A Prediction Model for Optimal Primary Debulking Surgery Based on Preoperative Computed Tomography Scans and Clinical Factors in Patients With Advanced Ovarian Cancer: A Multicenter Retrospective Cohort Study.

Objective This study assessed the predictive value of preoperative computed tomography (CT) scans and clinical factors for optimal debulking surgery (ODS) in patients with advanced ovarian cancer (AOC). Methods Patients with AOC in International Federation of Gynecology and Obstetrics (FIGO) stage III-IV who underwent primary debulking surgery (PDS) between 2016 and 2019 from nine tertiary Chinese hospitals were included. Large-volume ascites, diffuse peritoneal thickening, omental cake, retroperitoneal lymph node enlargement (RLNE) below and above the inferior mesenteric artery (IMA), and suspected pelvic bowel, abdominal bowel, liver surface, liver parenchyma and portal, spleen, diaphragm and pleural lesions were evaluated on CT. Preoperative factors included age, platelet count, and albumin and CA125 levels. Results Overall, 296 patients were included, and 250 (84.5%) underwent ODS. The prediction model included age >60 years (P=0.016; prediction index value, PIV=1), a CA125 level >800 U/ml (P=0.033, PIV=1), abdominal bowel metastasis (P=0.034, PIV=1), spleen metastasis (P 0.750), and the Hosmer-Lemeshow test indicated its stable calibration (P=0.600>0.050). With the aim of maximizing the accuracy of prediction and minimizing the rate of inappropriate explorations, a total PIV ≥5 achieved the highest accuracy of 85.47% and identified patients who underwent suboptimal PDS with a specificity of 100%. Conclusions We developed a prediction model based on two preoperative clinical factors and four radiological criteria to predict unsatisfactory debulking surgery in patients with AOC. The accuracy of this prediction model needs to be validated and adjusted in further multicenter prospective studies.