Skip to main content

Table 2 Comparison of predictive performance of different classification in training cohort

From: Dynamic nomograms combining N classification with ratio-based nodal classifications to predict long-term survival for patients with lung adenocarcinoma after surgery: a SEER population-based study

Model IDI (95%CI) P NRI (95%CI) P
Cancer-specific survival
 Model 1 (NPLN) −0.006 (−0.010 to −0.002) < 0.001 −0.000 (−0.054 to 0.029) 0.884
 Model 2 (LODDS) −0.006 (−0.009 to −0.003) < 0.001 −0.119 (−0.160 to 0.016) 0.093
 Model 3 (LNR) −0.002 (−0.004 to −0.000) < 0.001 −0.094 (−0.125 to −0.068) < 0.001
 Model 4 (LODDS+LNR) Reference   Reference  
Overall survival
 Model 1 (NPLN) −0.008 (−0.013 to −0.005) < 0.001 −0.111 (−0.132 to −0.050) < 0.001
 Model 2 (LODDS) −0.006 (−0.010 to −0.004) < 0.001 −0.067 (−0.131 to 0.039) 0.266
 Model 3 (LNR) −0.001 (−0.003 to −0.000) < 0.001 −0.088 (−0.136 to −0.066) < 0.001
 Model 4 (LODDS+LNR) Reference   Reference  
  1. IDI integrated discrimination improvement; NRI, net reclassification improvement; NPLN number of positive lymph nodes; LODDS log odds of positive lymph nodes; LNR lymph node ratio