Abstract
We evaluated the ability of our two laparotomy-based models to predict optimal primary debulking surgery (PDS) and long-term outcomes of stage IIIC epithelial ovarian cancer (EOC). Data of 400 IIIC EOC patients who underwent laparotomy were retrospectively analyzed. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and overall accuracy were calculated for 10 parameters. The parameters with a specificity ≥75%, PPV ≥50%, and NPV ≥50% were included in the final predictive index value (PIV) model. Peritoneal cancer index (PCI) was calculated summarizing lesion size scores (LSSs) of 13 regions. Receiver operating characteristic (ROC) curve was used to assessed the predictive value of PIV and PCI for optimal PDS. Univariate and multivariate analyses were performed to assess the prognostic value of PIV and PCI. After PDS, 223 (55.8%) patients with RD ≤1 cm had longer progression-free survival (PFS) and overall survival (OS) than patients with RD >1 cm (PFS: 22.4 vs. 15.4 months, respectively; P < 0.001 and OS: 48.6 vs. 35.6 months; P < 0.001). PCI better predicted optimal PDS than PIV (The area under the curve of ROC: PCI 0.79 vs. PIV 0.75). The predictive value of PIV and PCI models was verified using another cohort of 77 patients. And PIV and PCI models were demonstrated to be more powerful than the published laparoscopy-based predictive index (LPS-PI) model. Patients with a PIV ≥14 were more likely to undergo suboptimal PDS with a specificity of 100%. The median PFS and OS of patients with PIV < 3 were significantly longer than patients with PIV > 3 (PFS: 19.5 vs. 16.3 months, P = 0.007; OS: 46.1 vs. 37.0 months, P = 0.009). The median PFS and OS of patients with the PCI < 17.5 were significantly longer than patients with the PCI > 17.5 (PFS: 22.9 vs. 14.5 months, P < 0.001; OS: 54.3 vs. 31.5 months, P < 0.001). PCI could better predict optimal PDS compared with PIV. PCI was an independent prognostic factor for long-term outcome of IIIC EOC patients.