Table-S1: Basic information of series used for estimating fractions of tumor microenvironment cells; Table-S2: Relative Fractions of tumor microenvironment cells of 1524 gastric cancer patients with overall survival informations; Table-S3: Kmeans cluster and Hazard ratios of tumor microenvironment cells in gastric cancer; Table-S4: Cellular interaction of tumor microenvironment cells; Table-S5: Clinical infromation, TMEcluster, TMEscore and fractions of Tumor microenvironment cells of GC patients in the ACRG cohort; Table-S6: P-values of Pairwise comparison ; Table-S7: Differentially expressed genes of TMEclusters in the ACRG cohort; Table-S8: TME signature genes A and B; Table-S9: Functional and pathway enrichment analyses (Gene Ontology-Biological process) of TME signature genes A and B; Table-S10: Other gene Signatures enrolled in this study; Table-S11: Enrichment analysis of TME gene cluster A and C (Hallmark gene sets of Broad Institude); Table-S12: TMEscore of all gastric cancer patients with survival informations; Table-S13: TMEscore of patients from TCGA with survival informations; Table-S14: TMEscore of patients treated with immunotherapy
ARTICLE ABSTRACT
Tumor microenvironment (TME) cells constitute a vital element of tumor tissue. Increasing evidence has elucidated their clinicopathologic significance in predicting outcomes and therapeutic efficacy. Nonetheless, no studies have reported a systematic analysis of cellular interactions in the TME. In this study, we comprehensively estimated the TME infiltration patterns of 1,524 gastric cancer patients and systematically correlated the TME phenotypes with genomic characteristics and clinicopathologic features of gastric cancer using two proposed computational algorithms. Three TME phenotypes were defined, and the TMEscore was constructed using principal component analysis algorithms. The high TMEscore subtype was characterized by immune activation and response to virus and IFNγ. Activation of transforming growth factor β, epithelial–mesenchymal transition, and angiogenesis pathways were observed in the low TMEscore subtype, which are considered T-cell suppressive and may be responsible for significantly worse prognosis in gastric cancer [hazard ratio (HR), 0.42; 95% confidence interval (CI), 0.33–0.54; P < 0.001]. Multivariate analysis revealed that the TMEscore was an independent prognostic biomarker, and its value in predicting immunotherapeutic outcomes was also confirmed (IMvigor210 cohort: HR, 0.63; 95% CI, 0.46–0.89; P = 0.008; GSE78220 cohort: HR, 0.25; 95% CI, 0.07–0.89; P = 0.021). Depicting a comprehensive landscape of the TME characteristics of gastric cancer may, therefore, help to interpret the responses of gastric tumors to immunotherapies and provide new strategies for the treatment of cancers.
