RESUMEN
Background: Perioperative chemotherapy is an established treatment of advanced gastric cancer patients. Treatment selection is based on clinical staging (cT). We aimed to establish and validate a prognostic score including clinical and molecular factors, to optimize treatment decisions for these patients. Patients and methods: We analyzed 626 carcinomas of the stomach and of the gastro-esophageal junction from two academic centers including primarily resected and pre-/perioperatively treated patients. Patients were divided into a training (N = 269) and validation (N = 357) set. Expression of 11 target genes was measured by quantitative PCR in resected tumors. A risk score to predict overall survival (OS) was generated and validated. Intra-tumoral heterogeneity was assessed by analyzing 50 tumor areas from 10 patients. Results: A risk score including the expression of CCL5, CTNNB1, EXOSC3 and LZTR1 and the clinical parameters cT, tumor localization and histopathologic type suggested two groups with a significant difference in OS [hazard ratio (HR) 0.30; 95% confidence interval (CI) 0.17-0.52]. The risk score was successfully validated in an independent cohort (HR 0.32; 95% CI 0.21-0.51; P < 0.001) as well as in subgroups of primarily resected (HR 0.30; 95% CI 0.17-0.54; P < 0.001) and pre-/perioperatively treated patients (HR 0.37; 95% CI 0.17-0.81; P = 0.009). A significant difference in OS of high- and low-risk patients was also found in primarily resected patients with intestinal (HR 0.45; 95% CI 0.23-0.90; P = 0.020) and nonintestinal-type carcinomas (HR 0.1; 95% CI 0.02-0.42; P < 0.001). Intra-tumor heterogeneity analysis indicated a classification reliability of 95% for a supposed analysis of three biopsies. Conclusion: The identified risk score could substantially contribute to an improved management of gastric cancer patients in the context of perioperative chemotherapy.