P212. Inter- and intra-observer variation in histologic diagnosis of microscopic colitis and subgroups
A.-M.K. Fiehn1, C. Bjørnbak2, M. Warnecke3, P.J.H. Engel1, L.K. Munck2, 1Roskilde Sygehus, Dept. of Pathology, Roskilde, Denmark, 2Køge sygehus, Dept. of Int. Med., Køge, Denmark, 3Slagelse Sygehus, Dept. of Pathology, Slagelse, Denmark
The diagnosis of microscopic colitis (MC) is based on histological findings and includes collagenous colitis (CC) and lymphocytic colitis (LC). MC is defined by a mixed inflammatory infiltrate in lamina propria, intraepithelial lymphocytosis and surface epithelial damage. In LC >20 lymphocytes/100 epithelial cells is required and CC is characterized by an abnormal subepithelial band >10 my. A number of patients with chronic diarrhea do not meet the histological criteria and the terms MCincomplete (MCi) has been suggested. We investigated the inter- and intra-observer agreement on MC.
A single HE-stained slide from biopsies from 5 groups of 25 patients each with CC, LC, MCi, IBD/nonspecific reactive change and normal were randomly selected and blinded. Three pathologists independently reviewed the slides. The slides were relabeled and reinterpreted four months later. Inter- and intra-observer agreement was evaluated by kappa statistics where 0.41–0.60, 0.61–0.80 and 0.81–1.00 are considered to reflect moderate, good and very good agreement.
The overall agreement at first and second assessment was 59% and 67% for the five categories and 88% and 86% for three categories (MC/MCi versus IBD/nonspecific reactive change versus normal). Kappa values for each pair of pathologists varied between 0.60–0.85 and 0.81–0.89 for categorization into five and three diagnostic categories, respectively, corresponding to moderate to good and very good agreement. Intra-observer agreement was good to very good for the five diagnostic categories and very good for MC/MCi versus non-MC. Overall agreement was very high for discrimination between MC/MCi and non-MC, but intra- and inter-observer agreement on MCi was weak.
The incidence of MC is increasing, the quality of life in patients with active MC is poor, and an effective treatment available. The diagnosis relies on specific histological findings and it is therefore important that the pathological criteria are well defined and reproductive. Our results show that both the inter-observer and intra-observer agreement is high for discriminating between MC/MCi and non-MC, while agreement on MCi is weak. The histological criteria for MC subgroups including MCi need to be reevaluated and consensus obtained.