Table 3.
Results of radiomics nomogram, radiomics signature, and the clinical model predictive ability for distinguishing between aggressive ccRCC and non-aggressive ccRCC.
| Variables | AUC(95% CI) | Sensitivity* | Specificity* | Accuracy* | |
|---|---|---|---|---|---|
| Clinical model | Training cohort | 0.86(0.78–0.94) | 74.3(26/35) | 86.7(72/83) | 83.1(98/118) |
| Validation cohort | 0.90(0.80–0.99) | 60.0(9/15) | 91.4(32/35) | 82.0(41/50) | |
| Radiomics signature | Training cohort | 0.86(0.79–0.92) | 88.6(31/35) | 77.1(64/83) | 80.5(95/118) |
| Validation cohort | 0.85(0.73–0.97) | 73.3(11/15) | 82.9(29/35) | 80.0(40/50) | |
| Radiomics nomogram | Training cohort | 0.91(0.86–0.97) | 88.6(31/35) | 81.9(68/83) | 83.9(99/118) |
| Validation cohort | 0.92(0.85–0.99) | 100.0(15/15) | 77.1(27/35) | 84.0(42/50) |
CI, confidence interval.
*Numbers in parentheses were used to calculate percentages.