Tumor Mutation Burden: A Predictive Biomarker for Gastric Cancer Immunotherapy
Abstract Background: Few studies have focused on the underlying relationship between the prognosis of tumor mutation burden (TMB) and immune cell infiltration in gastric cancer (GC). This study aims to explore the relationship among TMB and various components in tumor microenvironment (TME). Methods: The transcription profiles and somatic mutation data of 375 tumor and 32 normal samples were obtained from TCGA. The specific mutation information was summarized and visualized with waterfall chart, then number of TMB per million bases of each GC sample was calculated. Immune/stromal scores and tumor purity were calculated by the ‘ESTIMATE’ package, and the fractions of 22 immune cells in each sample were evaluated by CIBERSORT algorithm. Finally, Lass regression analysis was utilized to generate a prognostic scoring signature with TCGA cohort as the training set, while GES84437 cohort as the validation set. Results: Higher TMB indicated favorable overall survival (OS, P = 0.043),better disease specific survival (P = 0.029), and longer progression free interval (P = 0.004). TMB was positively correlated with MSI and tumor purity, while negatively associated with immune/stromal scores. Moreover, TMBhigh group has lower T cells CD4 memory resting (P < 0.001) and T cells regulatory (P < 0.001), and more T cells CD4 memory activated (P < 0.001) and T cells follicular helper (P = 0.009). More importanly, the infiltration of dendritic cells activated predicted a worse OS, while T cells CD4 memory activated and T cells follicular helper meant a better OS. Finally, a nomogram combined TMB-related signature with clinicopathologic variables can successfully predict the OS with high accuracy and efficiency.Conclusion: TMB can effectively reveal the immune infiltration status in TME of GC, and might serve as a prognostic classifier for individualized treatment of clinical decision-making.