Development of a Prognostic Model for Gastric Cancer Based on Apoptosis- and Hypoxia-Related Genes: Predictive Insights into Survival and Immune Landscape

Abstract

Background: Gastric cancer (GC) stands as the fifth most prevalent malignancy worldwide and the fourth primary contributor to cancer-associated fatalities worldwide. The search for pivotal prognostic biomarkers and elucidation of underlying mechanisms driving GC progression necessitate innovative approaches. Apoptosis and hypoxia were intricately associated and interdependent in tumorigenesis. This work investigated the potential value of apoptosis- and hypoxia-related genes (AHRGs) in GC prognosis, with implications for immune therapy in GC patients. Methods: Differential expression analysis was performed on GC transcriptomic data from TCGA. Apoptosis-related genes (ARGs) and hypoxia-related genes (HRGs) were obtained from the MSigDB database, followed by intersection analysis with differentially expressed genes (DEGs) in GC. A prognostic model was constructed using univariate Cox regression, LASSO analysis, and multivariate Cox regression analyses. The model was validated using a GEO dataset, and DEGs between high- and low-risk groups were subjected to enrichment analysis. A nomogram was created by incorporating clinical information. Non-negative matrix factorization based on core prognostic genes from the multifactorial model was employed to cluster tumor samples. The subsequent analyses encompassed immune landscape, immunophenoscore, TIDE score, as well as chemosensitivity for distinct subtypes. Results: A prognostic model based on AHRGs was established, and its robust predictive capability was validated in external cohorts. Riskscore was determined as an independent prognostic factor, augmenting prognostic nomogram in conjunction with other clinical features.