Research Article

LASSO Model Better Predicted the Prognosis of DLBCL than Random Forest Model: A Retrospective Multicenter Analysis of HHLWG

Table 2

Multivariable analysis of OS based on LASSO and random forest.

VariablesHR95% CI

LASSO
 Age1.0321.022-1.042<0.001
 WBC1.0281.017-1.040<0.001
 HB0.9880.983-0.993<0.001
 CNS involvement2.2411.636-3.068<0.001
 Gender0.6590.524-0.829<0.001
 Ann Arbor stage1.6441.286-2.100<0.001
Random forest
 Age1.0311.021-1.041<0.001
 WBC1.0311.020-1.041<0.001
 HB0.9880.983-0.994<0.001
 CNS involvement1.9921.462-2.715<0.001
 ALB0.9840.969-0.9990.038
 ECOG1.2951.011-1.6590.040