Skip to main content
. Author manuscript; available in PMC: 2009 May 4.
Published in final edited form as: Leukemia. 2008 Jul 10;22(10):1917–1924. doi: 10.1038/leu.2008.188

Table 3.

Evaluation of the predictive power of candidate models by the C index criterion

Variables C index value (SD)
−IPI 0.62(0.11)
−COO 0.63 (0.09)
− IPI + COO 0.70 (0.12)
− APOBEC3G + RAB33A 0.69 (0.12)
− IPI + COO + APOBEC3G + RAB33A 0.77(0.11)
− LMO2 + MME + FOXP1 + LPP 0.76 (0.09)
− LMO2 + MME + LPP + FOXP1 + APOBEC3G + RAB33A 0.82 (0.08)
− IPI + LMO2 + MME + LPP + FOXP1 + APOBEC3G + RAB33A 0.84 (0.08)

A model 100% efficient to discriminate fatal versus non-fatal disease would have a C index value of 1. A model with no discrimination power would have a C index value of 0.5.