Table 6.

Comparison of performance between different risk prediction models.

	Development Cohort				Validation Cohort
	Hazard Ratio (95% CI)	p-Value	C-Index (SE)	p-Value *	Hazard Ratio (95% CI)	p-Value	C-Index (SE)	p-Value *
Radiomics_DCE_MR	3.583 (1.708–7.516)	0.001	0.726 (0.002)	0.030	0.908 (0.385–2.146)	1.000	0.537 (0.005)	0.001
Radiomics_Perfusion	2.891 (1.377–6.065)	0.009	0.687 (0.002)	0.002	2.120 (0.866–5.191)	0.157	0.640 (0.005)	0.008
HRS-only	2.811 (1.340–5.898)	0.011	0.694 (0.003)	0.002	3.274 (1.378–7.782)	0.014	0.699 (0.005)	0.074
Clinical	4.523 (2.155–9.492)	<0.001	0.768 (0.002)	0.342	3.232 (1.330–7.855)	0.018	0.748 (0.004)	0.456
Combined habitat	5.227 (2.490–10.973)	<0.001	0.793 (0.002)	NA	4.128 (1.744–9.769)	0.003	0.760 (0.003)	NA

The p-value in each row refers to the p-value of the hazard ratio model. The p-value * is the model comparison between each model with the combined habitat model using the C-index values.