Development and external validation of a predictive model for epithelial ovarian tumors based on MRI features and serum tumor markers

Ovarian cancer remains a major threat to women’s health, and early-stage disease is often asymptomatic, making timely diagnosis difficult. This study aimed to develop and externally validate a predictive model integrating magnetic resonance imaging (MRI) features and serum tumor markers to distinguish benign from malignant epithelial ovarian tumors. A total of 222 patients from Huzhou Maternity & Child Health Care Hospital were included as the training cohort, and 102 patients from Sir Run Run Shaw Hospital served as the external validation cohort. Independent predictors were identified using multivariate logistic regression, and a predictive model was constructed based on regression coefficients and externally validated. In the training cohort, 168 patients were benign and 54 were malignant. Tumor morphology, diffusion-weighted imaging (DWI) signal intensity, ascites, carbohydrate antigen 125 (CA125), and human epididymis protein 4 (HE4) were independent predictors (AUCs: 0.748–0.876). The combined model achieved an AUC of 0.965, with 94.4% sensitivity and 84.5% specificity, and showed good calibration ( P = 0.870). In the validation cohort (80 benign, 22 malignant), the model yielded an AUC of 0.972, with 90.9% sensitivity and 95.0% specificity. The proposed model integrating MRI features and serum tumor markers demonstrated excellent diagnostic performance and robust external validity. This practical and easily applicable tool may assist in preoperative risk stratification and clinical decision-making.