Multi-Omics Profiling in a Symptomatic Cohort Identifies Coordinated Biomarker Signatures in Ovarian Cancer Serum

Background/Objectives: Ovarian cancer (OC) is a leading cause of cancer-related mortality in women, largely driven by late-stage diagnosis. Five-year survival is just 30% for advanced-stage (III-IV) disease but exceeds 90% for early-stage disease, underscoring the critical need for effective early detection tools. Current standard-of-care biomarkers show limited sensitivity for early-stage OC and lack specificity in symptomatic populations. Most biomarker studies in OC serum evaluate single molecular classes or compare OC to healthy controls, limiting understanding of coordinated biological alterations in circulating proteins, lipids, and metabolites in clinically relevant populations. Methods: We performed integrated multi-omics profiling of serum from a retrospective, case-control cohort of women presenting with vague abdominal symptoms (VAS), including early- and late-stage OC, borderline tumors, benign gynecologic conditions including adnexal masses, GI disorders, and healthy controls. Protein biomarkers were quantified by ELISA, lipidomic profiling was performed by untargeted LC-MS, and ganglioside and metabolomic profiling were performed by semi-targeted LC-MS with metabolite annotation performed against a curated reference library. Results: Consistent with known limitations for early-stage OC detection, CA125 and HE4 levels overlapped substantially with benign gynecologic conditions. Additional proteins also showed limited separation in their expression between early-stage OC and symptomatic controls. In contrast, OC showed unique lipid and metabolite profiles: phospholipids and glycerolipids were decreased, and sphingolipid composition was altered. Borderline and benign conditions exhibited lipid profiles that fall between healthy and OC groups, suggesting a continuum of metabolic changes rather than distinct states between OC and non-OC controls. Sphingolipid alterations included changes in ceramides and sphingomyelins, along with broader dysregulation of ganglioside profiles, including an elevated GD2;O2-to-GD1;O2 ratio. Metabolic profiling showed decreased amino acids and enriched cysteine metabolism in OC, consistent with altered redox balance, along with changes in fatty acids and acyl-carnitines, suggesting altered lipid metabolism and inflammatory mechanisms. Lower levels of glycolytic and TCA cycle intermediates in OC suggested altered mitochondrial metabolism and energetic reprogramming. Pairwise comparisons revealed a gradient of significance between groups, with differences between OC and healthy controls across lipid classes (LPC, PC, PE, TG, SM), gangliosides (GD1, GD2, GD2/GD1 ratio), and metabolites (amino acids, Cys/CySS, TCA cycle); borderlines occupied an intermediate space. Integration of these datasets revealed coordinated cross-omics relationships, identifying links between metabolite, lipid, and protein features. Together, these connections highlight structured, system-level alterations related to lipid remodeling, redox balance, immune signaling, and energy metabolism that no single modality would have revealed in isolation. Conclusions: This study presents an integrated analysis of the lipidome, gangliosome, metabolome, and protein biomarkers within a single clinically relevant symptomatic cohort enriched with multiple stages and subtypes of OC. This multi-omics framework demonstrates that molecular alterations in OC are biologically interconnected across molecular classes. While these findings are discovery-based and require independent validation prior to clinical application, they support the development of clinically deployable multi-omics biomarker strategies for early detection and potential pathways for therapeutic intervention.