TY - JOUR T1 - ROC-king onwards: intraepithelial lymphocyte counts, distribution & role in coeliac disease mucosal interpretation JF - Gut JO - Gut SP - 2080 LP - 2086 DO - 10.1136/gutjnl-2017-314297 VL - 66 IS - 12 AU - Kamran Rostami AU - Michael N Marsh AU - Matt W Johnson AU - Hamid Mohaghegh AU - Calvin Heal AU - Geoffrey Holmes AU - Arzu Ensari AU - David Aldulaimi AU - Brigitte Bancel AU - Gabrio Bassotti AU - Adrian Bateman AU - Gabriel Becheanu AU - Anna Bozzola AU - Antonio Carroccio AU - Carlo Catassi AU - Carolina Ciacci AU - Alexandra Ciobanu AU - Mihai Danciu AU - Mohammad H Derakhshan AU - Luca Elli AU - Stefano Ferrero AU - Michelangelo Fiorentino AU - Marilena Fiorino AU - Azita Ganji AU - Kamran Ghaffarzadehgan AU - James J Going AU - Sauid Ishaq AU - Alessandra Mandolesi AU - Sherly Mathews AU - Roxana Maxim AU - Chris J Mulder AU - Andra Neefjes-Borst AU - Marie Robert AU - Ilaria Russo AU - Mohammad Rostami-Nejad AU - Angelo Sidoni AU - Masoud Sotoudeh AU - Vincenzo Villanacci AU - Umberto Volta AU - Mohammad R Zali AU - Amitabh Srivastava Y1 - 2017/12/01 UR - http://gut.bmj.com/content/66/12/2080.abstract N2 - Objectives Counting intraepithelial lymphocytes (IEL) is central to the histological diagnosis of coeliac disease (CD), but no definitive ‘normal’ IEL range has ever been published. In this multicentre study, receiver operating characteristic (ROC) curve analysis was used to determine the optimal cut-off between normal and CD (Marsh III lesion) duodenal mucosa, based on IEL counts on >400 mucosal biopsy specimens.Design The study was designed at the International Meeting on Digestive Pathology, Bucharest 2015. Investigators from 19 centres, eight countries of three continents, recruited 198 patients with Marsh III histology and 203 controls and used one agreed protocol to count IEL/100 enterocytes in well-oriented duodenal biopsies. Demographic and serological data were also collected.Results The mean ages of CD and control groups were 45.5 (neonate to 82) and 38.3 (2–88) years. Mean IEL count was 54±18/100 enterocytes in CD and 13±8 in normal controls (p=0.0001). ROC analysis indicated an optimal cut-off point of 25 IEL/100 enterocytes, with 99% sensitivity, 92% specificity and 99.5% area under the curve. Other cut-offs between 20 and 40 IEL were less discriminatory. Additionally, there was a sufficiently high number of biopsies to explore IEL counts across the subclassification of the Marsh III lesion.Conclusion Our ROC curve analyses demonstrate that for Marsh III lesions, a cut-off of 25 IEL/100 enterocytes optimises discrimination between normal control and CD biopsies. No differences in IEL counts were found between Marsh III a, b and c lesions. There was an indication of a continuously graded dose–response by IEL to environmental (gluten) antigenic influence. ER -