, also known as the Minnaar types, on plain film (Table I) (Fig. 3). The latter classification is more frequently used. Minnaar type 1 (about 2%) re-sembles a pseudarthrosis due to its incomplete fibrocartilaginous coali-tion. In type 2 (second most frequent, about 22%) there is an incomplete osseous fusion, whereas in type 3 (most frequent, about 75% ) the os-seous fusion is complete. If other carpal congenital abnormalities are associated with a complete osseous coalition it is considered as a type 4 (about 1%) (7).

Our purpose was to review the plain film findings in our series (n = 9 patients, 12 LTC), compare the data with the literature and define the (po-tential) role of MRI. The differential diagnosis will be discussed as well.

Materials and methods

We examined nine patients with a LTC (of whom four had a known bi-lateral LTC) and analyzed the age, gender, type, clinical manifestation and imaging findings of LTC. A stan-dard MRI protocol was used on a 1,5 Tesla MR with a dedicated wrist coil. A Field of View varying from 100-140 mm and a matrix varying from 256 x 180 to 512 x 360 were used. Coronal SE-T1 weighted (TR 470 msec, TE 22 msec, Slice Thickness (ST) 2 mm), Coronal PD-SE T2 weighted (TR 3930 msec, TE 13 and 93 msec, ST 2 mm) and Coronal GRE TR (23,59 msec, TE 8,31 msec, ST 0,5 mm) images were obtained. The number of averages varied from 1 to 3.

JBR–BTR, 2015, 98: 72-78.

LunotRiquetRaL coaLition, a noRMaL vaRiant that May RaReLy cause uLnaR sided wRist painM. Mespreuve1,2, F. Vanhoenacker1,2,3, K. Verstraete2

Lunotriquetral coalition (Ltc), the most frequent and often bilateral type of carpal coalition, is in general considered as asymptomatic. in rare cases – however – fibrocartilaginous Ltc may be an uncommon cause of ulnar sided pain in the wrist due to the pseudarthrosis or a post-traumatic disruption of Ltc. two rare cases of symptomatic Ltc are presented and the role of MRi is emphasized. MRi shows the pseudarthrosis and may additionally show bone marrow edema and subcortical cysts. in symptomatic cases surgical lunotriquetral fusion may be considered as treatment option.

Key-word: wrist, abnormalities.

From: 1. Department of Radiology, St.-Maarten General Hospital, Mechelen, Belgium, 2. Department of Radiology, University Hospital Ghent, Ghent, Belgium, 3. Department of Radiology, University Hospital Antwerp, Antwerp, Belgium.Address for correspondence: M. Mespreuve, St.-Maarten General Hospital, Leopold-straat 2, 2800 Mechelen, Belgium. E-mail marc.mespreuve@skynet.be

Fig. 1. — Carpal coalitions (arrows). Plain radiography. Thick arrows show the two most frequent coalitions (lunotrique-tral and capitate-hamate).

LUNOTRIQUETRAL COALITION — MESPREUVE et al 73

compliance with Minnaar type 3. Re-valuation on MRI redefined this case as type 2 Minnaar (Fig. 8). Two symptomatic patients with a Min-naar type 1 LTC had ulnar sided pain. MRI showed subchondral cysts and a clear bone marrow edema adjacent the LTC (Fig. 5 B-D, 6 B-D). There were no associated soft tissue ab-normalities. The surrounding joints were considered normal. These two symptomatic patients were success-fully treated, one conservatively with medication and the second was sur-gically treated (Fig. 9).

discussion

Lunotriquetral coalition (LTC) is the most frequent type of carpal co-alition, representing 90% of all carpal fusions (8). The general prevalence is about 0,1% in Caucasian popula-tion and the congenital variant is commonly bilateral. We found a bi-lateral appearance in all of the three patients from whom the wrist was examined bilaterally. The congenital fusion of the lunate and the triquetral

Fig. 2. — Isolated and combined coalitions. (A), (B), (C), (E) and (F) Plain radiography, (B) Coronal SE T1-WI and (D) Coronal SE T2-WI. A and B trapezoid-capitate coalition (arrows), C, and D capitate-hamate (long arrows) and trapezoid-capitate (short arrows). E. scapho-trapezium (long arrow) and capitate-hamate (short arrow). F. trapezo-trapezoideum (long arrow), capitate-hamate (short arrow) and LTC (double short arrow).

A C E

B D F

Fig. 3. — Minnaar types 1 to 4. Schematic view. Type 1: narrowed LTJ with irregular sclerotic margins (arrow). Type 2: incomplete osseous fusion with (small) mostly distal remnant of the joint space (arrow). Type 3: complete osseous fusion between the lu-nate and triquetral bone (arrow). Type 4: = Type 3 with other carpal congenital bony abnormalities.

74 JBR–BTR, 2015, 98 (2)

wrist abnormalities. Complicated carpal fusions are likely to be associ-ated with more widespread anoma-lies.

In our series we only found Min-naar type 1 (33%) and 3 (67%) LTC based on plain film.

present – more common on the left side. This is probably the reason why we encountered more LTC on the left side.

It is usually considered as an asymptomatic normal variant. LTC can be associated with other hand-

bone is more common in females which was in line with our series (Ta-ble II) (female to male ratio is 2:1, as in larger series), in African-Ameri-cans (1,6%) and up to 9% in West Af-rican natives (6). This coalition is mostly bilateral but – if unilaterally

Table I. — (de Villiers) Minnaar classification (plain film) with MR correlation, histopathology and clinical corrrelation.

Coalition Plain film (de Villiers) Minnaar

MRI Histopathology Clinical correlation

Type 1 – narrowed LTJ– irregular sclerotic margins– possibly subcortical cysts

– narrowed LTJ– irregular sclerotic margins– possibly subcortical cysts– possibly bone marrow edema– possibly damage to the surrounding cartilage– possibly concomitant pathology

fibro-cartilaginous (a) symptomatic

Type 2 – incomplete osseous fusion– (small) mostly distal remnant of the joint space

– incomplete osseous fusion– (small) mostly distal remnant of the joint space

incomplete osseous

asymptomatic

Type 3 – complete osseous fusion between the lunate and triquetral bone

– complete osseous fusion between the lunate and triquetral bone– dd. type 2 and 3

complete osseous asymptomatic

Type 4 – complete osseous fusion between the lunate and triquetral bone– other carpal congenital bony abnormalities

– complete osseous fusion between the lunate and triquetral bone– other carpal congenital bony abnormalities– possibly anomalies of the soft tissues

complete osseous other carpal congenital abnormalities

Table II. — Illustrative cases of LTC.

Patient Age(y)

SexF/M

Minnaar type L-side

Minnaar type R-side

Surgery Additional clinical information

Figure(s)

1 EJ 37 M 1 ? / symptomatic 5

2 AM 46 F 3 1 / asymptomatic (both sides) 7A, 7B, 8

3 DN 14 F 1 ? + symptomatic 6, 9

4 KV 56 F ? 3 / asymptomatic 4A, 4B

5 EM 24 F 1 ? / asymptomatic 4C, 4D

6 VK 12 F 3 3 / asymptomatic (both sides) 7C, 7D

7 VE 37 F ? 3 / asymptomatic /

8 BA 38 M 3 ? / asymptomatic /

9 VN 11 M 3 3 / asymptomatic (both sides), complicated carpal fusion

2F

MeanRange

3011 -56

6F / 3M

7L (2?) 5R (4?) 1 2 symptomatic / 7 asymptomatic (4 on both sides)

LUNOTRIQUETRAL COALITION — MESPREUVE et al 75

The type 1 coalition may become symptomatic due to the pseudar-throsis or – as for all structures – in case of fracture. The weaker fibrocar-tilaginous coalition appears to be more susceptible to stress or trauma. The deficient intra-articular cartilage formation at the lunotriquetral joint results in a clinical and anatomic condition similar to degenerative os-teoarthritis (9). The cartilage of the surrounding joints may become damaged as well due to disturbed motion. Our two symptomatic pa-tients had a type 1 Minnaar LTC. Al-though carpal fusion is mostly an asymptomatic condition, symptoms may appear in specific conditions such as after intense and repeated movements. A possible alteration of the normal biomechanics of the wrist may cause an abnormal stress on the contributing joints and the sur-rounding soft tissues.

In rare cases fibrocartilaginous type 1 LTC may be an uncommon cause of ulnar sided pain in the wrist due to the pseudarthrosis or a post-traumatic disruption of the fibrocarti-lagineous LTC (10). Two of our pa-

Fig. 4. — Asymptomatic variant. Plain radiography. A. and B. A LTC (arrows) type 3 was found during X-ray examination after major trauma (patient 4). C. and D. A LTC (long arrows) type 1 was found during X-ray examination after minor trauma (short arrows) (patient 5).

Fig. 5. — LTC type 1 (patient 1). A. Plain radiography, B. Coro-nal SE T1-WI, C. Coronal SE T2-WI and D. Coronal 3D-GRE. A. The LTJ is narrowed with irregular margins (long arrow) and subchondral cysts (short arrow). B, C and D: The LTJ is filled with fibrovascular tissue (long arrows). There is oedema ( B, C,short arrows) surrounding a subchondral cyst (C) (short ar-row) and the margins are irregular (D) (short arrow).

Fig. 6. — Symptomatic patient with LTC type 1 (patient 3). A: Plain radiography and B: Coronal SE T1-WI, C, SE proton density and D, SE T2-WI. Type 1 LTC with subchondral cysts (A) and sur-rounding edema (B-D) in a patient with ulnar sided wrist pain.

A

A

AB

B

B

C

C

C

D

D

D

76 JBR–BTR, 2015, 98 (2)

post-traumatic setting. This widen-ing is namely caused by a compen-satory thickening of the cartilage on the opposing surfaces of the scaph-oid bone and the lunate bone.

Examination of the opposite wrist may be useful as LTC is often bilat-eral (6) (about 60%) but not always of the same type.

pal congenital abnormalities are as-sociated with the complete osseous LTC.

Remarkably LTC has been associ-ated with a widening of the scaphol-unate joint space (12), which does not reflect a scapholunate ligament disruption. This differential diagno-sis may be particularly important in a

tients with type 1 LTC had ulnar sided pain. The other two with type 1 LTC were asymptomatic. Osseous LTC (type 2 and 3) of the lunate bone and the triquetral bone are in general considered as asymptomatic. How-ever a case of painful wrist move-ment on the side of a complete bony fusion has been reported (8), without clear explanation of the possible pathogenetic mechanism. All of our six patients with type 3 LTC (of whom two bilateral) were asymptomatic. Although fractures of osseous coali-tions also have been reported, some suggest that these might have been symptomatic fibrocartilaginous types (10).

Plain film findings

LTC type 1 resembles a pseudar-throsis with irregular sclerotic mar-gins. There is a narrowed joint space (LTJ) between the lunate bone and triquetral bone, possibly with sub-cortical cysts. This non-osseous co-alition may result in degenerative osteoarthritis due to abnormal joint mechanics and the thin cartilage be-tween the affected carpals (11).

In type 2 there is an incomplete osseous fusion with trabeculae tra-versing the lunotriquetral joint space and a (small), mostly distal (but oc-casionally also proximally) remnant of the joint space, whereas in type 3 the osseous fusion between the lu-nate bone and the triquetral bone is complete and forms a so called os lunatotriquetrum. In type 4 other car-

Fig. 7. — Bilateral LTC in the same patient. Plain radiography. A. Type 3 LTC on the left side and B and type 1 LTC on the right side (patient 2). C and D, LTC type 3 on both sides (patient 6).

Fig. 8. — Reevaluation on MRI compared to plain radiography (patient 2). A. Plain radiography, B: Coronal SE T1-WI, C: Coronal SE T2-WI and D: coronal 3D-GRE. A. Based on plain film evalua-tion LTC seems complete, in compliance with Minnaar type 3. B, C and D: on MRI of the same patient, a (small) distal notch can be seen, thus redefining this case as type 2 Minnaar.

Fig. 9. — Surgical treatment of a symptomatic LTC (patient 3). Plain radiography. A corticocancellous wedge from the dorsal side of the distal radius (short arrows) was interposed between the lunate and triquetral bone (long arrows) after resection of the pseudarthrosis.

A A

A

B B

B

C CD D

LUNOTRIQUETRAL COALITION — MESPREUVE et al 77

MRI findings

MRI in LTC type 1 shows the pseudarthrosis with irregular scle-rotic margins a narrowed joint space filled with fibrocartilage between the lunate and triquetral bone, subchon-dral cysts and may additionally show bone marrow edema on fluid-sensi-tive sequences adjacent to lunato-triquetral joint in symptomatic cases. After contrast administration the edema and the fibrovascular tissue

Fig. 11. — Secondary fusion due to surgery for a lunotriquetral ligament lesion. A: Plain radiography and B: Coronal SE T1-WI. There is a carpal “coalition” between the lunate bone and the triquetral bone (long arrows). Notice the bony defect after prel-evation of a bone graft in the distal radius (A, short arrow) and the susceptibility artifacts after surgery (B, short arrows).

