Table 3 The diagnostic performance of the different models for the MYCN amplification in the training and validation cohorts
AUC (95%CI) | Accuracy (95%CI) | Sensitivity (95%CI) | Specificity (95%CI) | |
|---|---|---|---|---|
Training cohort | ||||
LR | 0.742 (0.627–0.857) | 0.716 (0.599–0.815) | 0.724 (0.379–0.897) | 0.711 (0.400–0.844) |
DT | 0.806 (0.704–0.908) | 0.784 (0.673–0.871) | 0.621 (0.420–0.813) | 0.889 (0.692–0.967) |
SVM | 0.834 (0.721–0.948) | 0.892 (0.798–0.952) | 0.793 (0.000–0.897) | 0.956 (0.222–1.000) |
C-R | 0.672 (0.542–0.803) | 0.716 (0.599–0.815) | 0.345 (0.137–0.517) | 0.956 (0.689–1.000) |
C-R-R | 0.860 (0.757–0.963) | 0.878 (0.782–0.943) | 0.793 (0.474–0.931) | 0.933 (0.333–1.000) |
Validation cohort | ||||
LR | 0.741 (0.539–0.942) | 0.533 (0.343–0.717) | 0.750 (0.500–1.000) | 0.389 (0.165–1.000) |
DT | 0.775 (0.588–0.963) | 0.833 (0.653–0.944) | 0.667 (0.139–0.917) | 0.944 (0.355–1.000) |
SVM | 0.819 (0.632–1.000) | 0.867 (0.693–0.962) | 0.833 (0.000–1.000) | 0.889 (0.110–1.000) |
C-R | 0.681 (0.468–0.893) | 0.633 (0.439–0.801) | 0.500 (0.250–1.000) | 0.722 (0.500–1.000) |
C-R-R | 0.824 (0.657–0.992) | 0.867 (0.693–0.962) | 0.750 (0.000–1.000) | 0.944 (0.333–1.000) |