31 March 2026 : Clinical Research
Development and Validation of a Predictive Model Combining [18F] FDG PET/CT and the Suidan Score for Primary Optimal Cytoreduction in Advanced High-Grade Serous Ovarian Cancer: A Retrospective Cohort Study
Xiaohang Lu ACE 1, Xiaomei Wang CD 2, Shengnan Wang BC 1, Jiazhen Huang DE 1, Wei Wei EF 1, Fuli Kang EF 1, Ning Wang AFG 1*DOI: 10.12659/MSM.951183
Med Sci Monit 2026; 32:e951183
Table 2 Differences between the 3 models.
| Variable | AUC (95% CI) | Brier score | Sensitivity (95% CI) | Specificity (95% CI) | PPV (95% CI) | NPV (95% CI) | Accuracy (95% CI) |
|---|---|---|---|---|---|---|---|
| Suidan score | 77.3% (61.5–93.1%) | 0.233 | 81.8% (48.2–97.7%) | 38.7% (22.2–57.4%) | 32.1% (15.9–52.4%) | 85.7% (57.2–98.2%) | 50.0% (33.8–66.2%) |
| PET/CT model | 80.8% (66.5–95.1%) | 0.143 | 72.7% (39.0–94.0%) | 80.6% (62.5–92.5%) | 57.1% (28.9–82.3%) | 89.3% (71.8–97.7%) | 78.6% (62.2–90.0%) |
| Combined model | 83.6% (70.6%–96.6%) | 0.146 | 81.8% (48.2–97.7%) | 83.9% (66.3–94.5%) | 64.3% (35.1–87.2%) | 92.9% (76.5–99.1%) | 83.3% (67.2–93.6%) |
| * Performance metrics are reported for the temporally held-out validation set (n=40). AUC, area under the receiver operating characteristic curve; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value. The 95% CIs for AUC were calculated via 2000 bootstrap replicates. CIs for sensitivity, specificity, PPV, NPV, and accuracy are exact Clopper-Pearson confidence intervals. The optimal classification threshold for each machine learning model was determined by maximizing the Youden index. The Brier score assesses calibration (range 0–1, with lower scores indicating better agreement between predicted probabilities and observed outcomes). | |||||||