Fig. 10. — Secondary fusion due to arthritis. A and B: Plain radiography, C: Coronal SE T1-WI and D: Coronal SE T2-WI. There is a carpal “coalition” (here between the scaphoid and the lunate) but also an associated “coalition” with the radius (ar-rows). The scaphoid bone has lost a part of its volume and the proximal cortical delineation of the lunate bone is irregular. The fusion is secondary to a juvenile arthritis.

Table III. — Carpal coalitions in association with other abnormalities. Differential diagnosis.

Disease Abnormalities

Turner’s syndrome – dorsal and radial bowing of the radius (Madelung deformity)– short fourth metacarpal with a positive metacarpal sign (positive sign of Kosowicz)– carpal angle less than 117°

Ellis-van Creveld syndrome – fusion of the capitate and hamate bone– extra carpal bones– malformed carpals– broad, short middle phalanx– polydactyly (“six-fingered dwarfism”)

Holt-Oram syndrome – polydactyly– hypoplastic or triphalangeal thumb– partial or complete absence of the radial bone– accessory carpal bones

arthrogryposis – camptodactyly– ulnar deviation

diastrophic dysplasia – dwarfism– narrow joint spaces– hitchhiker thumb

Nievergelt-Pearlman syndrome – symphalangismfoetal alcohol syndrome – radio-ulnar synostosishand-foot-uterus syndrome – deformed scaphoid boneoto-palatodigital syndrome – anomalies of shape and position of carpal bones as comma shaped

trapezoid and transverse position of the capitates bone

symphalangism – partial or total absence of interphalangeal joints

A AB B

C D

78 JBR–BTR, 2015, 98 (2)

4. Cockshott W.P.: Carpal fusions. Am J Roentgenol Radium Ther Nucl Med, 1963, 89: 1260-1271.

5. Burnett S.: Hamate-pisiform coalition: morphology, clinical significance and a simplified classification scheme for carpal coalition. Clin Anat, 2011, 24: 188-196.

6. de Villiers Minnaar A.B.: Congenital fusion of the lunate and triquetral bones in the South African Bantu. J Bone Joint Surg Br, 1952, 34: 45- 48.

7. De Fazio M., Cousins B., Miversuski R., et al.: Carpal coalition. Hand, 2013, 8: 157-163.

8. Lotter O., Stahl S., Luz O., et al.: Bilat-eral paradoxically symptomatic luno-triquetral coalition: a case report. Open Access J Plastic Surg, 2010, 6 - www.eplasty.com

9. Gross S.C., Watson H.K., Strickland J.W., et al.: Triquetral-lunate arthritis secondary to synostosis. J Hand Surg Am, 1989, 14(1): 95-102.

10. Lotter O., Amr A., Stahl S., et al.: Pseudarthrosis after disruption of an incomplete luno-triquetral coalition: a case report. Ger Med Sci, 2010, 8, doc34 - http://www.egms.de/en/ journals/gms/2010-8/000123.shtml

11. Resnik C.S., Grizzard J.D., Simmons B.P., et al.: Incomplete carpal coali-tion. Am J Roentgenol, 1986, 147: 301-304.

12. Ganos D.L., Imbriglia J.E.: Symptom-atic congenital coalition of the pisi-form and hamate. J Hand Surg Am, 1991, 16: 646-650.

13. Metz V.M., Schimmerl S.M., Gilula L.A., et al.: Wide scapholunate joint space in lunotriquetral coalition: a normal variant? Radiology, 1993, 188: 557-559.

14. Stäbler A., Glacer C., Reiser M., et al.: Symptomatic fibrous lunato-trique-tral coalition. Eur Radiol, 1999, 9: 1643-1646.

15. Poznanski A.K., Holt, J.F.: The carpals in congenital malformation syn-dromes. Am J Roentgenol, 1971, 112: 443-459.

16. Van Schoonhoven J., Prommersberger K.J., Schmitt R.: Traumatic disruption of a fibrocartilage lunate-triquetral coalition – a case report and review of the literature. Hand Surg, 2001, 6: 103-108.

Herbert screw (16) and/or a cortico-cancellous wedge from the dorsal side of the distal radius (Fig. 9) or the iliac crest.

Normally the symptoms should disappear completely and mostly mobility is also restored, although sometimes there may be a residual a minor loss of range of motion (9).

conclusion

Lunotriquetral coalition (LTC) – the most frequent and often bilateral type of carpal coalition – is usually considered asymptomatic. In rare cases however fibrocartilaginous LTC type 1 may be an exceptional cause of ulnar sided pain in the wrist due to the pseudarthrosis or a post-traumatic disruption of the fibrocarti-laginous LTC. MRI shows the pseud-arthrosis more clearly and offers a more accurate classification than plain films. It directly evaluates the symptomatic variants that may ben-efit from (surgical) treatment by showing eventually present bone marrow edema and subcortical cysts. Associated bone and soft tis-sue abnormalities will be illustrated as well. Moreover, it offers a com-plete preoperative cartography of the surrounding joints. Symptomatic variants may be treated conserva-tively or by lunotriquetral surgical fu-sion which may result in pain relief and restored mobility.

References

1. Garn S.M., Frisancho R., Poznanski A.K., et al.: Analysis of Triquetral-Lunate fusion. Am J Physical Anthropology, 1971, 34: 431-434.

2. Poznanski A.K., Holt J.F.: The carpals in congenital malformation syn-dromes. Am J Roentgenol Radium Ther Nucl Med, 1971, 112: 443-459.

3. Henry M.B.: Anomalous fusion of the scaphoid and the greater multangular bone. Archives of Surgery, 1945, 50: 240-241.

in the synovium and the subcortical cysts will enhance (13), although this does not add any clear diagnostic in-formation.

MRI may provide the necessary information about the condition of the surrounding articular cartilage and may also exclude concomitant pathology of the wrist, which is to be evaluated certainly if any operative treatment is considered.

In type 2 LTC there is an incom-plete proximal osseous fusion and a (small) mostly distal notch filled with hyaline cartilage, whereas in type 3 the osseous fusion is complete. MRI will evaluate the LTJ more in detail. Due to the absence of osseous su-perposition, MRI allows a more accu-rate evaluation and classification than plain radiography. Type 3 coali-tions will often be reclassified on MRI as a type 2 LTC. However this does not seem to have any clinical importance.

In type 4 possibly additional soft tissue anomalies are illustrated as well.

Differential diagnosis

Carpal coalitions may be associ-ated with other hand-wrist abnor-malities (Table III). LTC specifically is most frequently seen in Turner’s syndrome, Ellis-van Creveld syn-drome, Holt-Oram syndrome and ar-throgryposis.

Anamnesis and clinical informa-tion should always exclude an ac-quired fusion secondary to arthritis (Fig. 10), trauma, surgery (Fig. 11), drug intake during pregnancy or a metaplastic conversion of intra-artic-ular structures (fibrous tissue, liga-ments or cartilage) (14).

Therapy

After resection of the pseudar-throsis, lunotriquetral surgical fu-sion (15) may be performed in symp-tomatic cases (i.e. type 1) using a

Two-thirds of primary liver tu-mors in the pediatric population are malignant, and malignant primary hepatic tumors account for 1%-2% of all childhood cancers (1). Primary malignant liver tumors in children comprise a heterogeneous group of neoplasms, the majority of them be-ing hepatoblastomas (HBL) and he-patocellular carcinomas (HCC). HBL occurs almost exclusively in patients younger than 5 years old, whereas HCC develops in older children and adolescents (2). On the other hand, transitional liver cell tumor (TLCT) is an aggressive tumor, which is histo-logically neither typical HBL nor HCC; however, it has several patho-logical findings in common with those tumors. To the best of our knowledge, there are only 10 pediat-ric patients reported with TLCT in the literature (3). In this regard, we re-port an 8-year-old boy with histologi-cally proven TLCT and describe the pertinent radiological findings.

Case report

An 8-year-old boy presented with a one-month history of right upper abdominal pain. He described a fall episode one-month before the onset of abdominal pain. The medical his-tory was otherwise unremarkable.

Physical examination revealed mild hepatomegaly. An abdominal ultrasound (US) showed mild hepa-tomegaly and heterogeneous mass with solid and cystic components in the right lobe of the liver (10 cm in anterior-posterior and 8 cm in trans-verse diameter). The right portal vein was encased within the mass and there was also 2 × 1 cm portal lymph-adenopathy. Laboratory tests showed anemia, leukocytosis (with

Encasement of the right hepatic vein could also be seen.

The US guided percutaneous bi-opsy was performed and histological examination revealed a tumor com-posed of cellular components with features of both HBL and HCC (Fig. 4). The tumor was diagnosed as TLCT. The patient received chemo-therapy. On repeat abdominal MRI (3 months later), the size of the tu-mor was found not to have changed -with further elevated AFP levels (93095.40 IU/ml). Thereafter, a right extended hepatectomy was per-formed. One week after surgery, AFP level dropped to 11901.30 IU/ml. The patient also received postoperative chemotherapy. Two months after the operation, abdominal US showed a 1 cm hypoechoic mass in the left hepatic lobe parenchyma. There was a 2×2 cm solid mass on the anterior aspect of the left hepatic lobe out-side the liver capsule and a 4 × 3.5 cm solid mass in the right lower quad-rant near the incision scar in the sub-cutaneous tissue. Doppler US re-vealed increased vascularity in both

neutrophilia) and elevated erythro-cyte sedimentation rate. Liver func-tion tests were normal. Serum α-fetoprotein (AFP) level was signifi-cantly elevated 37241.67 IU/mL (N: 0-6.67). Magnetic resonance imaging (MRI) revealed a multilobulated-large (9.5 cm in craniocaudal, 10.5 cm in anterior-posterior and 7 cm in transverse diameter), heterogeneous solid mass in the right liver lobe. The tumor involves segments V, VII and VIII (PRETEXT Classification II) (4). The tumor exhibits central necrotic areas with low signal intensity on T1-weighted and high signal intensity on T2-weighted images (Fig. 1 and 2). The tumor enhanced slightly, but less than adjacent liver parenchyma after gadolinium administration (Fig. 3) and showed diffusion restric-tion on diffusion-weighted images.

JBR–BTR, 2015, 98: 79-81.

A RARE MALIGNANT HEPATIC TUMOR OF CHILDHOOD: TRANSITIONAL LIVER CELL TUMOR REVISITEDH.N. Ozcan, B. Oguz, T. Salim, B. Talim, M. Haliloglu

Transitional liver cell tumor is an extremely rare entity and has a poor prognosis. It has similar histopathologic findings with hepatoblastoma and hepatocellular carcinoma. Up to now, only 10 cases have been reported in the literature. We report on an 8-year-old boy with histologically proven transitional liver cell tumor and describe the pertinent radiological findings.

Key-word: Liver neoplasms, in infants and children.

From: Department of Pediatric Radiology, Hacettepe University Medical School, Ankara, Turkey.Address for correspondence: Dr H. Nursun Özcan, M.D., Fakülteler Mah. Dirim sok 22/3 Cebeci, 06590 Ankara, Turkey. E-mail: drhnozcan@yahoo.com

Fig. 1. — Axial T1-weighted image shows hypointense tumor with lobulated margin (arrows) and cystic-necrotic area in the central part of the tumor (arrowhead).

80 JBR–BTR, 2015, 98 (2)

lesions. Thorax and abdominal con-trast enhanced CT showed multiple pulmonary metastases and multiple hypodense lesions in the left liver lobe, a solid mass in the anterior ab-dominal wall within the rectus ab-dominis muscle and another solid lesion in the right abdominal lateral wall (Fig. 5). At the last follow-up, AFP level was 17968.65 IU/ml.

Discussion

Primary malignant liver cell tu-mors in children represent a hetero-geneous group, whereby HBL is seen predominantly in children less than 5 years of age, and HCC in older chil-dren and adolescents. Additionally, Prokurat and Zimmermann denoted a distinct group of malignant hepato-cellular tumor named as TLCT, which

Fig. 4. — Hepatocellular carcinoma component with trabecu-lar pattern composed of cells with high nucleus-cytoplasmic

ratio (Hematoxylin and Eosin. Original magnification: ×100).

Fig. 3. — Axial gadolinium-enhanced T1-weighted MR images show heterogeneous enhancement of the tumor which is hypo-intense relative to the normal liver parenchyma (arrows in A and B). There is lack of enhancement of the central necrotic area (arrowheads in A and B).

A

B

Fig. 5. — Two months after operation post contrast axial CT scan demonstrates a solid lesion in the right abdominal lateral wall (arrows).

Fig. 2. — Axial T2-weighted MR image demonstrates a large hyperintense tumor in the right liver lobe (arrows). Note also the cystic-necrotic area in the central part of the tumor (arrowhead).

