Interobserver variation in the histopathological assessment of malt/malt lymphoma: towards a consensus

Background: The classification of mucosa associated lymphoid tissue (MALT) lymphoma is based on characteristic morphologic and immunophenotypic patterns, with distinctive chromosomal aberrations. The critical first step is diagnosis on evaluating H&E-stained sections. We performed an inter-observer study to determine the degree of agreement among pathologists in evaluating gastric lymphocytic infiltrates for MALT lymphoma. Methods: A set of 41 H&E-stained gastric sections (36 endoscopic biopsies and 5 surgically resected sections) that ranged from simple gastritis to primary gastric lymphoma was reviewed separately and independently by 17 participants including hematopathologists, pathologists with a special interest in gastrointestinal pathology, and general pathologists. The participants were from the United States, Europe, and Japan. Results were entered into a standardized data collection form and the results were analyzed using K statistics. Monte Carlo simulation was used to adjust for multiple biases. Results: Overall, interobserver reproducibility in the morphologic evaluation of gastric MALT was suboptimal. The Kappa statistic was 0.3 for simple gastritis, low-grade MALT and for high grade MALT lymphoma. Monte Carlo simulation suggested that the degree of disagreement was directly related to the pathologist's experience in evaluating gastric biopsies for MALT lesions. However, after conjointly reviewing all cases, the Houston workshop agreed on findings that would increase the reproducibility of diagnosis, especially for pathologists with limited experience with this disease. These included the availability of macroscopic data, extensive sampling, the presence of lymphoepithelial lesions, immunophenotyping and particularly abnormal mucosa localization of B-cells, in addition to other molecular finding such as monoclonality and translocationt (11;18). The group also agreed on the need for standardizing the terminology currently used to facilitate future comparison between studies. Conclusions: Though the study shows poor agreement on morphologic MALT lymphoma categorization, the Houston workshop suggested recommendations that should increase the diagnostic accuracy and reproducibility of MALT lymphoma diagnosis. A follow up workshop will be organized to measure the diagnostic reproducibility for MALT lymphoma using the suggested recommendations. (C) 2004 Elsevier Inc. All rights reserved.