Table 5.

Comparison of new risk models for overall survival using the Heng and MSKCC risk models with 2000 bootstraps.

Model	Harrell's C index	mean(difference), 95% CI (2.5%, 97.5% of difference)
Model A	0.708	Model B vs A: 0.02 (-0.011, 0.058)
Model B	0.727
Heng risk model	0.661	Heng vs Model A: -0.035 (-0.088, 0.008) Heng vs Model B: -0.055 (-0.112, -0.004)
mMSKCC risk model	0.612	mMSKCC vs Model A: -0.097 (-0.153, -0.043) mMSKCC vs Model B: -0.117 (-0.174, -0.066)
Model C = Heng risk model + AST	0.676	Heng vs (Heng + SGOT): -0.011 (-0.031, 0.004)
Model D = Heng risk model + Alkaline phosphatase + DRR	0.697	Heng vs (Heng + De Ritis ratio): -0.035 (-0.083, 0)
		(Heng + SGOT) vs (Heng + De Ritis ratio): -0.024 (-0.07, 0.011)
Model E = mMSKCC risk model + Neutrophil + AST	0.658	mMSKCC vs Model E: -0.049 (-0.098, -0.013)
Model F = mMSKCC risk model + NLR + Alkaline phosphatase +DRR	0.691	mMSKCC vs Model F: -0.084 (-0.149, -0.034) Model E vs Model F: -0.034 (-0.092, 0.014)

MSKCC, Memorial Sloan Kettering Cancer Center; AST, aspartate transaminase; NLR, neutrophil-to-lymphocyte ratio; CI, confidence interval.

Model A = Nephrectomy, Liver mets, Hb, Neutrophil, Alkaline phosphatase.

Model B = Nephrectomy, Hb, NLR, Alkaline phosphatase, DRR.