TRANSITIONAL LIVER CELL TUMOR — OZCAN et al 81

References

1. Chung E.M., Lattin G.E. Jr., Cube R., et al.: From the archives of the AFIP: Pediatric liver masses: radiologic-pathologic correlation. Part 2. Malig-nant tumors. RadioGraphics, 2011, 31: 483-507.

2. Isaacs H. Jr.: Fetal and neonatal he-patic tumours. J Pediatr Surg, 2007, 42: 1797-1803.

3. Prokurat A., Chrupek M., Kazmirczuk R., et al.: Primary malignant liver cell tumours in children different treat-ment strategies. Ann Diagn Pediatr Pathol, 2006, 10: 17-22.

4. Roebuck D.J., Aronson D., Clapuyt P., et al. International Childrhood Liver Tumor Strategy Group.2005 PRE-TEXT: a revised staging system for primary malignant liver tumours of childhood developed by the SIOPEL group. Pediatr Radiol. 2007, 37: 123-132.

5. Prokurat A., Kluge P., Kosciesza A., Danuta P., Kappeler A., Zimmermann A.: Transitional liver cell tumors (TLCT) in older children and adoles-cents: A novel group of aggressive hepatic tumors expressing beta-catenin. Med Pediatr Oncol, 2002, 39: 510-518.

6. Moore S.W., Hesseling P.B., Wessels G., Schneider J.W.: Hepatocellular carcinoma in children. Pediatr Surg Int, 1997, 12: 266-270.

T2-weighted images with delayed enhancement.

The imaging characteristics of HBL reflect its histologic composi-tion. Histologically, HBL is classified into two types: the epithelial type and the mixed epithelial and mesenchymal type. Epithelial HBLs demonstrate a more homogeneous appearance, while mixed tumors are more heterogeneous in attenuation. At MR imaging, epithelial HBLs are homogeneously slightly hypo-intense on T1-weighted images and hyperintense on T2-weighted imag-es relative to adjacent liver paren-chyma. Mixed tumors demonstrate more heterogeneous signal intensity characteristics.

In our patient, the giant tumor in-cluded a central necrosis and serum AFP level was very high expected than HCC or HBL. MR images dem-onstrated slightly contrast enhance-ment after gadolinium administra-tion but less than adjacent liver parenchyma. Diffusion restriction was seen on diffusion-weighted im-ages.

In conclusion, presenting this rare case of ours, we underscore the im-portance of multidisciplinary ap-proach for prompt diagnosis of TLCT whereby the onward treatment will be tailored accordingly.

share common histological features with both of the aforementioned two types (5). They occur in older chil-dren and adolescents and markedly express beta-catenin, typically with a mixed nuclear and cytoplasmic pat-tern (5).

The usual presentation is that of a large or multifocal and/or unresect-able primary mass most commonly seen in the right liver lobe. TCLT dis-plays an expanding growth pattern, sometimes exhibiting a large central necrosis. In cases of TLCT, in con-trast to HBL and HCC, the tumor dis-plays a poor response to initial che-motherapy with higher serum AFP levels at diagnosis. In the previous reports, due to the giant size of the tumors, surgery was connected with many technical difficulties and re-sulted in a high percentage of pa-tients in whom microscopic radicali-ty of surgery was impossible (3).

Concerning the differential diag-nosis, HCC usually occurs in children aged 10-14 years and 40-60% of the cases have elevated serum AFP lev-els (6). On post-contrast MR images, HCC typically demonstrates early ar-terial phase enhancement and may wash out with relative low signal in-tensity during the portal venous phase. If present, the tumor capsule is usually hypointense on T1- and

Erdheim-Chester disease (ECD) is a rare non-Langerhans cell histiocy-tosis. The aetiology of this disease is unknown (1, 2). The disease was first described as “lipoid-granuloma-tous” in two patients by William Chester and Jakob Erdheim in 1930 (2). Histologically, ECD is char-acterized by infiltration by foamy non-Langerhans cell histiocytes, Tou-ton-type giant cells and mixed lym-phoid infiltrates (3). The commonest involvement sites are bone, skin, or-bit, pituitary, retroperitoneal space, pericardium, and lung (1, 2). Perivas-cular region, central nervous system, thyroid, testis, liver, and spleen are the rarer sites of ECD involve-ment (4). The symptoms of the dis-ease include bone pain, diabetes in-sipidus, exophthalmos, dyspnoea, and neurologic symptoms. The prog-nosis of the disease depend on ex-tend and severity of extraskeletal findings. Of 57% patients die second-ary to pulmonary or cardiac involve-ment (2). Isolated breast involve-ment in ECD has been rarely described. To the best of our knowl-edge, there are only a few case re-ports of ECD of the breast. We pres-ent a case of ECD presenting as bilateral clinically malignant breast and axillary masses, with imaging findings suggestive of bilateral breast cancer.

Case report

In September 2013, a 62-year-old woman presented with palpable breast masses in the upper outer quadrant of both breasts and axillar regions. She did not have any trau-ma and did not report nipple dis-

biopsy was applied to the lesions on her axillary regions. In the histopa-thology evaluation, the histiocytes had positive staining with antibodies for CD68 and were negative for S-100 protein (Fig. 2). After clinic, radiolog-ic and histopathology evaluations of breasts, additionally systemic radio-logical and clinic evaluations were applied but there was no sign of ECD involvement in any other part of the body. As a result a diagnosis of iso-lated breast involvement of ECD was then proposed.

Discussion

ECD is an extremely rare disease of unknown aetiology, with distinct clinic, pathologic, and radiological findings in the absence of detectable serum lipid abnormalities. The clini-cal manifestations of ECD are not well defined, but men and women in the 50th through 70th decades of life appear to be affected (2). The clinical presentation is largely determined by the sites of involvement as a re-sult of mass lesions, local pain, or functional compromise. Usually the most commonly reported symptoms are juxtaarticular bone pain of knee and ankle. There is no response to analgesics. Other symptoms include exophthalmos, diabetes insipidus, constitutional symptoms, retroperi-toneal or ureteric/renal involve-ment (2). However less frequently symptoms are neurologic symp-toms, body cavity effusions, cutane-ous xanthomas and diffuse intersti-tial lung involvement (2). Isolated breast involvement by ECD is ex-ceedingly rare and has rarely been described in the literature (3, 5, 6). There are only six cases published in the English literature (7). Our case is one of the few cases in the literature.

In histological determination, there are mainly xanthomatous non-Langerhans histiocyte cells. The histiocytes in this disease do not

charge. The patient had been admit-ted to the hospital with a complaint of nodular lesions on her eyelids 6 years ago. At that time these lesions were evaluated and biopsy was per-formed. Histopathology result was xanthomatous granuloma. In the meantime she did not have any spe-cific complaints except cervical and lumbar pain. Blood and biochemical examinations, abdominal ultraso-nography (US), cervical and lumbar magnetic resonance imaging (MRI) examinations were normal. In radiol-ogy department of our institute, the patient was evaluated with mam-mography. There were irregular opacities on both outer quadrants and axillary regions (Fig. 1A). After that she was examined with US. In US evaluation there were multiple, hypoechoic, macrolobulated solid nodular lesions. Also there were hy-poechoic nodular lesions with hyper-echogenic area which were thought as central hilum of lymph nodes at her both axillary regions (Fig. 1B). Af-ter these evaluations the patient was investigated with MRI examination. On MRI images there were well de-fined but irregular lesions in her both breasts extending to the axillary re-gions (Fig. 1C). The lesions were lo-cated on the anterior part of pectoral muscle and they were separated from the muscle. After contrast ma-terial administration, the lesions were enhanced and in pharmacoki-netic evaluations type 2 curves were obtained (Fig. 1D). The lesions had diffusion restrictions in diffusion weighted MRI (Fig. 1E).

After these evaluations the pa-tients was referred to biopsy proce-dure and with 18G needle, tru-cut

JBR–BTR, 2015, 98: 82-84.

ErDhEim ChEstEr DisEasE PrEsEntED isolatED BrEast anD axillary involvEmEntI. Basara1, E. Yavuz1, P. Balci1, E.B. Tuna2, I. Sari3

Erdheim-Chester disease (ECD) is a rare non-langerhans cell histiocytosis of unknown aetiology. the most common sites of involvement are the long bones, skin, orbit, pituitary and retroperitoneal space. isolated breast involvement is rare in the literature. ECD of the breast has been rarely reported. ECD should be considered in the differential diag-nosis of histiocytoid breast lesions, including fat necrosis and histiocytoid invasive mammary carcinoma. in this case report, we present an unusual presentation isolated breast involvement of ECD with radiological and histopathology findings.

Key-word: liopogranulomatosis.

From: Department of 1. Radiology, 2. Pathology, 3. Internal Medicine, Dokuz Eylul University School of Medicine, Izmir, Turkey. Address for correspondence: Dr I. Basara, M.D., Department of Radiology, Dokuz Eylul University School of Medicine, 35340, Izmir, Turkey. E-mail: isilbasara@gmail.com

ISOLATED ERDHEIM CHESTER DISEASE OF BREAST — BASARA et al 83

the breast (11). It is represented by a histiocytic proliferation with pale acidophilic cytoplasm, mildly atypi-cal round vesicular nuclei and so-called lymphocytophagocytosis. The histiocytes are strongly positive for S100-protein as just the opposite of ECD. The most important differential diagnosis is the histiocytoid variant of invasive lobular carcinoma (12, 13). In this pathology, additionally there can be cutaneous metastasis such as eyelids mimicking xanthelas-mas of ECD. Helpful diagnostic fea-tures favouring ECD versus histiocyt-oid carcinoma both in the primary and metastatic depositories are the lack of nuclear atypical in the former as well as absence of the immune reactive for cytokeratin (13).

Immune suppressive drugs, pred-nisolone can be the first option in

In our case, the breast lesions were radiological intermediate espe-cially with axillary involvements. After all the radiologic evaluations, the lesions were accepted as BIRADS 4 lesions. In breast lesions, the differential diagnosis on histology includes fat necrosis, panniculitis such as lupus profundus mastitis, other histiocytic lesions including Langerhans cell histiocytosis and Rosai-Dorfman disease, and infec-tion. Other lesions that can mimic histiocytes, particularly the histiocyt-oid variant of invasive lobular carci-noma, must be excluded (10). In fat necrosis, there should be a trauma history and architectural skin distor-tion of the breast. Rosai-Dorfman disease is a rare histiocytic prolifera-tion, primarily nodal-based but with extra-nodal involvement, including

express CD1a or S-100 protein (but may display weak and focal positivi-ty by immunohistochemistry for S-100 protein), and lack intra cyto-plasmic Birbeck granules ultra-struc-turally. The cells are typically positive for CD68 and lysozyme (2, 8, 9). The xanthomatous cells are diffusely in-filtrative, apparently resulting in col-lagenous fibrosis of tissues, and are usually associated with sparse lym-phocytic infiltrates and Touton-type giant cells. All cells lack atypical cyto-logical features in the lesion (2, 3, 8, 9). The symptoms of the disease such as bone pain, pathologic bone frac-tures and pancytopenia depend on replacement of the marrow space by xanthomatous histiocytes typically. The course of disease generally re-flects the extend and sites of extra skeletal involvement (2, 9).

Fig. 1. — A. On mammographic examination of right and left breasts, there are irregular lesions extending from both upper outer quadrant to axillar sites. B. On ultrasonograhic examination, the images at the upper are breast lesions, and the lesions at the lower part are axillar lesions. The breast lesions are hypoechoic, solid lesions with irregular contours. There are hyperechogen areas at the center of the lesions which were thought as hilum (arrows). C. On MR views, the images at the upper part are, axial fat saturated T2 weighted and the images at the lower part are axial post-enhanced subtracted images. There are irregular, enhanced lesions at the anterior of bilateral pectoral muscles extending to both axillar parts (arrows). D. There are type 2 pharmacokinetic curves obtained from the breast lesions. E. At the upper part there are b-1000 diffusion weighted images and at the lower part apparent diffusion coefficient images. There are marked diffusion restriction at the lesions (Circles and arrows).

A

C

B

D E

84 JBR–BTR, 2015, 98 (2)

disease. Although, radiological eval-uations are helpful in diagnosis, his-topathology evaluation should be applied for correct diagnosis.

references

1. Allen T.C., Chevez-Barrios P., Shetlar D.J., Cagle P.T.: Pulmonary and ophthalmic involvement with Erdheim-Chester disease: a case re-port and review of the literature. Arch Pathol Lab Med, 2004, 128: 1428-1431.

2. Veyssier-Belot C., Cacoub P., Caparros-Lefebvre D., Wechsler J., Brun B., Remy M., et al.: Erdheim-Chester disease. Clinical and radio-logic characteristics of 59 cases. Medi­cine (Baltimore). 1996, 75: 157-169.

