Pathology findings and validation of gastric and esophageal cancer cases in a European cohort (EPIC/EUR-GAST)

Objective. Cardia, non-cardia and intestinal and diffuse subtypes of gastric cancer may have different trends and etiological factors. However, the available information is not always collected in population cancer registries, and heterogeneous criteria have been applied for the histopathological classification of tumors. We describe the pathological features of incident gastric and esophageal cancers identified within the European Prospective Investigation into Cancer and Nutrition (EPIC). Material and methods. In an investigation on gastric and esophageal cancer (EUR-GAST) in the EPIC project, a validation study of diagnoses reported by EPIC centers was conducted by a European panel of pathologists. Original pathology reports, stained slides of tumors and the respective paraffin blocks were requested from the centers. Results. The whole series encompassed 467 cancer cases (gastric and esophageal cancers). Material was available for histopathological validation in 263 cases (56%); in the remaining cases, information was retrieved from the original reports (n = 110; 24%) or codes provided by the EPIC centers (n = 94; 20%). Among cases submitted to histopathological validation reported originally as unknown histotype or unknown site, a specific diagnosis was made in 95% and 74% of the cases, respectively. In cases for which only the original reports were available, the respective percentages were 46% and 67%. Gastric adenocarcinomas were classified according to site (cardia (29.4%), non-cardia (48.2%) and unknown (22.4%)) and histological type (intestinal (33.4%), diffuse (33.7%) and mixed, unclassified or unknown (32.9%)). Frequency of cardia was higher in Northern countries (35%) than in Mediterranean countries (18%). Conclusions. In addition to providing epidemiological data within the EPIC cohort on gastric and esophageal adenocarcinomas, the results reported here confirm the relevance of a validation study, notably for multicenter studies.