Evaluation of gastric biopsies for neoplasia: differences between Japanese and Western pathologists

Abstract

Cited variations in the evaluation of gastric endoscopic biopsies for neoplasms between pathologists in Japan and those in the United States and Europe (the West) may have stemmed from several causes. The five-tiered group classification of the Japanese Research Society for Gastric Cancer (JRSGC) for interpretation of biopsies is not used in the West. Some differences may also exist in the morphologic criteria to reach a diagnosis of dysplasia or carcinoma. The goals of this study were to test the Western and Japanese classifications of gastric dysplasia and adenocarcinoma and to assess the differences between four Japanese and seven Western pathologists. One hundred biopsies, 20 from each of the five categories of the JRSGC scheme as determined by one observer, were collected. The Japanese observers used the JRSGC system, expressed in Roman numerals, whereas Western pathologists used a five- or six-tiered scheme expressed in diagnostic terms. Pairwise agreement was evaluated using k statistics within both groups. Consensus diagnosis on each biopsy was accepted as the opinion of the majority. The sensitivity and specificity of each reviewer for a certain diagnosis were also assessed. The intragroup agreements were moderate for both the Japanese (mean k = 0.663) and the Westerners (mean k = 0.652). The pairwise agreements between Japanese and Western observers were low (mean k = 0.542). Overall, the sensitivity was low for all Japanese observers for the diagnosis of dysplasia (38.7% vs 92.5%), and the sensitivity for the diagnosis of adenocarcinoma was high in both groups but higher among the Japanese observers (93.9% and 85.2%, respectively). Overall, the Japanese-Western interobserver agreement was moderate. The JRSCG scheme did not translate into higher interobserver agreement among Japanese observers. The sensitivity for the diagnosis of gastric adenocarcinoma was high for both groups, but the specificity was low among the Japanese. The cause seemed to be centered around the diagnosis of dysplasia in the Western system, which was a lesion frequently interpreted as carcinoma in Japan because of the different definitions of carcinoma in each system. Such a discrepancy might be important because it may explain some of the differences in the prevalence and prognosis of early gastric cancer between Japan and the West. An international effort is needed to harmonize morphologic criteria and analyze whether therapeutic consequences may stem from such discrepancies.