3. Andrade V.P., Nemer C.C., Prezotti A.N., Goulart W.S.: Erdheim-Chester disease of the breast associ-ated with Langerhans-cell histiocyto-sis of the hard palate. Virchows Arch, 2004, 445: 405-409.

4. Sheu SY., Wenzel R.R., Kersting C., Merten R., Otterbach F., Schmid K.W.: Erdheim-Chester disease: case report with multisystemic manifestations in-cluding testes, thyroid, and lymph nodes, and a review of literature. J Clin Pathol, 2004, 57: 1225-1228.

5. Ferrozzi F., Bova D., Tognini G., Zuccoli G. Pseudotumoral bilateral in-volvement of the breast in Erdheim-Chester disease: CT appearance. J Comput Assist Tomogr, 2000, 24: 281-283.

6. Tan A.P., Tan L.K., Choo I.H.: Erdheim-Chester disease involving breast and muscle: imaging findings. AJR Am J Roentgenol, 1995, 164: 1115-1117.

7. Furuta T., Kiryu S., Yamada H., Hosoi M., Kurokawa M., Morikawa T., Shibahara J., Ohtomo K.: Erdheim-Chester disease with an 18F-fluorode-oxyglucose-avid breast mass and BRAF V600E mutation. Jpn J Radiol, 2014, 32: 282-287.

8. Lenahan S.E., Helm K.F., Hopper K.D.: Erdheim-Chester disease. J Cutan Med Surg, 2003, 7: 129-132.

9. Loeffler A.G., Memoli V.A.: Myocardi-al involvement in Erdheim-Chester disease. Arch Pathol Lab Med, 2004, 128: 682-685.

10. Provenzano E., Barter S.J., Wright P.A., Forouhi P., Allibone R., Ellis I.O.: Erdheim-chester disease presenting as bilateral clinically malignant breast masses. Am J Surg Pathol., 2010, 34: 584-588.

11. Ng S.B., Tan L.H., Tan P.H.: Rosai- Dorfman disease of the breast: a mimic of breast malignancy. Patholo­gy, 2000, 32: 10-15.

12. Gupta D., Croitoru C.M., Ayala A.G., Sahin A.A., Middleton L.P.: E-cadherin immunohistochemical analysis of histiocytoid carcinoma of the breast. Ann Diagn Pathol, 2002, 6: 141-1417.

13. Tomasini C., Soro E., Pippione M.: Eyelid swelling: think of metastasis of histiocytoid breast carcinoma. Der­matology, 2002, 205: 63-66.

In atypical breast lesions, a high grade of suspicion is needed to es-tablish the correct diagnosis. It is im-portant to keep in mind that patients harbouring systemic illness and breast lumps should be carefully investigated to exclude a breast involvement by the generalized

medical treatment. Furuta et al re-ported tyrosine kinase inhibitors for cases with BRAF V600E mutation (7). However, our patient receipt im-mune suppressive medical treat-ment, the lesions regressed and is followed-up period at our depart-ment.

Fig. 2. — A. Core biopsies show extensive infiltration by foamy histiocytes with no nuclear atypical, and a patchy mature lym-phoid infiltrate (H&E x20). B. The histiocytes are strongly posi-tive for CD68, and C. negative for S100.

A

B

C

Brunner’s gland hamartomas (BGHs) are uncommon duodenal le-sions, and account for approximate-ly 5% of all duodenal masses (1). BGHs contain cysts on rare occa-sions, and only a very small number of cases of cystic BGH with accom-panying computed tomography (CT) or magnetic resonance imaging (MRI) documentation have been re-ported (2-4). Imaging diagnosis of BGHs is usually a challenge, due to their nonspecific appearance. The presence of cysts can limit the differ-ential diagnosis to cystic duodenal lesions. Characterization of the cysts in BGHs may facilitate their differen-tiation from other cystic duodenal le-sions. Herein, we present a case of duodenal cystic BGH and review the cases reported to date, to determine the differences between the cysts of BGHs and those of other cystic duo-denal lesions.

Case report

A 39-year-old man who had been experiencing postprandial epigastric pain and vomiting for three days vis-ited our emergency department. The results of his physical examination, laboratory tests, and abdominal so-nography were unremarkable. An endoscopic examination of the up-per gastrointestinal tract revealed a shallow gastric ulcer and a submu-cosal tumor in the second portion of the duodenum causing marked luminal narrowing (Fig. 1). Contrast-enhanced abdominal CT showed a 2.2 × 4.5 cm mass with multifocal, cyst-like low densities occupying the second portion of the duodenum (Fig. 2). An endoscopic biopsy of the duodenal tumor was performed and pathologic examination revealed

Discussion

Brunner’s glands are normally lo-cated in the duodenal submucosa, and their concentration gradually de-creases towards the distal part of the duodenum. These glands produce an alkaline secretion to protect the duodenum from injury by gastric acid and optimize the pH for pancre-atic enzymes. The pathogenesis of BGHs is unknown. BGHs are glandu-lar hyperplasias or hamartomas, rather than true neoplasms. Because these lesions do not have features of cellular atypia, the terms Brunner’s gland “hyperplasia” or “hamarto-ma” are preferable to “adenoma.” The distinction between Brunner’s gland hyperplasia and Brunner’s gland hamartoma is arbitrary. Ac-cording to the guidelines used at the Armed Forces Institute of Pathology (Washington DC, USA), if the lesion is less than 5 mm in size it is called a hyperplasia, and if it is greater than 5 mm it is called a hamartoma (5).

that the cells and lobules of the Brunner’s glands were enlarged, but otherwise normal in appearance with no cellular atypia (Fig. 3). On the basis of the histologic findings, a diagnosis of a Brunner’s gland hamartoma was made. During hos-pitalization, oral intake was withheld and esomeprazole and metoclo-pramide were administered intrave-nously. Three days after admission, his abdominal pain and vomiting re-solved, and the patient was dis-charged from the hospital. Surgical excision of the duodenal hamartoma was not performed. The patient has followed up twice in the outpatient department in the 1 month since his discharge, and remains symptom-free.

JBR–BTR, 2015, 98: 85-87.

Cysts in a Brunner’s GlanD Hamartoma: a Clue to DiaGnosisY.-K. Fan, Y.-P. Liu, Y.-L. Lin, W.-K. Su1

the appearance of cystic Brunner’s gland hamartomas (BGHs) on computed tomography (Ct) or magnetic resonance imaging (mri) has only been reported in a very small number of cases. imaging diagnosis of cystic BGHs is usually difficult. We present a case of cystic BGH and characterize it in conjunction with previously reported cases. We found that the cysts of BGHs are smaller than those of other cystic duodenal lesions. the presence of cysts in BGHs can limit the differential diagnosis to cystic duodenal lesions, and our observations may assist others in the discrimina-tion of cystic BGHs from other cystic duodenal lesions.

Key-words: Brunner’s gland – Hamartoma – Duodenum — Cyst.

From: 1. Dpt of Radiology, Mackay Memorial Hospital, Taiwan, R.O.C.Address for correspondence: Dr Y.-K. Fan, Dpt of Radiology, Mackay Memorial Hospi-tal, N° 690, Sec. 2, Guangfu Rd, East Dist., Hsinchu, Taiwan 300, R.O.C.E-mail: fyk5358@gmail.com

Fig. 1. — Endoscopic findings. A submucosal tumor with intact mucosa (asterisk) in the second portion of the duodenum causes luminal narrowing.

86 JBR–BTR, 2015, 98 (2)

agnosis because the differential di-agnosis is limited to cystic duodenal lesions. Cysts in BGHs differ from those in other cystic duodenal le-sions. Duplication cysts and lymph-angiomas are thin-walled cystic masses without mural nodules or soft tissue components. These cysts are larger than those in cystic BGHs (7, 8). In contrast with conven-tional soft tissue and central nervous system schwannomas, cysts are un-usual in gastrointestinal schwanno-mas (9). If present, they are fewer in number and larger than those in BGHs of similar size (4, 10). Cystic change is not uncommon in gastro-intestinal stromal tumors (GISTs),

literature to date (2-4). These cysts are dilated acini or ducts of Brunner’s glands. The CT findings of the previ-ously reported cases and the current case are summarized in Table I. The smaller hamartomas (< 2 cm) con-tain a solitary cyst, and the larger ones (≥ 4 cm) manifest as a soft tis-sue mass containing multiple cysts. These cysts are round, ovoid, or elongated in shape. We found that the short axes of all cysts in these BGHs do not exceed the diameter of the abdominal aorta at the same lev-el, in axial images. Diagnosis of BGHs is usually difficult, due to non-specific imaging findings. However, the presence of cysts is helpful in di-

BGHs may be symptomatic, and symptomatic lesions are often larg-er. The most common presentations in symptomatic patients are gastro-intestinal hemorrhage (37%) and obstructive symptoms (37%) (6). In a series of 27 patients, 70% of BGHs were located in the duodenal bulb, 26% in the second portion of the duodenum, and 4% in the third portion, and most (89%) were pe-dunculated (6).

Cysts can only be identified in BGHs via CT or MRI on rare occa-sions. To our knowledge, only 4 cases demonstrating the CT or MRI ap-pearance of cystic BGHs have been reported in the English-language

Fig. 2. — Axial and coronal contrast-enhanced CT images. A mass (white arrows) in the second portion of duodenum contains multiple small cysts.

Duodenal lumen: black and white arrowheads, pancreas: asterisks.

Fig. 3. — Photomicrographs of the BGH (hematoxylin and eosin, original magnification ×80 in A and ×200 in B) shows proliferation of Brunner’s glands and cystic dilatation of a gland (asterisk in B).

A B

CYSTS IN A BRUNNER’S GLAND HAMARTOMA — FAN et al 87

8. Levy A.D., Cantisani V., Miettinen M.: Abdominal lymphangiomas: imaging features with pathologic correlation. AJR Am J Roentgenol, 2004, 182: 1485-1491.

9. Levy A.D., Quiles A.M., Miettinen M., Sobin L.H.: Gastrointestinal schwan-no mas: CT features with clinico-pathologic correlation. AJR Am J Roentgenol, 2005, 184: 797-802.

10. Bayraktutan U., Kantarci M., Ozgokce M., Aydinli B., Atamanalp S.S., Sipal S.: Education and Imaging. Gastrointestinal: benign cystic schwannoma localized in the gastroduodenal ligament, a rare case. J Gastroenterol Hepatol, 2012, 27: 985.

11. King D.M.: The radiology of gastro-intestinal stromal tumours (GIST). Cancer imaging, 2005, 5: 150-156.

12. Naitoh I., Okayama Y., Hirai M., Kitajima Y., Hayashi K., Okamoto T., et al.: Exophytic pedunculated gastrointestinal stromal tumor with remarkable cystic change. J Gastro­enterol, 2003, 38: 1181-1184.

13. McCowin M.J., Federle M.P.: Computed tomography of pancreatic pseudocysts of the duodenum. AJR Am J Roentgenol, 1985, 145: 1003-1007.

14. Vullierme M.P., Vilgrain V., Flejou J.F., Zins M., O’Toole D., Ruszniewski P., et al.: Cystic dystrophy of the duodenal wall in the heterotopic pancreas: radiopathological correla-tions. J Comput Assist Tomogr, 2000, 24: 635-643.

references

1. Botsford T.W., Crowe P., Crocker D.W.: Tumors of the small intestine. A review of experience with 115 cases including a report of a rare case of malignant hemangio-endothelioma. Am J Surg, 1962, 103: 358-365.

2. Park B.J., Kim M.J., Lee J.H., Park S.S., Sung D.J., Cho S.B.: Cystic Brunner’s gland hamartoma in the duodenum: a case report. World J Gastroenterol, 2009, 15: 4980-4983.

3. Hur S., Han J.K., Kim M.A., Bae J.M., Choi B.I.: Brunner’s gland hamartoma: computed tomographic findings with histopathologic correlation in 9 cases. J Comput Assist Tomogr, 2010, 34: 543-547.

4. Lee J., Park C.M., Kim K.A., Lee C.H., Choi J.W., Shin B.K., et al.: Cystic lesions of the gastrointestinal tract: multimodality imaging with patho-logic correlations. Korean J Radiol, 2010, 11: 457-468.

5. Patel N.D., Levy A.D., Mehrotra A.K., Sobin L.H.: Brunner’s gland hyper-plasia and hamartoma: imaging fea-tures with clinicopathologic correla-tion. AJR Am J Roentgenol, 2006, 187: 715-722.

6. Levine J.A., Burgart L.J., Batts K.P., Wang K.K.: Brunner’s gland hamarto-mas: clinical presentation and patho-logical features of 27 cases. Am J Gastroenterol, 1995, 90: 290-294.

