Abstract
Liquid biopsy is a promising approach for early detection of gynecological malignancies. In the PERCEIVE-I study, gynecological Cancer cases (n = 249) and age-matched non-cancer controls (n = 249) are randomly divided into training and test sets at a 1:1 ratio. Data derived from multi-omics assays are obtained including a cell-free DNA methylation panel targeting ≈490 000 CpG sites, a mutation panel comprising 168 genes, and eight tumor protein markers. The results showed that the methylation model outperformed the protein and mutation models, demonstrating higher sensitivity (77.2%) while maintaining similar specificity. The multi-omics model combining methylation and protein markers achieved improved sensitivity (81.9%) with a good specificity (96.9%). The sensitivity varied across different stages, ranging from 66.7% to 100%. The model accurately identified the tissue of origin in 72.1% of cases. The superior performance of the methylation model highlights the potential of integrating multi-omics for non-invasive early detection of gynecological malignancies.