Effect of RECIST revision on classification of target lesions and overall response in advanced gastric cancer patients

Abstract

The Response Evaluation Criteria in Solid Tumors (RECIST) was revised in 2009, based on a large dataset of 6512 patients from 16 trials. However, no gastric cancer patients were included in those data. The purpose of this study was to clarify the difference between RECIST version 1.0 and version 1.1 in advanced gastric cancer. From 2004 to 2009, 129 consecutive patients with advanced gastric cancer received S-1 plus cisplatin as first-line treatment at the National Cancer Center Hospital East. Ninety-seven of these patients who had had baseline and post-treatment computed tomography scans performed were included in this study. Measurements of tumors were conducted retrospectively. At the baseline of first-line chemotherapy, 172 lymph nodes in 54 patients were considered to be candidate target lesions by RECIST version 1.0. However, only 38% of the lymph nodes were classified as target lesions by RECIST version 1.1, with 47% classified as non-target lesions and 15% classified as non-pathological. By RECIST version 1.0, the proportion of patients with target lesions at the baseline of first-line chemotherapy was 67% (65/97), and this percentage was significantly reduced according to RECIST version 1.1 (53%; 51/97) (McNemar's exact test, P < 0.001). The findings at the baseline of second-line chemotherapy were similar (reduced from 62 to 49%; McNemar's exact test, P = 0.002). Overall response rates of first-line chemotherapy were 52% (34/65) according to RECIST version 1.0 and 55% (28/51) according to version 1.1. The revision of RECIST significantly reduced the proportion of patients classified with target lesions at the baselines of first-line and second-line chemotherapies. No obvious difference in overall response rates was observed.