Abstract
OBJECTIVES: This study aimed to investigate the distribution of molecular signatures in endometriosis- associated ovarian cancer (EAOC) and to develop a prognostic model based on these molecular signatures. METHODS: We retrospectively analyzed EAOC patients treated at Beijing Obstetrics and Gynecology Hospital between December 2015 and July 2023. Progression-free survival (PFS) and overall survival (OS) were compared across molecular subtype groups. Cox regression analysis identified independent recurrence risk factors in EAOC, and a nomogram was constructed using these factors. RESULTS: The cohort included 191 patients. Pathological classification (clear cell carcinoma vs. endometrioid), advanced FIGO stage (III–IV vs. I–II), bilateral ovarian involvement, and MMRd status were identified as independent factors associated with recurrence risk (all P < 0.05). A nomogram incorporating these four variables demonstrated strong predictive performance, with a C-index of 0.844. The areas under the curve (AUCs) for predicting 1-, 3-, and 5-year PFS were 0.838, 0.912, and 0.898, respectively. Calibration curves showed excellent agreement between predicted and observed recurrence probabilities at 1, 3, and 5 years. Bootstrap internal validation confirmed the model’s robust discriminatory power. CONCLUSIONS: Advanced FIGO stage, clear cell carcinoma histology, bilateral ovarian tumors, and MMRd molecular signatures were independent risk factors for EAOC recurrence. The molecular signature-integrated nomogram exhibited strong discrimination and calibration, offering a reliable tool for clinical decision-making in EAOC management. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s13048-025-01937-3.