Table 2.

Performance evaluation of the models in the training cohort and validation cohort

	AUC (95%CI)	Sensitivity	Specificity	PPV	NPV	Accuracy	Youden Index
Training cohort
INTRA	0.77(0.71–0.83)	0.82	0.67	0.55	0.88	0.72	0.49
IntraPeri2mm	0.76(0.70–0.82)	0.84	0.60	0.51	0.89	0.68	0.44
IntraPeri4mm	0.67(0.60–0.74)	0.89	0.38	0.41	0.88	0.55	0.27
IntraPeri6mm	0.68(0.62–0.75)	0.83	0.48	0.44	0.85	0.59	0.31
IntraPeri8mm	0.70(0.64–0.77)	0.67	0.68	0.51	0.81	0.68	0.36
DeepL	0.75(0.69–0.81)	0.60	0.83	0.63	0.81	0.75	0.42
RAITS	0.87(0.82–0.92)	0.71	0.92	0.82	0.87	0.85	0.63
validation cohort
INTRA	0.60(0.48–0.72)	0.42	0.83	0.64	0.67	0.66	0.25
IntraPeri2mm	0.63(0.52–0.75)	0.55	0.68	0.55	0.68	0.63	0.23
IntraPeri4mm	0.65(0.54–0.76)	0.92	0.43	0.54	0.88	0.64	0.36
IntraPeri6mm	0.61(0.49–0.72)	0.55	0.68	0.55	0.68	0.63	0.23
IntraPeri8mm	0.66(0.54–0.77)	0.89	0.40	0.52	0.84	0.60	0.29
DeepL	0.65(0.54–0.76)	0.76	0.51	0.53	0.75	0.62	0.27
RAITS	0.78(0.69–0.88)	0.66	0.83	0.74	0.77	0.76	0.49

AUC, area under the receiver operating characteristics curve; CI, confidence interval; PPV, positive predictive values; NPV, negative predictive values; INTRA, intra-tumoral signature; IntraPeri2mm, peritumoral radiomic signature extracted from the 2 mm peritumoral area; IntraPeri4mm, peritumoral radiomic signature extracted from the 4 mm peritumoral area; IntraPeri6mm, peritumoral radiomic signature extracted from the 6 mm peritumoral area; IntraPeri8mm, peritumoral radiomic signature extracted from the 8 mm peritumoral area; DeepL, radiomics signature constructed by deep learning features; RAITS, Radiomics Integrated TLSs System