BACKGROUND: We conducted a retrospective study to investigate the adequacy of the Efficacy Criteria for Primary Lesions in the Japanese Classification of Gastric Cancer (Japanese criteria) for evaluating the anti-tumor efficacy of chemotherapies and the relationship between tumor regression and the prognosis of gastric cancer.METHODS: The data for 90 patients with inoperable ad-vanced gastric cancer who received various chemotherapies, consisting of fluorinated pyrimidines and cisplatin, were retrospectively analyzed. Based on the Japanese criteria, we investigated the efficacy of the chemotherapies and the relationship between the response in primary lesions and survival. We also compared the efficacy of chemotherapies evaluated by the Japanese criteria to that evaluated by the WHO criteria.RESULTS: All 90 patients were evaluable by the Japanese criteria. The overall response rate was 53.3% (Partial response [PR] in 48 patients and no change + progressive disease [NC + PD] in 42 patients). The primary lesions were classified as measurable (a-lesions) in 27 patients, evaluable but not measurable (b-lesions) in 31 patients, and diffusely infiltrating (c-lesions) in 32 patients. Overall median survival time (MST) was 9.4 months. The MSTs of the responders and non-responders were 12.6 and 7.8 months, respectively. In contrast, by the WHO criteria, 49 patients (54.4%) were evaluable; the other 41 patients had gastric primary lesions alone but were not measurable by WHO criteria. The overall response rate was 67.3% (33/49), and overall MST was 9.4 months. The MSTs of the responders evaluated by both sets of criteria were both 12.6 months.CONCLUSIONS: We suggest that the Japanese criteria are useful for evaluating the anti-tumor effect of gastric cancer chemotherapies and that prospective studies to reconfirm their usefulness are warranted in Japan, and in Western countries.