Table 3.

Evaluation indicators of multiple models of training and test sets.

	Model	AUC	95% CI	accuracy	sensitivity	specificity	NPV	PPV
TS	Radomics	0.807	0.753–0.854	78.80	69.32	85.19	55.68	91.36
	Clinical-semantic	0.615	0.551–0.675	68.40	42.05	81.68	90.74	27.27
	Combined	0.840	0.788–0.883	81.60	77.27	81.48	90.74	64.77
VS	Radomics	0.786	0.665–0.879	74.60	54.55	92.68	85.37	54.55
	Clinical-semantic	0.683	0.554–0.795	71.43	50.00	80.49	90.24	36.36
	Combined	0.809	0.690–0897	74.60	81.82	65.85	85.37	54.55

TS：training set, VS: test set