Table 1 The optimal single-factor predictors with AUC, sensitivity and specificity for three morphological indices using four machine learning methods
Index | ΔLPI | ΔCTI | ΔMPVI | |||
|---|---|---|---|---|---|---|
Method | Factor | AUC (Spe, Sen) | Factor | AUC (Spe, Sen) | Factor | AUC (Spe, Sen) |
RF | Critical PWSn | 0.856 (0.928, 0.749) | MinCT | 0.749 (0.858, 0.555) | MinCT | 0.785 (0.863, 0.656) |
DA | Plaque area | 0.875 (0.872, 0.677) | MinCT | 0.818 (0.868, 0.613) | MinCT | 0.752 (0.830, 0.646) |
SVM | Wall thickness | 0.883 (0.947, 0.653) | MinCT | 0.697 (0.853, 0.334) | MinCT | 0.727 (0.866, 0.404) |
EL | Plaque area | 0.776 (0.927, 0.767) | MinCT | 0.719 (0.852, 0.530) | MinCT | 0.766 (0.864, 0.654) |