Integrative Bioinformatic Analysis of Cellular Senescence Genes in Ovarian Cancer: Molecular Subtyping, Prognostic Risk Stratification, and Chemoresistance Prediction

Background: Ovarian cancer (OC) is a heterogeneous malignancy associated with a poor prognosis, necessitating robust biomarkers for risk stratification and therapy optimization. Cellular senescence-related genes (CSGs) are emerging as pivotal regulators of tumorigenesis and immune modulation, yet their prognostic and therapeutic implications in OC remain underexplored. Methods: We integrated RNA-sequencing data from TCGA-OV (n = 376), GTEx (n = 88), and GSE26712 (n = 185) to identify differentially expressed CSGs (DE-CSGs). Consensus clustering, Cox regression, LASSO-penalized modeling, and immune infiltration analyses were employed to define molecular subtypes, construct a prognostic risk score, and characterize tumor microenvironment (TME) dynamics. Drug sensitivity was evaluated using the Genomics of Drug Sensitivity in Cancer (GDSC)-derived chemotherapeutic response profiles. Results: Among 265 DE-CSGs, 31 were prognostic in OC, with frequent copy number variations (CNVs) in genes such as STAT1, FOXO1, and CCND1. Consensus clustering revealed two subtypes (C1/C2): C2 exhibited immune-rich TME, elevated checkpoint expression (PD-L1, CTLA4), and poorer survival. A 19-gene risk model stratified patients into high-/low-risk groups, validated in GSE26712 (AUC: 0.586-0.713). High-risk patients showed lower tumor mutation burden (TMB), immune dysfunction, and resistance to Docetaxel/Olaparib. Six hub genes (HMGB3, MITF, CKAP2, ME1, CTSD, STAT1) were independently predictive of survival. Conclusions: This study establishes CSGs as critical determinants of OC prognosis and immune evasion. The molecular subtypes and risk model provide actionable insights for personalized therapy, while identified therapeutic vulnerabilities highlight opportunities to overcome chemoresistance through senescence-targeted strategies.