Identification and validation of stromal-tumor microenvironment-based subtypes tightly associated with PD-1/PD-L1 immunotherapy and outcomes in patients with gastric cancer

Immunotherapies targeting programmed cell death 1 (PD-1) and programmed death-ligand 1 (PD-L1) have been approved for gastric cancer (GC) patients. However, a large proportion of patients with T-cell-inflamed tumor microenvironment do not respond to the PD-1/PD-L1 blockade. The stromal component of the tumor microenvironment has been associated with immunotherapy. This study aims to explore the clinical significance of the non-immune cells in the tumor microenvironment and their potential as biomarkers for immunotherapy. A total of 383 patients with GC from the Cancer Genome Atlas (TCGA) cohort, 300 patients with GC from the GSE62254 cohort in Gene Expression Omnibus (GEO) were included in the study. A stromal score was generated using the ESTIMATE algorithm, and the likelihood of response to PD-1/PD-L1 immunotherapy of GC patients was predicted using the TIDE algorithm. The prognostic value of the stromal score from GC cases was evaluated by the Kaplan-Meier method and Cox regression analysis. Gene set enrichment analysis (GSEA) was also conducted. The stromal score showed significant differences in different molecular subtypes and T stages. Multivariate analyses further confirmed that the stromal score was an independent indicator of overall survival (OS) in the two cohorts. The low stromal score group showed higher tumor mutation burden (TMB) and micro-satellite instability (MSI), and was more sensitive to immune checkpoint inhibitor according to the TIDE algorithm. Activation of the transforming growth factor and epithelial-mesenchymal transition were observed in the high stromal score subtype, which is associated with T-cell suppression, and may be responsible for resistance to PD-1/PD-L1 therapy. BPIFB2 was confirmed as a hub gene relevant to immunotherapy. The stromal score was associated with cancer progression and molecular subtypes, and may serve as a novel biomarker for predicting the prognosis and response to immunotherapy in patients with GC.