Construction and evaluation of a prognostic risk assessment model of gastric cancer by using hypoxia features

In this study, mRNA expression of gastric cancer tissue and clinical data of patients in TCGA-STAD dataset were used, together with the hypoxia-related gene sets in the MsigDB database, to screen hypoxia-related differentially expressed genes (DEGs) in GC. Thereafter, univariate and multivariate Cox regression analyses were carried out on hypoxia-related DEGs. The optimal feature genes related to prognosis were obtained to construct a prognostic risk assessment model. According to the model, the riskScore of GC patients was measured, and GC samples were assigned into high- and low-risk groups in accordance with the median riskScore. Based on the Kaplan-Meier curve and Receiver operating characteristic curve, validity of the prognostic risk assessment model was measured. Gene set enrichment analysis was performed on the two risk groups through Gene set enrichment analysis software. The results revealed that in the high-risk group, 9 signaling pathways were remarkably activated in several terms, like focal adhesion, extracellular matrix receptor interaction, Cell adhesion molecules cams, Cytokine-cytokine receptor interaction, TGF-beta signaling pathway, NOD-like receptor signaling pathway, JAK-STAT signaling pathway, Toll-like receptor signaling pathway and MAPK signaling pathway. In combination with riskScore and clinical factors, univariate and multivariate Cox regression analyses verified the independence of the model. Meanwhile, a nomogram was constructed to predict the 1-, 3- and 5-year survival of GC patients. The calibration curve indicated that the survival status predicted by the nomogram fitted better with actual survival status. On the whole, the prognostic risk model of GC on the basis of hypoxia-related genes demonstrated good predictive ability. It can provide more powerful technical support for clinicians to make prognostic determination and therapeutic plans.