Table 3.

Performance of the clinical and radiomics model in the training and validation cohorts

	AUC (95% CI)		Accuracy	Sensitivity	Specificity	PPV	NPV	F-l score
Radiomic model	The training cohort	0.952 (0.907–0.996)	0.8636364	0.8979592	0.8205128	0.8627451	0.8648649	0.88
	The validation cohort	0.898 (0.850–0.946)	0.8409091	0.8775510	0.7948718	0.8431373	0.8378378	0.86
Clinical model	The training cohort	0.670 (0.5904–0.75)	0.6534091	0.7653061	0.5128205	0.6637168	0.6349206	0.71
	The validation cohort	0.678 (0.5535–0.801)	0.6184211	0.7142857	0.5000000	0.6382979	0.5862069	0.67
Combined model	The training cohort	0.906 (0.8632–0.9498)	0.8522727	0.8877551	0.8076923	0.8529412	0.8513514	0.87
	The validation cohort	0.957 (0.9183–0.9963)	0.8947368	0.8333333	0.9705882	0.9722222	0.8250000	0.90

PPV: positive predictive value; NPV: negative predictive value