7. Macpherson R.I.: Gastrointestinal tract duplications: clinical, pathologic, etiologic, and radiologic consider-ations. Radiographics, 1993, 13: 1063-1080.

and is more commonly seen in large tumors. Cysts in GISTs are signifi-cantly larger than those in BGHs (11, 12). The duodenum is an uncommon site for pancreatic pseudocyst formation. In the cases reported by McCowin and Federle (13), these pseudocysts were solitary and larger than the cysts of BGHs, and in all patients, CT revealed other evidence of pancreatitis in addition to duode-nal pseudocysts. Cystic dystrophy in heterotopic pancreas (CDHP), most commonly located in the second portion of the duodenum, has a simi-lar appearance to cystic BGHs. These cysts are multiple and small, and are very similar to those seen in larger cystic BGHs. However, inflammatory changes such as periduodenal edema, periduodenal effusion, and enlarged lymph nodes are observed in most cases with CDHP (14).

Conclusion

The size of cysts is the key to diag-nosing cystic BGHs. The cysts in BGHs are smaller than the cysts of other cystic duodenal lesions, except CDHP. When a mass containing a small solitary cyst or multiple small cysts is found in the first and second portion of the duodenum, without evidence of periduodenal inflamma-tory change or pancreatitis, cystic BGH should be considered.

Table I. — Summary of the CT findings of cystic BGHs in the current case and the literature.

Case Age/Sex Size of tumors (cm)a Number of cysts Reference

1 39/M 4.5 7 Current case

2 30/M 4.0 ≥ 10c (2)

3 49/F 1.8 1 (3)

4 66/F 6.0 ≥ 5c (3)

5 48/M < 2.0b 1 (4)

a Largest dimension.b The exact size of the BGH is not available in case 5. However, the largest dimension of the hamartoma approximates the diameter of abdominal aorta, and the diameter of abdominal aorta is estimated to be less than 2 cm. c Because limited images are available in these cases, exact number of cysts cannot be determined.

Case report

A 63-year-old man treated for bi-lateral pneumonia for two days was admitted to the emergency depart-ment with a temperature of 39°C, tachycardia and left basi-thoracic chest pain. Blood tests showed a C-reactive protein (CRP) of 92 mg/L (reference < 5 mg/L) without leuko-cytosis.

Chest radiography at the admis-sion showed bilateral opacities in the bases of the lungs.

Patient’s condition worsened de-spite large spectrum antibiotherapy. A repeated chest radiograph, two days after admission, showed a com-plete opacification of the left lung.

A chest computed tomography (CT) was performed and showed a huge left sided pleural effusion (Fig. 1). The ultrasonography show-ed multiloculated pleural fluid. The diagnostic was a rapid development pleural empyema. A left thoraco-scopy was performed and the collec-tion was drained.

Three months later, a control CT-scan was performed and showed that the left pleural empyema had completely resolved but it revealed in the left basal trunk, a nodular le-sion, with homogeneous fat density and a bronchus parietal defect (Fig. 2).

Bronchoscopy revealed a subtotal obstruction of the basal segmental bronchi of the left lower lobe by a yellowish polypoid lesion protruding into the bronchial lumen (Fig. 3).

Histopathological examination of the biopsies confirmed a prolifera-tion of mature fat tissue. There was no evidence of malignancy, and thus, a diagnosis of endobronchial lipoma was established.

JBR–BTR, 2015, 98: 88-90.

BronChial lipoma: an unusual Cause of pleural empyemaS. Lanotte1, R. Frognier2, O. Van Cutsem2, P. Mailleux1

We report a case of rapidly growing pleural empyema due to endobronchial lipoma. The diagnosis was established by chest computed tomography (CT). endobronchial lipoma is a rare benign tumor of the tracheobronchial tree which can cause irreversible damage to the distal lung parenchyma if diagnosis and treatment are not carried out in time.

Key-word: lipoma and lipomatosis.

From: 1. Department Radiology and Medical Imaging, 2. Department of Pneumology, Clinique Saint-Luc, Bouge (Namur), Belgium.Address for correspondence: Dr S. Lanotte, M.D., Department Radiology and Medical Imaging, Clinique Saint-Luc, Bouge (Namur), Belgium.E-mail: solennelanotte@gmail.com

Fig. 1. — A. Coronal reconstruction CT scan shows a huge left pleural empyema (white star). B. On a control CT performed three months later, meticulous analysis of bronchial tree revealed the presence of an adipose mass occluding the left basal trunk, including a homogeneous fat den-sity area (- 98 HU), with bronchus parietal defect (white arrow).

A

B

BRONCHIAL LIPOMA — LANOTTE et al 89

toms due to upper airways obstruc-tion that occur when more than 50%-75% of the luminal diameter is occluded. All of the symptoms are nonspecific and include persistent cough, dyspnea, chest pain, hemop-tysis and recurrent pneumonia.

Endobronchial lipoma is a benign tumor but it can cause irreversible damages to the distal lung parenchy-ma, unless the diagnosis and treat-ment are carried out in time (3).

Chest radiographs usually show nonspecific changes related to the

sively of mature fat tissue (1) arising from the submucosal fat of large bronchus.

Benign neoplasm of the tracheo-bronchial tree is quite rare, account-ing for less than 10% of all airways neoplasms, while endobronchial lipoma is extremely rare.

The tumors are more frequent in middle-age men (mean age, 60 years). Smoking and obesity are significant risk factors (2).

Lipoma, like other endo-bronchial tumors, produces respiratory symp-

The patient was treated with endo bronchial resection by laser and cryotherapy.

The patient underwent broncho-scopic regular follow up to detect a possible recurrence because of the bronchus parietal defect at the level of the lipoma.

Discussion

Endobronchial lipoma is a benign tumor of the tracheobronchial tree, defined as a mass composed exclu-

A

C

B

D

Fig. 2. — CT scan with contrast injection of iodinated contrast (KV 120, mA 180, DLP 390 mGycm). Axial CT views at the level of left basal trunk show an endobronchial lipoma (- 111 HU) in the left basal trunk (white star) with bronchus parietal defect (black arrow-heads) through bronchial wall (white arrows).

90 JBR–BTR, 2015, 98 (2)

is permanent distal damage or any feature suggesting a possible malig-nant process (2).

Conclusion

Endobronchial lipoma is a rare and benign tumor of the lung. During the analysis of a chest CT scan, a systematic examination of the endo-bronchial tree should be performed, especially when imaging shows per-sistent changes due to upper airway obstruction or rapid development pleural empyema. Early diagnosis and endoscopic resection help to prevent irreversible distal lung dam-age.

references

1. Park C.M., Goo J.M., Lee H.J., Kim M.A., Lee C.H., Kang M.J.: Tumors in the tracheobronchial tree: CT and FDG PET features. Radio-graphics, 2009, 29: 55-71.

2. Muraoka M., Oka T., Akamine S., et al.: Endobronchial lipoma: review of 64 cases reported in Japan. Chest, 2003, 123: 293-296.

3. Ouadnouni Y., Bouchikh M., Bekarsabein S., et al.: Endobronchial lipoma a rare cause of pleural empy-ema: a case report. Cases Journal, 2009, 2: 6377.

4. Simmers T.A., Jie C., Sie B.: Endo-bronchial lipoma posing as carcino-ma. Neth J Med, 1997, 51: 143-145.

5. Wilson R.W., Kirejezyk W.: Pathologi-cal and radiological correlation of endobronchial neoplasm. Part I. Benign tumors. Ann Diagn Pathol, 1997, 1: 31-46.

6. Rodrigues A.J., Coelho D., Dias Junior S.A., Jacomelli M., Scordamaglio P.R., Fiqueiredo V.R.: Minimally invasive bronchoscopic re-section of benign tumors of the bron-chi. J Bras Pneumol, 2011, 37: 796-800.

identified on CT, differential diagno-sis should include lipoma and fatty hamartoma that also can appear as a fatty mass (5). Pathological analysis of the resected tissue is necessary for accurate diagnosis but the dis-tinction of endobronchial lipomatous hamartoma from lipoma is of minor interest as both are rare benign mes-enchymal tumors in clinical practice

Definitive diagnosis is made by bronchoscopy and biopsy (6). Bron-choscopic examination typically re-veals a yellow polypoid mass with a smooth, regular and soft surface.

Bronchoscopic resection is the treatment of choice as it helps pre-serving lung parenchyma. However, surgical resection is required if there

bronchial obstruction such as atelec-tasis or pneumonia. Less frequently, pleural effusion was observed on chest radiography. Our patient pre-sented pneumonia of the left lung with rapid development pleural em-pyema. Pleural empyema associated to endobronchial lipoma was only recorded in four cases, and this is the fifth English-language reported (3).

CT typically shows a homoge-neous mass with fat density (-70 UH to -140 UH) and no contrast enhance-ment (4). Because of his high speci-ficity and sensitivity in fat detection, computed tomography (CT) has a key role in establishing the diagnosis of endobronchial lipoma. However, when a fatty endobronchial mass is

Fig. 3. — Bronchoscopy revealed a smooth, polypoid lesion protruding into the bronchial lumen, obstructing the orifice of the left basal trunk (white arrows), below the Nelson bronchus (black arrows).

‘Backfill’ of the sacroiliac joint space in spondyloarthritis

F. Laloo1, N. Herregods1, H. Cypers2, K. Verstraete1, L. Jans1

Three patients of the outpatient rheumatology clinic of our hospital with inflamma-tory type low back pain suggestive for spondyloarthritis were referred for MRI of the sacroiliac joints. MRI showed high T1 signal within the SI joint (arrows) in all three patients, filling the extended erosions of the iliac bone in two patients (Figs. A and B) whereas a more petechial appearance in the sacroiliac joint space was seen in the third patient (Fig. C). The diagnosis of ‘backfill’ of erosions and of the sacroiliac joint space in spondyloarthritis was made.

Comment

The prevalence of spondyloarthritis is estimated 1.5%. MRI of the sacroiliac joints is a cornerstone in the diagnosis, classification and follow-up of the disease since it de-picts active inflammatory lesions long before radiographic changes become evident. Moreover, MRI may demonstrate late structural changes such as erosions, sclerosis, fat deposition and ankylosis.

As new bone formation is a hallmark of spondyloarthritis, ankylosis of the sacroiliac joint is the well-known end-stage of the disease. Ankylosis may occur at sites where new bone formation is present. Early new bone formation may be seen as high T1 signal within erosions or within the sacroiliac joint itself. As this tissue reossifies the eroded bone and fills the joint space, this high T1 signal is called ‘backfill’ and is very specific for the diagnosis of spondyloarthritis.

Reference

1. Weber U., et al.: Can erosions be reliably detected in patients with ankylosing spondylitis? A cross- sectional study. Arthritis Res Ther, 2012, 14: R124.

JBR–BTR, 2015, 98: 91.

IMAGES IN CLINICAL RADIOLOGY

1. Department of Radiology, 2. Department of Rheumatology, Ghent University Hospital, Gent, Belgium.

A

B

C

Rare presentation of Langerhans cell histio-cytosis

C. Gieraerts1, P. Vandaele1, R. Schildermans2, L. Daveloose3, K. Ramboer1

A 51-year old woman consulted at our hospital with rightsided chest pain. A thoracic radiograph revealed multiple nodules in both lungs. These nodules were not present on a radiograph taken two months earlier during a routine check-up. A computed tomography of the tho-rax demonstrated multiple nodules varying from 3 to 15 mm (Fig. A). The nodules were spiculated and randomly distributed with predomi-nance for the middle and upper lung zones (Fig. B). The mediastinal and hilar lymph nodes were slightly enlarged. Furthermore, a right sided osteolytic bone lesion was present anterolaterally in the eighth rib (Fig. C). Since the radiological differential diagnosis was very broad, a surgical lung biopsy and partial rib resection was performed. Histo-logically, the nodules showed infiltration of lymphocytes, eosinophils and Langerhans’ cells. The osteolytic bone lesion showed typical find-ings of an eosinophilic granuloma. Both lesions were consequently a presentation of the same disease and the diagnosis of Langerhans’ cell histiocystosis (LCH) with pulmonary involvement was made. Immuno-suppressive therapy with glucocorticoids was initiated resulting in partial regression of the pulmonary nodules on a follow-up scan. Unfortunately, the patient refused to stop smoking and a recent scan showed disease progression.

