Establishment of a predictive survival prognosis model for epithelial ovarian cancer: a study based on the SEER database

Background: This study aimed to develop a predictive model for survival prognosis in epithelial ovarian cancer (EOC) patients using the Surveillance, Epidemiology, and End Results (SEER) database and identify key factors influencing survival outcomes, providing a reference for personalized treatment and clinical decision-making. Methods: Data from EOC patients diagnosed between 2000 and 2021 were analyzed. Variables included age, pathological grade, International Federation of Gynecology and Obstetrics (FIGO) stage, operation, surgical intervention, marital status, income level, tumor size, time from diagnosis to treatment, radiotherapy, chemotherapy and lymph node positivity. Prognostic factors were identified through univariate and multivariate COX proportional hazards regression. A nomogram predictive model was constructed using significant variables. Internal validation assessed the model’s performance using calibration plots and Receiver Operating Characteristic (ROC) curves. Results: COX regression identified pathological grade, FIGO stage, operation, radiotherapy status and time from diagnosis to treatment as significant prognostic factors. The nomogram demonstrated strong calibration between predicted and observed outcomes. ROC analysis yielded Area Under Curve (AUC) values of 0.853 (1-year), 0.840 (2-year) and 0.847 (3-year), indicating good predictive accuracy. The model visually quantified each variable’s contribution to survival. Conclusions: Pathological grade, FIGO staging, radiotherapy status and treatment delay are critical prognostic indicators in EOC. The developed nomogram provides valuable guidance for individualized treatment planning and supports evidence-based clinical decisions.