Background: In Western countries, the response of gastric cancer to chemotherapy is evaluated by assessing measurable metastatic lesions (MMLs) according to the response evaluation criteria in solid tumors (RECIST). In Japan, the response of primary lesions is assessed according to local Japanese criteria. We compared the response to chemotherapy as evaluated by these two sets of criteria.
Methods: Patients with unresectable, advanced gastric cancer who had primary lesions and had received first-line chemotherapy were studied. Responses of MMLs were evaluated with RECIST. Responses of primary lesions were evaluated with the Japanese criteria. Median survival times (MSTs) were compared according to treatment response by each set of criteria.
Results: Data from 341 patients were analyzed. Of the 242 patients with MMLs, 108 were MML responders and 134 were MML nonresponders. MST was significantly longer in MML responders (293 days; 95% confidence interval [CI], 244-342) than in MML nonresponders (159 days; 95% CI, 127-191; P < 0.0001). According to the Japanese criteria, there were 128 primary-lesion responders and 213 primary-lesion nonresponders. MST was significantly longer in responders (304 days; 95% CI, 266-342) than in nonresponders (168 days; 95% CI, 143-193, P < 0.0001). Of the 99 patients without MMLs, 26 were primary-lesion responders and 73, primary-lesion nonresponders; MST was significantly longer in the former (300 days; 95% CI, 266-334) than in the latter group (173 days; 95% CI, 111-235; P = 0.019).
Conclusion: The responses of primary lesions according to the Japanese criteria and the responses of MMLs according to the RECIST were both significantly related to the MST. Use of the RECIST alone might bias the evaluation of treatment response because response cannot be evaluated in patients without an MML.