Comment

Langerhans cell histiocytosis is a rare histiocytic disorder character-ized by accumulations of large mononuclear cells forming granulomas in various organs. The disease can be divided into two groups based on single or multisystem involvement. Multisystemic disease is most common in young children whereas single organ involvement is more common in adults. When the lung is the primarily affected organ, the disease is called pulmonary Langerhans cell histiocytosis (PLCH), pre-viously known as Histiocystosis X or eosinophilic granuloma of the lung. In the whole spectrum of LCH, PLCH is considered as somewhat atypical since it is associated with cigarette smoking in more than 90 percent of cases. Patients with PLCH usually present with respira-

tory or constitutional symptoms. Frequently the diagnosis is made incidentally on a routine chest radiograph. Our patient shows multisystemic disease but the pulmonary changes are most notably. Typically, in the first stage of pulmonary disease, multiple small nodules are present with middle and upper lung zone predominance. The nodules are normally around 5 mm in diameter. With disease progression, the nodules start to excavate and form cystic lesions. The advanced form is the most commonly seen by radiologists: multiple pulmonary cysts, usually less than 10 mm in diameter but often confluent, with variable wall thickness and middle and upper lung zone predominance, sometimes associated with some nodules. Multiple nodules lesions in the absence of cysts are only present in the early stage of LCH and are seldom seen in imaging. Multisystemic presentations like in our patient with a concomitant bone lesion are rare in older patients. Next to smoking cessation, glucocorticoid treatment is often effective in the treatment of this disease (1).

Reference

1. Mogulkoc N., Veral A., Bishop P.W., Bayindir U., Pickering C.A., Egan J.J.: Pulmonary Langerhans’ cell histiocytosis: radiologic resolution following smoking cessation. Chest, 1999, 115: 1452.

JBR–BTR, 2015, 98: 92.

IMAGES IN CLINICAL RADIOLOGY

1. Department of Radiology, 2. Department of Pneumology, AZ Sint-Lucas, Bruges, Belgium, 3. Medical student, Ghent University, Ghent, Belgium.

A

B

C

A glomus tumor

P. Kulczycka1, B. Dallaudière1, O. Barbier2, B. Vande Berg1, A. Larbi1

A 54-year-old woman presented with a two-year-history of ten-derness and cold sensitivity of the subungueal region of her 3rd left finger. Physical examination revealed a pinkish red spot under the nail. Ultrasonography of the 3rd left finger shows a well-delimited hypoechoic and hypervascularised mass in the subungueal area (Fig. A). X-ray of the 3rd left finger shows well-defined bony erosion with sclerotic margins in the distal phalanx (scalloping) which indi-cating a chronic growth lesion (Fig. B). The diagnosis proposed was the glomus tumor. A tumor excision was performed and the histo-logic examination confirmed the glomus tumor diagnosis (Fig. C). The localized pain and point tenderness disappeared after surgical excision. The patient did not experience any recurrence or further complications.

Comment

Glomus tumor is rare, often-benign neoplasms arising from a neuromyoarterial structure called a glomus body that controls blood pressure and temperature (Masson’s tumor). It can appear in any part of the body, although it mostly appears in the extremities, especially in the hand at the subungueal area. Usually, it is a soli-tary tumor, but in 10% of all cases there are multiples lesions.

Glomus tumors are difficult to diagnose clinically and the diag-nosis is frequently delayed. Classic clinical symptoms are aching pain, focal tenderness and cold hypersensitivity. Bony erosion with sclerotic margins can be seen on X-Rays but this finding is rare (Fig. B). An ultrasound examination is a good tool for detecting glomus tumor, showing a well-delimited hypo echoic and hyper vascularised mass (Fig. A). It can however underestimate the size of a glomus tumor.

MRI can be used but is not necessary for the diagnosis. It shows a well-delineated mass with a low signal on T1-weighted sequence and high signal intensity on T2-weighted sequence. Glomus tumor enhances early, intense and rapid after intravenous gadolinium administration.

The treatment of a glomus tumor should be surgical and the dramatic pain relief after excision is characteristic. These tumors have a high recurrence rate because it is difficult for the surgeon to distinguish it from the adjacent tissues.

Reference

1. Tang C.Y.K., Tiptoe T., Fung B.: Where is the Lesion? Glomus Tumours of the hand. Archives of Plastic Surgery, May 2013.

JBR–BTR, 2015, 98: 93.

IMAGES IN CLINICAL RADIOLOGY

1. Department of Radiology, 2. Department of Orthopedy and traumatogy, Cliniques Universitaires St-Luc, Brussels, Belgium.

A

B

C

Dirty shadowing in emphysematous pyelonephri-tis

A. Eeckhoudt1,2, F.M. Vanhoenacker1,2,3

A 62-year-old female presented with progressive pain at the left loin re-gion, chills and fever for 1 week.

Laboratory results showed an elevation of CRP and raise of white blood cell count. Creatinine and urea were normal. Midstream urine sample showed an elevation of white blood cells, as well as detection of bacteria. The patient had no prior medical history of diabetes mellitus or urinary obstruction. Glycemia was within normal range.

On ultrasound, visualization of the left kidney was largely impaired due to dirty shadowing caused by subcapsular air (Fig. A, arrows). CT scan of the abdomen showed an enlarged left kidney, with multiple subcapsular air bubbles (Fig. B and C, arrows). There was inhomogeneous enhancement of the left kidney with formation of subcapsular fluid collections at the upper and middle pole.

Based on the combination of characteristic clinical and imaging findings, the diagnosis of an emphysematous pyelonephritis was made.

The patient was treated with intravenous antibiotics and left nephrostomy.Clinical evolution was favorable, although the left kidney showed a

shrunken appearance on follow-up studies.

Comment

Emphysematous pyelonephritis is an acute necrotizing infection of the renal parenchyma and its surrounding tissues. Mortality due to emphysema-tous pyelonephritis is most commonly associated with septic complications. Mortality rate is approximately 20%. In 95% of the cases, there is an underly-ing undiagnosed diabetic mellitus. Other risk factor are urinary tract obstruc-tion, immunosuppression, polycystic kidneys and end stage renal disease. Our patient had none of these predisposing factors. Escherichia coli is the causative pathogen in most cases (70%). It produces gas via fermentation of glucose and lactate.

The clinical presentation of emphysematous pyelonephritis is similar to non-complicated pyelonephritis. Signs and symptoms that may be present are: fever, dysuria, nausea and flank pain. Loin tenderness is frequently noticed on clinical examination. Sometimes crepitus around the renal region or in the scrotum can be felt. Because of its nonspecific clinical presentation, emphysematous pyelonephritis is a radiological diagnosis.

Sings of emphysematous pyelonephritis on ultrasonography (US) are an enlarged kidney, with hyperechogenic foci within the parenchyma and dirty shadowing, corresponding to gas. Impaired visualization of the kidney may result in underestimation of the disease. CT is the preferred imaging tech-nique to allow a specific diagnosis and evaluation of the extent of infection. The presence of bubbly and/or linear gas bubbles within the renal parenchy-ma, combined with renal enlargement and heterogeneous enhancement with wedge-shaped hypodense areas and abscess formation, are the signature of emphysematous pyelonephritis. It is important to distinguish emphysema-

tous pyelonephritis from emphysematous pyelitis, because emphysematous pyelitis has a far better prognosis, and can be treated with mere antibiotics. Emphysematous pyelitis can be recognized by gas formation that is limited to the renal collecting system. The most successful treatment strategy for emphysematous pyelonephritis is a combination of antibiotics with percutaneous or surgical nephrostomy. In conclusion, emphysematous pyelonephritis is a potentially life threatening condition. It is an imaging diagnosis, with CT as the first choice imaging technique.

Reference

1. Ubee S.S., McGlynn L., Fordham M.: Emphysematous pyelonephritis. BJU Int, 2011, 107: 1474-1478.

JBR–BTR, 2015, 98: 94.

IMAGES IN CLINICAL RADIOLOGY

1. Department of Radiology, AZ Sint-Maarten, Duffel-Mechelen, 2. Department of Radiology, University Hospital Ghent (UZ Gent), 3. Department of Radiology, University Hospital Antwerp (UZA).

A

B

C

MDCT findings of polymicrobial descending necrotizing mediastinitis

C. Karanikas, P. Lampropoulou, D. Karakiklas, C.S. Baltas, A. Demertzis, C. Drosos1

A 43-year-old Caucasian male with an uneventful medical history presented to ER complaining of sore throat and fever of 38.3°C. Clinical examination revealed unilateral right sided tonsillitis. Over the next 4 days, he developed malaise, shortness of breath, right-sided neck swell-ing and chest pain and he was admitted. The patient was systematically unwell with clinical manifestations of persistent neck swelling and pain radiating to the right shoulder, fever, slight discoloration of overlying skin, subcutaneous crepitation, mild trismus, odynophagia and decreased oral intake, in the setting of a peritonsillar abscess.

Within a few hours after hospitalization, the patient presented upper airway obstruction due to massive edema, resulting in acute respiratory insufficiency that necessitated tracheal intubation and transfer to the ICU, while planning for surgical intervention.

MDCT scan of the cervical and thoracic region (Fig. A,B,C) showed a parapharyngeal abcess descending into the mediastinum, spreading to-wards the diaphragm, with bilateral pleural effusions, along with diffuse necrotizing fasciitis, collection of gas in mediastinal compartments and soft tissue infiltration with loss of normal fat planes. The patient under-went combined extensive neck and thorax drainage in conjunction with limited surgical debridement consisting of cervicotomy. Swabs identified a polymicrobial infection with predominant species of Streptococcus pyogenes, anaerobic Peptostreptococci and Streptococcus viridans. The empirical antibiotic regimen included piperacillin – tazobactam and vancomycin.

After 29 days of intubation and ventilation in ICU, the patient was extubated and transferred to the ward in order to complete the 6-week-course of IV antibiotic therapy. By this time the patient presented no residual deficits.

Comment

Numerous series of patients suffering of mediastinitis have been re-ported, especially the descending necrotizing type both focal and diffuse, that lead Endo et al to develop a classification system of the extent of the disease on the basis of the CT findings. Type I disease represents mediastinitis confined above the carina, type IIA disease extends to the anterior lower mediastinum and type IIB disease involves both anterior and posterior mediastinum. Patients with type I mediastinitis may not require drainage at all, while type IIA patients require drainage without sternotomy (via the subxiphoid approach) and type IIB patients require drainage with open thoracotomy.

MDCT is the study of choice for evaluation of mediastinal emergencies such as DNM, since it provides sagittal/coronal reconstructions, is readily available, non invasive and easy to perform. There are both primary and

secondary CT features of DNM, both of which were present in our case. Primary features include free gas bubbles in the mediastinum and/or localized fluid collections, or even abscess, and secondary CT findings include increased attenuation of mediastinal fat, pleural and/or pericardial fluid, enlargement of lymph nodes and rarely lung parenchy-mal abnormalities.

Reference

1. Weaver E., Nguyen X., Brooks M.A.: Descending necrotizing mediastinitis: two case reports and review of the literature. Eur Respir Rev, 2010, 19: 141-149.

JBR–BTR, 2015, 98: 95.

IMAGES IN CLINICAL RADIOLOGY

1. Radiology Department, “G. Gennimatas” General Hospital of Athens, Athens, Greece.

A

B

C

Tentorium hypoplasia with partial occipital lobe herniation

E. Thomaere1, S. Schepers2, B. Termote2, R. Vanwyck2, G. Souverijns2

A 50-year-old woman was referred by her general practitioner to our hospital with complaints of vertigo and headaches. Magnetic resonance (MR)-imaging, axial (Fig. A) and coronal (Fig. B) T1 weighted, revealed a partial right occipital lobe herniation (inferior precuneus), secondary to a congenital focal hypoplasia of the tentorium, located on the anterosuperior part of the cerebellum, a rare incidental finding.

Comment

The tentorium cerebelli is the second largest dural reflection. It extends horizontally between the cerebellum and the cerebral hemispheres and thereby divides the cranial cavity into supra-tentorial and infratentorial. Agenesis or hypoplasia of the tentorium cerebelli is usually associated with extensive central nervous sys-tem malformations, such as Dandy-Walker Syndrome and Arnold- Chiari Malformation.

Isolated hypoplasia has only been described in four cases before in adults. In the first case described by Tanohata a computed to-mography (CT) showed a focal hypoplasia of the left tentorium cer-ebelli with secondary protrusion of the temporal lobe into the supe-rior cerebellar and quadrigeminal cisterns. The symptoms of the patient (galactorrhea) were not related to the imaging findings.

The other three cases are described by Abi-Jaoudeh and Chevrette. In the first case CT and MRI investigations showed a hypoplasia of the right tentorial leaf with protrusion of the isthmus of cingulate gyrus and the medial occipitotemporal gyrus into the superior cerebellar cistern. In the second case a CT showed a small hypoplasia of the right tentorial leaf with discrete protrusion of the medial occipitotemporal gyrus and the parahippocampal gyrus into the cerebellar cistern. In the third case an angio-MRI for cere-bral aneurysm screening, showed a hypoplasia of the right tento-rial leaf with protrusion of the parahippocampal gyrus into the quadrigeminal cistern. In all three cases the symptoms were unre-lated to these imaging findings.

The most accepted hypothesis about the development of the tentorium is an abnormal fusion of the tentorium by Tanohata. The tentorium develops as the second dural reflection, first the medial parts develop and secondly the lateral parts. Afterwards the medial parts involute and the lateral parts, which consists of the caudo-lateral and the rostrolateral parts, fuse. Abnormal fusion of the lateral parts can possibly lead to this anomaly. Birth traumas or perinatal insults can also be considered as the cause of this

anomaly. Most probably symptoms depend on the severity and location of the tentorial defect, but the exact clinical significance is still uncertain.

