Development of multiple cell death pattern features to predict prognosis and drug sensitivity in gastric adenocarcinoma

Abstract

Background

Regulated cell death (RCD) pathways play a crucial role in stomach adenocarcinoma (STAD), influencing apoptosis and potentially offering new therapeutic opportunities. This study aimed to analyze the immune microenvironment, mutation landscape, and develop a prognostic framework based on RCD-related gene expression for personalized treatment in STAD.

Methods

Datasets from TCGA and GEO (GSE84426) were used. RCD-related genes were analyzed for differential expression with the "limma" R package, and unsupervised clustering was performed using "ConsensusClusterPlus." Immune infiltration was assessed via ssGSEA, and pathway variations were examined with GSVA. A risk-scoring system was developed using Cox regression. Prognostic accuracy was validated through ROC curves, C-index, and multivariate analyses. Drug sensitivity was predicted using the "oncoPredict" package with GDSC pharmacogenomic data.

Results

Twenty-six RCD-associated genes showed significant prognostic relevance. Two clusters were identified, with cluster B demonstrating better overall survival. The prognostic signature, PCD Score, identified CTSV, GCSH, LZTS1, SERPINE1 as unfavorable, and DDIAS, TRAF2 as protective. Validation of the model showed strong predictive power. Functional enrichment analysis revealed key pathways like vascular smooth muscle contraction and dilated cardiomyopathy. Immune profiling revealed elevated immune activity in low-risk patients, consistent with patterns associated with immune checkpoint blockade responsiveness. Computational drug screening suggested a potential susceptibility to temozolomide in high-risk patients, offering a testable hypothesis for future research.

Conclusion

This study presents a robust prognostic model based on RCD-related genes, offering insights into immune response and drug sensitivity for personalized STAD treatment.