. 2021 Oct;10(10):4375–4386. doi: 10.21037/tcr-21-702

Table 2. Comparison of performances of Rad-Score and nomogram.

Groups	Training Cohort				Validation Cohort
Groups	AUC_αbin	AUC_emp	AP_αbin	AP_emp	AUC_αbin	AUC_emp	AP_αbin	AP_emp
Rad-Score	0.854 (0.849–0.859)	0.852 (0.848–0.856)	0.822 (0.820–0.824)	0.808 (0.801–0.815)	0.792 (0.790–0.794)	0.809 (0.803–0.812)	0.730 (0.718–0.742)	0.689 (0.681–0.697)
Nomogram	0.911 (0.903–0.919)	0.909 (0.906–0.912)	0.885 (0.876–0.894)	0.869 (0.865–0.873)	0.873 (0.866–0.880)	0.866 (0.861–0.871)	0.827 (0.820–0.834)	0.807 (0.796–0.818)
P value	0.001^ζ	0.001^ε	0.001^η	0.001^θ	0.001^ψ	0.001^σ	0.001^ρ	0.001^μ

Note: All the data in parentheses are 95% confidence interval (CI). Subscript emp and αbin were empirical-based and α-binormal-based, respectively, AUC or AP. ^{ζ, ψ} The comparison of AUC_αbin, ^{ε, σ} The comparison of AUC_emp, ^{η, ρ} The comparison of AP_αbin, and ^{θ, μ} The comparison of AP_emp between Rad-Score and nomogram in both training and validation cohort.