Table 2 Predictive performance of clinical-only, radiomics-only, and combined clinical-radiomics models
Models | Training dataset (N = 147) | Validation dataset (N = 63) | ||||||
|---|---|---|---|---|---|---|---|---|
AUC (95%CI) | Accuracy | Sensitivity | Specificity | AUC (95%CI) | Accuracy | Sensitivity | Specificity | |
Conventional CT signature | 0.842 (0.779–0.906) | 0.735 | 62.6% | 91.1% | 0.833 (0.707–0.958) | 65.1% | 53.7% | 86.4% |
Radiomic signature | 0.878 (0.813–0.943) | 0.850 | 90.1% | 76.8% | 0.853 (0.735–0.970) | 81.0% | 82.9% | 77.3% |
Radiomic nomogram | 0.911 (0.858–0.965) | 0.871 | 87.9% | 85.7% | 0.857 (0.723–0.991) | 84.1% | 85.4% | 81.8% |