It is important to recognise a tentorial hypoplasia and to know and report that there are no known clinical symptoms associated with this finding. In this case no treatment was needed, because the symptoms were unrelated to the imaging findings.

Reference

1. Abi-Jaoudeh N., Chevrette E.: Isolated hypoplasia of the tentorium. J Comput Assist Tomogr, 2006, 30: 131-134.

JBR–BTR, 2015, 98: 96

IMAGES IN CLINICAL RADIOLOGY

1. Department of Radiology, UZ Gasthuisberg, Leuven, 2. Department of Radiology, JESSA ziekenhuis, Hasselt, Belgium.

A

B

320-row-detector CT angiography findings in a case with myocardial bridging in the three main coronary arteries

S. Akay1, U. Bozlar2, Demirkol Sait3, M. Tasar2

A 46-year-old man with the complaining of palpitation was admitted to our hospital. Polymorphic ventricular early beats were observed on electro cardiography, The patient was referred to our department for coronary artery computed tomography angiography (CTA) for a probable congenital anomaly.

On coronary CTA performed by a 320-Row-Detector scanner; left main coronary artery was short. The myocardial bridging (MB) causes approxi-mately 75% luminal stenosis in a 2.5-cm segment was observed in the middle segment of left anterior descending artery (Fig. A, white arrow). The circumflex artery was continuing as the first obtus margin artery and this branch was separating to four branches in the middle part. All of these branches were coursing subepicardially in the middle and distal part (Fig. B, white arrowheads). The right main coronary artery was separating to two branches in the proximal part, and the thinner one was showing MB in its middle part (Fig. C, black arrow). This branch had a subepicar-dial course as well, in its distal part along the right atrium (Fig. C, black arrowheads).

Eventually, all the cardiologic examinations and imaging findings were reevaluated. Based on those findings, antiarrhythmic drug therapy was prescribed by cardiology department, and the patient was called for regu-lar follow-up.

Comment

CTA study was performed by using a 320 row-detector CT scanner (Aq-uilion One; Toshiba Medical Systems, Ottawara, Japan). Because of the heart rate of the patient was 60 beats per minute, we did not give beta-blocker drug. Prospective ECG-triggered dose modulation was used. The tube voltage was 120 kV. A biphasic injection of intravenous contrast was used and the total amount of 70 ml non-ionic contrast agent (Iohexol-350; GE Healthcare, USA) was injected into the antecubital vein at a flow rate of 5.0 ml, followed by a saline flush. In order to synchronise the arrival of

the contrast agent and the scan, bolus arrival was detected using automated peak enhancement detection in the ascending aorta using a threshold of 200 HU. An initial data set was reconstructed with a slice thickness of 0.5 mm and a reconstruction interval of 0.25 mm. The images were transferred to an image postprocessing workstation (Vitrea 2.0; Vital Images, Minnetonka, MN). Then axial raw, multiplanar, curved and rotated reformatted images were evaluated along the course of each coronary artery and major side branches, and thin-slab maximum-intensity projection images created at the workstation.

The clinical importance of the MB is controversial. Although the bridging is asymptomatic in most of the cases, it can rarely cause myocardial ischemia and related complications. While MB was observed in 0.5-2.5% of the catheter angiography studies, it was reported as ranging rates from 15% to 85% in the autopsy series. MB is seen most commonly in the middle segment of left descending artery. The other main coronary arteries and their branches are affected less frequently.

In 320-detector row CT, the entire heart is imaged with temporal uniformity. The temporal resolution of an multi-detector CT scanner reflects the ability to freeze cardiac motion, thus producing motion-free images. The 320-detector scanner has a standard temporal resolution of approximately 175 ms which is a very short time compared with the lower detector row CT scanners. Wide-area coverage multidetector CT, such as 320-detector row scanner, has enabled volumetric imaging of the entire heart free of stair-step artifacts at a single time point within one cardiac cycle.

MB is not a rare situation in the routine clinical practice. But bridging in all the three main coronary arteries is very uncommon. Multidetector coronary CTA, especially by using a 320-row-detector scanner is an effective and non- invasive imaging modality for understanding the normal coronary anatomy and detecting the congenital anomalies of the coronary arteries.

References

1. Möhlenkamp S., Hort W., Ge J., Erbel R.: Update on myocardial bridging. Circulation, 2002, 106: 2616-22.2. Hsiao E.M., Rybicki F.J., Steigner M.: CT coronary angiography: 256-slice and 320-detector row scanners. Curr Cardiol Rep, 2010,

12: 68-75.

JBR–BTR, 2015, 98: 97.

IMAGES IN CLINICAL RADIOLOGY

1. Dpt of Radiology, Sirnak Military Hospital, Sirnak, 2. Dpt of Radiology, 3. Dpt of Cardiology, Gulhane Military Medical School, Ankara, Turkey.

A

B

C

The single right coronary artery

E. Christiaanse1, D. Verdries2, K. Tanaka2, S. Carlier3, J. de Mey2

An asymptomatic obese 60-year-old man with an increased cardio-vascular risk profile (arterial hypertension and hypercholesterolemia) was checked prior to hip surgery. Resting ECG was normal. Exercise test was not possible. A dipyridamole myocardial perfusion scintigraphy showed inferolateral ischemia and a preserved left ventricular (LV) function. Coro-nary angiography demonstrated moderate atherosclerosis in the proximal segment of the right coronary artery (RCA) and a moderate stenosis in the distal segment of the posterolateral branch (Fig. A). The left anterior de-scending artery (LAD) and diagonal branches were visualized simultane-ously. No additional coronary ostium could be found. The contrast en-hanced cardiac CT was performed and it confirmed the absence of a left coronary ostium and a single RCA with two large right ventricular (RV) branches continuing to the left coronary artery (LCA). The first RV branch continued to the diagonal branch. The second RV branch continued to the LAD. Interestingly, in the anterior interventricular sulcus, the LAD had a bidirectional course proximally and distally, with proximally the emer-gence of an hypoplastic left circumflex artery (LCX) (Fig. B, C). None of the branches were running between the pulmonary artery and the ascending aorta, corresponding with a R-II A classification according to Lipton/ Yamanaka.

Comment

A single RCA is an extremely rare congenital anomaly with an estimat-ed incidence of 0.066% of the population. In a single coronary artery (SCR) by definition only one coronary artery originates from the aortic trunk and supplies the entire heart. The current classification system was introduced by Lipton et al. and modified by Yamanaka et al. In this classification the origin of the SCA from the right sinus of Valsalva is defined as ‘R’ and from the left sinus of Valsalva as ‘L’. The normal anatomical course of the SCA is defined as ‘I’. The type ‘II’ SCA provides the contralateral coronary artery and crosses the base of the heart to assume its inherent normal position. In the type ‘III’ SCA, after leaving the right coronary sinus of Valsalva LAD and LCX arise separately from proximal part of the artery. A further subdivision can be made according to the anatomical course. In type ‘A’ the main left or right coronary passes anterior to the pulmonary artery, ‘B’ passes between the aorta and pulmonary artery and ‘P’ poste-rior to the aorta. Yamanaka et al. distinguished a septal type ‘S’ with the course of the SCA through the interventricular septum and a combined type ‘C’.

Patients with a SCA are usually asymptomatic and the anomaly found incidentally during angiography. The clinical significance mainly depends on the course of the SCA in relation to the great arteries with a course between the pulmonary artery and the aorta being at risk to arrhythmias, ischemia and sudden death especially during exercise.

According to our knowledge, this is the first anatomical description of right ventricular branches of a single RCA continuing to a diagonal branch, a bidirectional LAD and a hypoplastic LCX.

Reference

1. Lipton M.J., Barry W.H., Obrez I., Silverman J.F., Wexler L.: Isolated single coronary artery: diagnosis, angiographic classification, and clinical significance. Radiology, 1979, 130: 39-47.

JBR–BTR, 2015, 98: 98.

IMAGES IN CLINICAL RADIOLOGY

1. Department of Radiology, ZorgSaam Zeeuws-Vlaanderen, Terneuzen, The Netherlands, 2. Department of Radiology, Universitair Ziekenhuis Brussel, Brussels, Belgium, 3. Department of Cardiology, CHU Ambroise Paré & Faculty of Medicine and Pharmacy, University of Mons, Mons, Belgium.

A

B

C

Symptomatic coracoclavicular joint

S. Lanotte1, S. Van Den Broeck1

A 47-year-old woman was referred to our radiology department for right shoulder arthroscanner evaluation because of anterior shoulder pain. She had previously been practising professional dance. Computed Tomography (CT) (Fig. A) did not show any shoulder pathology but meticulous analysis revealed the presence of a joint between the clavicle and the coracoid process with degenerative changes. In retrospect, the coracoclavicular joint was visible on plain radiographs (Fig. B).

A coracoclavicular joint arthrogram (Fig. C) confirmed a diarthrosis with articular cartilage covering both articulation surfaces surrounding by an articular capsule.

The patient was treated with local intraarticular injection of steroids under image guidance.

Comment

Coracoclavicular joint (CCJ) is a true synovial diarthrosis between the coracoid process of the scapula and the conoid tubercle of the clavicule. This relatively rare anatomical structure is reported by 0.04% to 3.0% of radiological inves-tigations. The presence of such a joint is mostly asymptom-atic but may cause shoulder pain. There are so far only 17 other reports of symptomatic CCJ but this incidence is likely underestimated due to underreported and underdiagnosed cases. Patients commonly suffer from diffuse shoulder pain radiating to the arm. This pain can be accompanied by paraesthesia and is aggravated by movement.

Plain radiographs are typical of a joint and reveal a bony outgrowth from the undersurface of the clavicle forming an articular surface with a tubercle on the dorsomedial surface of the coracoid process. Arthrograms typically show the two processes forming a real joint surrounded by a capsule and the two bony surfaces being covered by cartilage.

In conclusion, CCJ is an unusual cause of shoulder pain which can be recognised on plain radiographs and ought to be considered in the differential diagnosis of unexplained shoulder pains.

Reference

1. Singh V.M., Singh P.K., Trehan R., Thompson S., Pandit R., Patel V.: Symptomatic coracoclavicular joint: incidence, clinical significance and available management options. Int Orthop, 2011, 35: 1821-1826.

JBR–BTR, 2015, 98: 99.

IMAGES IN CLINICAL RADIOLOGY

1. Department Radiology and Medical Imaging, Clinique Saint-Luc, Bouge (Namur), Belgium.

A

B

C

Dear Editor in Chief,

Mesenteric panniculitis (MP) is a benign inflammatory condition of unknown etiology that involves the adipose tissue of the proximal mes-entery and has received consider-able attention over the last two de-cades. In a recent couple of papers published in volume 94 (2011) of JBR-BTR we successively tried to ex-plore the typical imaging findings and to prospectively study the preva-lence and natural course of MP (1,2).

The radiological CT diagnosis of MP is classically based on typical if not pathognomonic features clearly described and documented in the literature (1-6). They comprise the presence of a well-defined “mass effect” on neighboring structures (sign 1) constituted by mesenteric fat tissue of inhomogeneous higher attenuation than adjacent retro peri-toneal or mesocolonic fat (sign 2) and containing small soft tissue nodules (sign 3). These nodules are typically surrounded by a hypo-attenuated fatty “halo sign” (sign 4) and a hyperattenuating pseudo-capsule may also surround the all entity (sign 5). The last two signs (4 and 5) are considered inconstant but extremely specific.

We first concluded that the preva-lence of mesenteric panniculitis was much higher than previously re-ported ranging from 3, 42% to 7, 83% of patients – following the number of CT signs (3 or 5) considered as efficient to achieve the diagnosis – in which an abdominal CT was per-formed for various reasons or symp-toms (2). This high prevalence could probably explain the spontaneous association with the numerous and probably unrelated clinical situations

One of our most important but still controversial conclusions was that the value of MP in term of pre-dictivity of an associated neo plasm was probably not relevant. Indeed except a discretely higher prevalence found in patients presenting with bladder and/or prostatic neoplasms and with

found in the literature in many single cases reports. In fact we also con-cluded that the vast majority of cases could be considered as idiopathic, benign an asymptomatic (2). More-over follow-up CT studies showed a great stability of the CT findings of mesenteric panniculitis in about 85% of cases (2).

JBR–BTR, 2015, 98: 100-102.

