Integrating Ultrasound-CT-MR for Preoperative Multi-Task Prediction in Ovarian Cancer: Achieving Diagnostic Parity with Multidisciplinary Team Consensus

Ovarian cancer, with its high mortality rate, requires precise preoperative integration of ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI) for guide therapy, yet current clinical workflows rely on specialized multidisciplinary team (MDT) evaluations, a barrier in resource-limited settings. Current artificial intelligence (AI)-driven studies, primarily focused on single-modality analysis and distinguishing malignant from benign ovarian lesions, fail to address the complexity of multimodal decision-making. Here, we developed OVUCM, a multi-modality AI system integrating Ultrasound/CT/MRI via intermediate fusion, using radiomics and machine learning on a retrospective cohort of 1,742 patients from National Regional Medical Center for Obstetrics and Gynecology. The OVUCM system simultaneously predicts five clinical endpoints: benign vs. non-benign (AUC 0.929), borderline vs. malignant (0.889), non-epithelial vs. epithelial (0.897), International Federation of Gynecology and Obstetrics (FIGO) staging I-II vs. III-IV (0.853), and non-high-grade serous ovarian cancer (HGSOC) vs. HGSOC (0.847). This system achieved diagnostic parity with MDT consensus (ΔAUC 0-0.25; Delong test, P  > 0.05 in five tasks), while outperforming independent gynecologists in some tasks. External validation in 102 patients from a general hospital confirmed its generalizability (AUCs: 0.899–0.974 in five tasks). This multi-task and multi-modality system standardizes preoperative workflows by bridging the gap between single-modality tools and the comprehensive, multidisciplinary decision-making required for personalized therapy. Its ability to replicate MDT expertise in resource-limited settings positions it as a transformative tool for global health equity.