Concordance between pathologists in the diagnosis of gastric atrophy according with the OLGA system

BACKGROUND: The intestinal gastric cancer is preceded by a sequence of pathological changes whose link is mucosal atrophy. The modified Sydney system for atrophy is a parameter not reproducible among pathologists. AIM: To know the interobserver variability using the OLGA system (Operative Link on Gastritis Assessment). METHODS: We selected 116 histologic slides. Sixty cases of both types of atrophy and 56 without atrophy were included. Three general pathologists, interested in gastrointestinal biopsies independently review the slides and established a diagnosis. For statistical analyses we employed singles frequencies in order to describe the individual diagnosis and the kappa test for evaluate the concordance between 2 and 4 observers with 2 and 3 categories. RESULTS: The global concordance has a kappa index of 0.48 (IC 95% 0.4-0.57). When we compared two pathologists the kappa index varies from 0.82(IC 95% 0.73-0.91) to 0.36 (IC 95% 0.22-0.5). The consensus among three pathologists was achieved in 25 out 30 slides in the metaplastic variety and 11 out 30 for the non-metaplastic type. The concordance for the atrophy scale has a kappa index between 0.2 and 0.5. CONCLUSION: The problematic atrophic evaluation with the Sydney system justify every effort to improve the interobserver evaluation. The OLGA system seems reproducible, although laborious,it requires a careful application, but with daily practice it could be applied easier. The clinician acceptation becomes crucial.