Table 2 Predictive performance of the CT diagnostic criteria and generated classifiers
| Models | Dataset | Accuracy (%) | Sensitivity (%) | Specificity (%) | AUC ± 95% CI |
|---|---|---|---|---|---|
| Clinical model | Training | 65.38 | 83.87 | 47.13 | 0.703 ± 0.03 |
| Test | 62.82 | 84.62 | 41.03 | 0.772 ± 0.05 | |
| Patient-demographic model | Training | 67.31 | 62.58 | 71.97 | 0.706 ± 0.03 |
| Test | 73.08 | 69.23 | 76.92 | 0.773 ± 0.05 | |
| Radiomic-derived model | Training | 81.09 | 83.87 | 78.34 | 0.882 ± 0.02 |
| Test | 79.49 | 74.36 | 84.62 | 0.825 ± 0.05 |