Table 1.

Performance of the radiomics signature, C-R and C-R-R models

	AUC	Accuracy	Sensitivity	Specificity
Development
LR	0.783 (0.731-0.836)	0.728 (0.673-0.778)	0.774 (0.664-0.849)	0.682 (0.514-0.743)
DT	0.768 (0.716-0.819)	0.731 (0.677-0.781)	0.637 (0.496-0.726)	0.824 (0.695-0.877)
SVM	0.810 (0.761-0.859)	0.748 (0.695-0.797)	0.726 (0.568-0.801)	0.770 (0.648-0.831)
C-R	0.776 (0.724-0.828)	0.711 (0.655-0.762)	0.720 (0.675-0.794)	0.608 (0.455-0.693)
C-R-R	0.849 (0.805-0.893)	0.786 (0.734-0.831)	0.808 (0.685-0.870)	0.764 (0.642-0.845)
Validation
LR	0.778 (0.695-0.862)	0.738 (0.652-0.812)	0.773 (0.560-0.924)	0.700 (0.533-0.834)
DT	0.761 (0.679-0.844)	0.762 (0.678-0.833)	0.652 (0.368-0.731)	0.883 (0.624-0.953)
SVM	0.796 (0.717-0.876)	0.746 (0.661-0.819)	0.742 (0.499-0.909)	0.750 (0.583-0.850)
C-R	0.739 (0.652-0.826)	0.714 (0.627-0.791)	0.712 (0.448-0.803)	0.717 (0.467-0.817)
C-R-R	0.835 (0.761-0.909)	0.802 (0.721-0.867)	0.773 (0.485-0.879)	0.833 (0.617-0.917)

AUC, area under the curve; C-R, clinical-radiological; C-R-R, clinical-radiological-radiomics; DT, decision tree; LR, logistic regression; SVM, support vector machine.