A novel model for predicting immunotherapy response and prognosis in NSCLC patients

Table 2 The C-index of OS/PFS for RF model, Nomogram, SIS, and LIPI

Survival prediction	C-index	95 CI%	p
Training cohort
OS
RF model	0.803	0.763–0.843
Nomogram model	0.507	0.448–0.566
SIS	0.541	0.488–0.593
LIPI	0.534	0.488–0.580
RF model vs Nomogram model			<0.001
RF model vs SIS			<0.001
RF model vs LIPI			<0.001
PFS
RF model	0.663	0.617–0.710
Nomogram model	0.502	0.455–0.549
SIS	0.508	0.461–0.555
LIPI	0.528	0.492–0.565
RF model vs Nomogram model			<0.001
RF model vs SIS			<0.001
RF model vs LIPI			<0.001
Validation cohort
OS
RF model	0.712	0.652–0.772
Nomogram model	0.501	0.415–0.586
SIS	0.557	0.497–0.618
LIPI	0.501	0.442–0.560
RF model vs Nomogram model			<0.001
RF model vs SIS			<0.001
RF model vs LIPI			<0.001
PFS
RF model	0.711	0.658–0.765
Nomogram model	0.511	0.443–0.580
SIS	0.515	0.453–0.577
LIPI	0.512	0.462–0.562
RF model vs Nomogram model			<0.001
RF model vs SIS			<0.001
RF model vs LIPI			<0.001

The concordance index is denoted as C-index. P values are computed by applying the function rcorrp.cens in the Hmisc package, based on normal approximation

ISSN: 1475-2867