Predicting residual disease and extent of surgery after neoadjuvant chemotherapy for ovarian cancer: Does the MSKCC model from the primary setting apply?

BACKGROUND: Complete gross resection (CGR) a key determinant of survival in advanced ovarian cancer (AOC). Preoperative imaging is used to predict residual disease (RD) and surgical complexity in the primary setting, but its performance after neoadjuvant chemotherapy (NACT) remains unclear. We evaluated the ability of computed tomography (CT)-based models in predicting RD and surgical complexity at interval debulking surgery (IDS). METHODS: This multicentre retrospective cohort study included 246 patients with FIGO stage IIIC-IV AOC undergoing NACT-IDS between 2016 and 2021. The Memorial Sloan Kettering (MSK) CT-based predictive model for RD was applied and recalibrated. Pre- and post-NACT CT scans were independently reviewed by six radiologists for 18 predefined disease sites. Logistic regression identified radiologic and clinical predictors of RD and advanced surgical procedures. RESULTS: CGR was achieved in 61% of patients; 35% had ≤1 cm RD, and 4% had >1 cm. The MSK model demonstrated limited discrimination for RD in the IDS setting (c-statistic 0.670). A revised model incorporating optimized age, CA-125 thresholds together with selected radiologic features improved discrimination (c-statistic 0.711). A model predicting the need for advanced surgical procedures achieved moderate performance (c-statistic 0.736). Subcapsular and perihepatic lesions were associated with diaphragm procedures; however absence on CT imaging did not preclude diaphragmatic stripping. CONCLUSIONS: MSK algorithm demonstrated limited discrimination in predicting CGR during IDS. A revised model incorporating optimized age and CA-125 thresholds improved performance. While specific CT findings may help anticipate the need for targeted surgical techniques and multidisciplinary involvement, their absence does not preclude the need for such procedures.