Table 3.
Classification performance achieved on the training and test set of the hold-out validation scheme by the radiomic-based model trained using the shifted slices from those selected by our radiologist.
| Training test (30 pz.) |
Test set (8 pz.) |
|
|---|---|---|
| Number of selected features | 4 | 4 |
| AUC (%) | 77.88 (71.00–80.75) | 73.33 |
| Accuracy (%) | 85.00 (80.00–86.67) | 75.00 |
| Sensitivity (%) | 95.00 (85.00–95.00) | 80.00 |
| Specificity (%) | 70.00 (50.00–70.00) | 66.67 |
The median value and the interquartile range (1st and 3rd quantile) of AUC, accuracy, sensitivity, and specificity are reported in percentage values.