Identification of An Immune Signature Predicting Prognosis Risk of Patients in Gastric Cancer

Background: Gastric cancer is a common but lethal cancer owing to deficient in effective treatment. Substantial evidences have proved that immune infiltration plays a key role in progression of gastric cancer. This study aimed to establish a signature based on immune related genes that can predict clinical outcomes and therapeutic efficacy. Methods The expression data from The Cancer Genome Atlas database and 4617 immune related genes from previously published 160 immune gene sets were collected for development and validation of the signature. Cox proportional hazard regression model was used to construct the signature. The reliability and forecasting ability were evaluated by two independent datasets from GEO. Results A gene model consisting of 47 immune related genes was used as our signature. Risk scores were calculated based on the coefficient and the expression level of each gene in this model. The low risk score group had an obviously favorable prognosis than the other group in all cohorts. Both of univariate and multivariate analysis suggested that our immune gene signature was an independent prognostic factor. Single sample gene Set Enrichment Analysis (ssGSEA) revealed that high risk score was associated with high Th17 cell infiltration, low mast cell and pro- angiogenesis immune cell infiltration. More importantly, patients with high risk score presented high tumor mutation burden (TMB), which is an essential element for predicting therapeutic efficacy of immune check point inhibitor. Conclusion This signature is a promising tool to predict prognosis and screen out population who can get benefit from immune check point inhibitor.