Study on the Relationship Between Adverse Events and Efficacy of Immunotherapy for Gastric Cancer and Target Prediction

Author(s): Zidong Zhao, Xu Zhang, Yu Tan, Dandan Zhao, Satoshi Endo, Yanwen Liu

Objective: This study aimed to dissect molecular mechanisms of immune-related adverse events (irAEs) using tumor genomics data. We screened key regulatory pathways via immune enrichment analysis and identified core gene biomarkers through integrated bioinformatics and machine learning approaches, providing a basis for early irAE warning, efficacy evaluation, and personalized treatment.

Methods: RNA sequencing data of gastric adenocarcinoma (TCGA-STAD) were retrieved from TCGA. Six irAE-related immune pathways (Neutrophils, T cell receptor [TCR], Eosinophils, etc.) were selected. Pathway activity was validated by GSVA and GSEA. Key genes were screened and their diagnostic value evaluated using UpSet analysis, Boruta algorithm, and XGBoost model.

Results: The six pathways were significantly enriched in high irAE score samples, with Neutrophils (NES=1.97), TCR (NES=1.96), and Eosinophils (NES=1.84) showing highest enrichment. Five key genes (C11orf21, RGS19, etc.) were identified. RGS19 was highly expressed in gastric cancer tissues with optimal diagnostic performance (AUC=0.799).

Conclusion: Six irAE-related pathways (Neutrophils, TCR, etc.) and five key genes (including C11orf21, RGS19) were identified from TCGA. RGS19, highly expressed in gastric cancer with prominent diagnostic efficacy, provides clinical and molecular evidence for irAE prediction and early diagnosis.