LETTER TO THE EDITOR

A cAsE Of sympTOmATIc mEsEnTERIc pAnnIcuLITIs pREsEnTIng wITH unusuAL pOsITIvE fDg pET/cT nODuLAR cOmpOnEnTs: An ATypIcAL ImAgIng sTRATEgy wITH HIsTOpATHOLOgIc cORRELATIOn.B. Coulier1, I. Bueres2, F.C Deprez3, F. Richelle4, R. Rubay5, I. Gielen6, C. Fervaille7

Fig. 1. - Axial images (A-C) illustrate rather typical CT features of mesenteric pannicu-litis (MP) comprising mesenteric fat tissue of slightly higher attenuation than adjacent retroperitoneal or mesocolonic fat, a “mass effect” on neighbou ring structures, diffuse small soft tissue nodes and a subtle “halo sign” around several nodes. Nevertheless unusual and sometimes supracentimetric large nodes (black arrows) are also associated in the mesentere itself and in the preaortic retroperitoneal fat. Fortunately a previous abdominal CT performed 30 months earlier is available in our PACS (D-F) already

showing MP signs and subtle peripheral nodes. We first conclude that the process has an extremely slow growth reinforcing the high probability of a begnin process. For safety reasons a PET/CT is proposed (G-I) to reinforce the diagnosis of benign MP. Nevertheless the results appear very ambiguous and paradoxically pejorative with a marked hypermetabolism of the largest atypical nodules.

From: Department of 1. Diagnostic Radiology, 2. Gastroenterology, 4. Nuclear Medecine, 5. Visceral Surgery, Clinique St Luc, 5004 Bouge (Namur), Belgium, Department of 3. Diagnostic and Interventionnal Radiology, 7. Pathology, CHU Mont-Godinne, UCL, Belgium, 6. Institute of Pathology and Genetics, Gosselies, Belgium.Address for correspondence and reprint requests: Bruno Coulier, Department of Diagnostic Radiology, Clinique St Luc, Rue St Luc 8, 5004 Bouge (Namur) Belgium. E-mail: bcoulier.md@gmail.com

A B C

D E F

G H I

LETTER TO THE EDITOR — COULIER et al 101

Ultrasound and laboratory tests were normal. Abdominal CT demon-strated typical CT findings of mesen-teric panniculitis (MP) (Fig. 1A-C) but some unusual and sometimes supra-centimetric large nodules were also associated in the mesentery itself and in the preaortic retroperitoneal fat.

A previous abdominal CT per-formed 30 months earlier was available in our PACS (Fig. 1D-F). Typical MP was already clearly visible and had been described. The nodules described today were also already present but at this time there were infracentimetric. It was thus first concluded that the process has an extremely slow growth, reinforcing the high probability of a benign process. However for safe-ty reasons PET/CT was performed (Fig. 1 G-I) to reinforce the diagnosis of benign MP. Nevertheless the results appeared very ambiguous and para doxically pejorative with a marked hypermetabolism of the largest atypical nodules.

Celioscopic mesenteric biopsy was first performed and illustrated typical histopathologic features of MP.

Nevertheless we remained con-vinced that the histology did not correspond to the nodule that had been supposed to have been biop-sied. Indeed the fat content of the histologic specimen was too im-portant to correspond to the positive PET-CT dense nodule. In fact the histology was more corresponding to the more classical and usual inflammatory fat that constitutes most of the volume of the mesentery in cases of MP (Fig. 3 A, B). Effectively the nodule was still visible on a control CT performed two months later (Fig. 2B) and the scar of the unsuccessful surgical biopsy was clearly seen near it (Fig. 2A). The nodule was clearly visible during focused abdominal ultrasound (Fig. 2C, D) and pressure on the lesion was painful. Percutaneous biopsy under ultrasound guidance and supported by Conebeam CT revealed a benign histology of dense inflammatory fibrosis (Fig. 3C, D).

The patient was finally discharged with symptomatic treatment (anal-gesic and anti-inflammatory drugs)

This case demonstrates that while most cases of MP have pathog-nomonic CT appearances there are atypical presentations where more invasive diagnostic strategies are needed. Our case also offers the opportunity to illustrate different histological patterns of the same benign disease.

being useful to correctly exclude mesenteric tumoral involvement in doubtful or ambiguous cases (2).

Nevertheless even with an excel-lent experience in CT diagnosis of MP, unusual and ambiguous situa-tions may be found in several patients necessitating the deploy-ment of more unusual and/or inva-sive strategies to obtain an unam-biguous certified diagnosis.

So we also recently read with interest the paper of Garg (11) who performed CT-guided percutaneous administration of Spot sterile carbon stain to a single F-18 FDG positive mesenteric lymph node to allow identification during subsequent laparoscopic resection in a patient presenting with an atypical PET-positive mesenteric panniculitis.

We recently experienced a similar atypical and ambiguous case. A 68 year-old presented to the department of gastroenterology with complaints of fluctuating epigastric abdominal pain. Gastroscopy, upper abdominal

lymphoma the general prevalence of mesenteric panniculitis in our study didn’t significantly differ in our two “neoplastic” and “non neoplasic” large cohorts of patients (2). This important conclusion remains am-biguously controversial for several authors (7, 8) but nevertheless and recently been comforted by a large matched case-control study (9).

Another of our conclusions, which remained more speculative, was that the potential progression of MP to the heavy more debilitating retractile mesenteritis was extremely difficult to predict. It seemed to remain a very rare event that was not encountered in our rather long period of obser-vation (1,2). Therefore such a pro-gres sion may also be considered as doubtful. A real kinship or a complete independence between the two entities remains thus an open question which cannot be unequi-vocally answered.

Finally PET/CT as reported by Zissin (10) was also comforted as

Fig. 2. - On a control CT performed 2 months (A, B) after surgical exploration it clearly appeared (as suspected in the light of the histologic result) that the surgical biopsy whose trace was still perfectly visible (white arrow on A) had been made in the predominantly lipomatous portion of the MP. The most superficial supracentimetric hypermetabolic nodule was still present (B). During complementary abdominal Ultrasound this nodule appears as a hypoechoic, hyperattenuating and almost avascular mass (C). The nodule was successfully biopsied under combined ultrasound and ConeBeam CT guidance (black arrow on D).

A

C

B

D

102 JBR–BTR, 2015, 98 (2)

References

1. Coulier B.: Mesenteric panniculitis. Part 2: prevalence and natural course: MDCT prospective study. JBR-BTR, 2011, 94: 241-246.

2. Coulier B.: Mesenteric panniculitis. Part 1: MDCT – pictorial review. JBR-BTR, 2011, 94: 229-240.

3. Sabaté J.M., Torrubia S., Maideu J., et al.: Sclerosing mesenteritis: imag-ing findings in 17 patients. AJR, 1999, 172: 625-629.

4. Horton K.M., Lawler L.P., Fishman E.K.: CT findings in sclerosing mesenteritis (panniculitis): spectrum of disease. Radiographics, 2003, 23: 1561-1567.

5. Mata J.M., Inaraja L., Martin J., Olazabal A., Castilla M.T.: CT features of mesenteric panniculitis. Comput Assist Tomogr, 1987, 11: 1021-1023.

6. Okino Y., Kiyosue H., Mori H., et al.: Root of the small-bowel mesentery: correlative anatomy and CT features of pathologic conditions. Radio-graphics, 2001, 21: 1475-1490.

7. van Putte-Katier N., van Bommel E.F., Elgersma O.E., Hendriksz T.R.: Mesenteric panniculitis: prevalence, clinicoradiological presentation and 5-year follow-up. Br J Radiol, 2014, 87 (1044): 20140451.

8. Daskalogiannaki M., Voloudaki A., Prassopoulos P., et al.: CT evaluation of mesenteric panniculitis: prevalence and associated diseases. AJR, 2000, 174: 427-431.

9. Gögebakan Ö., Albrecht T., Osterhoff M.A., Reimann A.: Is mesenteric pan-niculitis truely a paraneoplastic phe-no menon? A matched pair analysis. Eur J Radiol, 2013, 82: 1853-1859.

10. Zissin R., Metser U., Hain D., Even-Sapir E.: Mesenteric panniculitis in oncologic patients: PET-CT findings. Br J Radiol, 2006, 79: 37-43.

11. Garg V., Alvarado N., Raju R.: CT-guided percutaneous administration of Spot sterile carbon stain to a single F-18 FDG positive mesenteric lymph node to allow identification during subsequent laparoscopic resection. Abdom Imaging, 2014, 39: 1134-1136

12. Nomura Y., Naito Y., Eriguchi N. et al.: A case of IgG4-related sclerosing mesenteritis. Pathol Res Pract, 2011, 207: 518-521.

13. Chen T.S., Montgomery E.A.: Are tumefactive lesions classified as sclerosing mesenteritis a subset of IgG4-related sclerosing disorders? J Clin Pathol, 2008, 61: 1093-1097.

14. Belghiti H., Cazals-Hatem D., Couvelard A., Guedj N., Bedossa P.: Sclerosing mesenteritis: can it be a IgG4 dysimmune disease? Ann Pathol, 2009, 29: 468-474.

15. Coulier B., Montfort L., Beniuga G., Pierard F., Gielen I.: Small bowel ob-struction caused by peritoneal immu-no globulin g4-related disease mim-icking carcinomatosis: case report. Korean J Radiol, 2014, 15: 66-71.

not mandatory for the diagnosis but may be of valuable assistance. More-over, not every entity with increased IgG4+ plasma cells and a high histo-logic IgG4/IgG ratio can be con-sidered to belong to the IgG4-RD spectrum. Most authors agree that a definitive diagnosis of IgG4-RD re-quires histo logic confirmation in-clud ing the presence of characteristic histo pathologic features such as dense lymphoplasmacytic infiltra-tion, stori form fibrosis, obliterative phlebitis, mild to moderate eosino-philia and germinal center formation along with immunohistochemical staining demonstrating an increased number of IgG4+ cells (15). Except the ab sence of obliterative phlebitis most of these classical histo patho-logic were undoubtedly found in the re ported case.

One thing is certain, future research is still needed to continue to affirm or refute the possibility that mesenteric panniculitis MP (alias sclerosing mesenteritis) be-longs to the ever growing group of IgG4-RD.

Until now the precise etiology of MP (the more accurate histological name being “sclerosing mesen-teritis”) remains unknown. However, as shown in the reported cases, many histological features such as the presence of lymphocytic infil-trates, and to a lesser degree of plasmocytic and eosinophilic reac-tions in “active” nodules and but also the development of massive fibrosis which succeeds when the lesions become more quiescent is not unlike the appearance of auto-immune diseases.

In this autoimmune hypothesis the possibility that MP could repre-sent another additional expression of the constantly increasing group of IgG4 related diseases (IgG4-RD) has already been suggested by several authors (12, 13) but rejected by others (14) because of a relatively low rate of IgG4+ plasma cells during immunohistochemistry.

Nevertheless today there is no published international consensus on the diagnostic criteria for IgG4-RD (15). A raised serum IgG4 level is

Fig. 3. - Photomicrographs. A (Hematoxylin and Eosin, × 5) & B (Hematoxylin and Eosin, × 20) illustrate the typical histopathology of the predominantly adipous component of the mesenteric panniculitis obtained during celioscopy. The mesenteric adipous tissue (white stars) is dislocated by fibrous bands (black star on a). A minimal inflammatory reaction is present with a lymphoid focus (black arrow). Figure B shows an area of adiponecrosis with macrophagic reaction (white arrow) surrounded by fibrosis (black stars). C, D illustrate the benign histopathology of the hypermetabolic nodule biopsied under ConeBeam CT guidance. Extensive fibrosis (grey star) is intermingled with a predominantly lymphocytic inflammatory reaction. Some plasmo-cytes and eosinophiles are also present on figure D.

A

C

B

D

JBR–BTR, 2015, 98: 103.

ANNOUNCEMENT

CHARTER YOUNG RADIOLOGISTS SECTION (YRS)

• The Young Radiologists Section (YRS) is a subdivision of the Belgian Society of Radiology

(BSR) dedicated to residents + recently (< 5 years) graduated radiologists, who are

member of the BSR.

• Representation by each of the 7 universities (Universiteit Antwerpen, Université

Catholique de Louvain, Université de Liège, Universiteit Gent, Université Libre de

Bruxelles, Katholieke Universiteit Leuven, Vrije Universiteit Brussel)

• Two Presidents (one Dutch + one French speaking), Two Vice-Presidents (one Dutch + one

French speaking) and One Secretary. The two Vice-Presidents and the Secretary are

elected every year. The positions of President will be filled in by the two Vice-Presidents

of the previous year.

• Key Goals:

• Education: standardized education and examination between the universities

• Communication: communication with external organizations (e.g. ECR)

• Scientific: co-organization of the Annual Symposium of the BSR

• Communication language: English

• By general consensus 4-5 meetings/year in a central location in Belgium with subsequently

a report of the meeting by the Secretary. Every member of the YRS is welcome to join the

meetings of the Scientific Board of the BSR (most interesting for the members who

participate in the (co)organization of the annual symposium of the BSR)

• The Presidents, Vice-Presidents and Secretary will be the communication bridge between

the General/Scientific Board of the BSR and the members of the YRS.