Comprehensive analysis of prognosis and drug sensitivity of programmed cell death pattern-related genes in gastric cancer patients.

Currently, effective prediction models for patients with advanced and postoperative gastric cancer (GC) are lacking. Programmed cell death (PCD) plays a crucial role in the development and metastasis of malignant tumors. This study aimed to investigate the underlying PCD-related molecular mechanisms and develop predictive models for GC. GC profiles were collected from TCGA-STAD, GSE84433, GSE62254, and GSE183904 databases. Differential expression analysis was conducted to identify PCD-related genes (differentially expressed genes (DEGs)), which were then subjected to functional analyses. Cox proportional hazards analyses were used to select PCD-related prognostic DEGs, and a cell death index (CDI) model was proposed. The performance of this model, tumor molecular subtypes, and the tumor microenvironment were assessed. Additionally, drug sensitivity and immune checkpoint expression were examined based on the CDI model. A total of 345 PCD-related DEGs were identified, enriched in processes such as autophagy, apoptosis, necroptosis, ferroptosis, and signaling pathways including p53, NOD-like receptor, IL-17, NF-kappa B, and PI3K-Akt. Subsequently, a CDI model comprising 17 PCD-related prognostic DEGs was constructed, demonstrating superior predictive capability. GC samples were classified into three distinct clustering subtypes, with cluster 1 exhibiting the best overall survival, followed by cluster 3 and cluster 2. Eight immune cell types were significantly associated with the CDI risk score. Furthermore, the CDI risk score exhibited positive correlations with most drugs (except for BMS.754807). Additionally, the expression of immune checkpoint genes PDCD1, CD274, and IDO1 was notably upregulated in the low-risk CDI group. Our developed CDI model, based on 17 PCD-associated prognostic genes, can be employed for risk assessment and prognosis prediction in patients with GC.