Integrative machine-learning analysis of omics data for platinum sensibility predictive profiling in unresectable high-grade serous tubo-ovarian cancer

High-grade serous ovarian carcinoma (HGSC) is characterized by marked genomic instability and heterogeneous clinical outcomes. Although BRCA status and homologous recombination deficiency (HRD) currently guide therapeutic strategies, they do not fully explain the variability in treatment response and prognosis. This study investigated HGSC through an integrated genomic and transcriptomic approach. A clinically annotated patient cohort was analyzed using targeted next-generation sequencing to assess somatic variants and copy number variations, alongside RNA sequencing for unsupervised transcriptomic clustering. Recurrent genomic alterations were identified, including widespread copy number changes. Transcriptomic analysis revealed distinct molecular signatures associated with clinical features and survival outcomes. Notably, specific expression patterns showed potential prognostic relevance beyond BRCA/HRD status. These findings support the need for a multilevel model integrating genomic, transcriptomic, and clinical data to refine patient stratification in the evolving therapeutic landscape of HGSC.