Abstract / Summary
Ovarian cancer is the most lethal gynecological malignancy, due to late-stage diagnosis, tumor heterogeneity, and lack of biomarkers for early detection. Extracellular vesicles (EVs) in plasma and ascites offer a minimally invasive source of tumor-derived information, yet their lipid composition and relationship to tumor heterogeneity remain poorly defined. We performed lipidomic profiling of EVs from plasma and ascites of high-grade serous ovarian cancer patients and from ovarian cancer and control cell lines to identify signatures of tumor progression and microenvironmental interactions. A lipid panel enriched in ovarian cancer EVs was evaluated in EVs and plasma across two independent cohorts (n = 146). Combined with CA125, this panel improved detection, particularly against benign adnexal masses, achieving an AUROC of 0.99 in the verification cohort in both formats. Ovarian cancer EVs preferentially accumulated triglycerides, including ether-linked species, whereas parental cells were enriched in phospholipids and ceramides. EV lipid signatures reflected tumor-associated metabolic reprogramming and captured tumor heterogeneity across cell models. These findings show that EV lipidomics from accessible biofluids offers a non-invasive window into ovarian cancer metabolism and candidate biomarkers for patient stratification and monitoring.