A novel centrosome-based signature stratifies prognosis and therapeutic sensitivity in gastric cancer

Centrosome dysfunction drives genomic instability and tumor progression, yet its clinical exploitation in gastric cancer (GC) remains limited due to uncharacterized molecular determinants. We therefore profiled centrosomal gene dysregulation to establish mechanistic links with GC progression and uncover clinical prognostic value. Centrosome-related candidate genes were identified by intersecting differentially expressed genes (DEGs) from GC vs. normal transcriptomes with a centrosome-related gene set. Prognostic signature construction combined univariable Cox regression with LASSO feature selection. Independent prognostic factors were integrated into a nomogram. Patients were stratified into low- and high-risk groups based on median risk score for comparative assessment of immune microenvironment and drug sensitivity. Differential expression of prognostic biomarkers was validated in clinical specimens using RT-qPCR, western blotting, and immunohistochemistry. DYNC1I1, GNAI1, and SNCG were identified as prognostic genes for GC. The centrosome-based prognostic signature robustly stratified GC patients. Integration in a nomogram significantly enhanced prognostic precision. The high-risk group exhibited significantly increased infiltration of seven immune cell subsets (including memory B cells) and heightened sensitivity to 19 chemotherapeutic agents (e.g., AZD6738 and BMS-345541). The downregulation of DYNC1I1, GNAI1, and SNCG was observed in bulk tumor tissue compared to normal tissue, while the upregulation in high-risk group tumors, suggesting intra-tumoral heterogeneity. The centrosome-based signature stratifies GC patients into prognostic risk groups with distinct immune evasion phenotypes and targeted drug sensitivity profiles. This identifies high-risk subgroups likely to benefit from specific therapeutic interventions, advancing precision oncology in GC management.