Table 3.
Predictive models for platinum-resistant recurrent ovarian cancer
| Model I | Model II | Model III | |
|---|---|---|---|
| Sensitivity | 81.82% | 72.70% | 95.50% |
| Specificity | 75.51% | 85.70% | 65.30% |
| Positive predictive value | 60.00% | 69.57% | 55.26% |
| Negative predictive value | 90.24% | 87.50% | 96.97% |
| AIC | 70.60 | 68.44 | 56.49 |
| Youden index | 0.57 | 0.58 | 0.61 |
| AUC | 0.78 | 0.79 | 0.89 |
| 95% CI for AUC | 0.66–0.89 | 0.67–0.92 | 0.81–0.97 |
A. Stromal cell ratio based on automated image analysis
B. Normalization of CA125 level after 3 cycles of chemotherapy
Model I: A
Model II: B
Model III: A + B
AIC Akaike information criterion
AUC Area under curve