MSI-GC-01: Individual patient data (IPD) meta-analysis of microsatellite instability (MSI) and gastric cancer (GC) from four randomized clinical trials (RCTs)

66 Background: In CLASSIC and MAGIC, MSI was a good prognostic factor, and adjuvant/perioperative chemotherapy had null/detrimental effect. Given the low prevalence of MSI in GCs and its association with other good prognostic variables, larger datasets are needed to draw more robust evidences on its specific prognostic/predictive impact. Methods: This was a multinational IPD meta-analysis of resectable GC pts enrolled in MAGIC, CLASSIC, ARTIST, ITACA-S. Data on pts’ demographics (age, sex, and race), primary site (stomach versus junctional), histotype (intestinal vs. other), T/N stage (7th TNM), treatment received (multimodal therapy vs. surgery alone) and MSI were pooled. Univariable and multivariable associations with disease-free survival (DFS)/overall survival (OS) were assessed. The predictive role of MSI according to treatment received was assessed overall and in the 2 RCTs with a surgery alone arm (MAGIC+CLASSIC). Results: We pooled 1,552 pts with available MSI status: 121 (7,8%) were MSI, 572 Caucasian/980 Asian. In MSI versus MSS subgroups, 5-y DFS was 71.8% (95% CI: 63.8-80.7%) versus 52.3% (49.6-55.0%) (HR = 0.50, 95% CI 0.35-0.72; p < 0.001); 5-y OS 77.4% (69.9-85.8%) versus 59.2% (56.6-62.0%) (HR = 0.50, 95% CI 0.34-0-74; p < 0.001). In multivariable analyses, MSI was independently associated with DFS (HR = 0.48 [0.33-0.70]; p < 0.001) and OS (HR = 0.48 [0.29-0.81]; p = 0.005), as T/N/race/treatment. Conclusions: In resectable primary GC, MSI is an independent good prognostic marker that should be adopted as stratification factor in future RCTs. Chemotherapy omission and/or immune checkpoint blockade should be prospectively investigated in MSI-high GCs according to the clinically-defined risk of relapse. [Table: see text]