Table 4 Performance of three radiomics model in the training and external validation set
Model | AUC (95% CI) | Accuracy | Sensitivity | Specificity | |
|---|---|---|---|---|---|
Training set | EN model | 0.770 (0.716–0.815) | 0.695 (0.646–0.741) | 0.602 (0.529–0.668) | 0.821(0.760–0.879) |
PEZ model | 0.744 (0.691–0.794) | 0.703 (0.657–0.749) | 0.716 (0.654–0.776) | 0.686 (0.612–0.757) | |
Combined model | 0.853 (0.811–0.890) | 0.779 (0.738–0.823) | 0.725 (0.665–0.785) | 0.853 (0.795–0.909) | |
External validation set | EN model | 0.732 (0.615–0.846) | 0.691 (0.588–0.794) | 0.730 (0.579–0.864) | 0.645 (0.472–0.812) |
PEZ model | 0.715 (0.581–0.833) | 0.676 (0.559–0.794) | 0.649 (0.500–0.806) | 0.710 (0.548–0.862) | |
Combined model | 0.809 (0.696–0.909) | 0.779 (0.676–0.882) | 0.757 (0.606–0.892) | 0.806 (0.655–0.933) |