Table 2.

The performance of radiomic signatures.

Classifier stage	Signature	AUC	95%CI	ACC	SEN	SPE
Classifier #1	GTR	0.708	0.603, 0.812	0.624	0.784	0.531
	PTR	0.710	0.609, 0.811	0.653	0.757	0.594
	GPTR(1)	0.726	0.625, 0.827	0.663	0.838	0.562
	GPTR(2)	0.680	0.574, 0.786	0.634	0.811	0.531
Classifier #2	GTR	0.806	0.667, 0.944	0.730	0.333	0.800
	PTR	0.752	0.583, 0.921	0.757	0.333	0.929
	GPTR(1)	0.770	0.616, 0.923	0.730	0.667	0.750
	GPTR(2)	0.742	0.578, 0.906	0.649	0.778	0.607

AUC, area under the curve; CI, confidece interval; ACC, accuracy; SEN, sensitivity; SPE, specificity; GTR, gross tumor region; PTR, peritumoral region; GPTR, Gross and peri tumoral volume.

GPTR₍₁₎, the GPTR signature developed by logistic regression using GTR and PTR signatures.

GPTR₍₂₎, The GPTR radiomic signature developed by radiomic features extracted from GTR and PTR combination region.