Table 3.

Diagnostic performances of different models on internal and external test sets

	Models	Accuracy	Sensitivity	Specificity	PPV	NPV	AUC (95% CI)
Internal test set	Peritumoral	0.714	0.750	0.673	0.724	0.702	0.786 (0.695, 0.872)
	Intratumoral	0.657	0.679	0.633	0.679	0.633	0.652 (0.538, 0.749)
	Original	0.686	0.643	0.735	0.735	0.643	0.694 (0.591, 0.792)
	Intratumoral + peritumoral	0.771	0.804	0.735	0.776	0.766	0.788 (0.699, 0.871)
	Clinical	0.733	0.732	0.735	0.759	0.706	0.794 (0.703, 0.874)
	Ensemble	0.800	0.804	0.796	0.818	0.780	0.832 (0.753, 0.908)
External test set	Peritumoral	0.732	0.786	0.676	0.714	0.754	0.789 (0.712, 0.861)
	Intratumoral	0.688	0.657	0.721	0.708	0.671	0.750 (0.667, 0.831)
	Original	0.681	0.543	0.824	0.760	0.636	0.784 (0.701, 0.859)
	Intratumoral + peritumoral	0.732	0.771	0.691	0.720	0.746	0.794 (0.718, 0.866)
	Clinical	0.725	0.657	0.794	0.767	0.692	0.806 (0.737, 0.871)
	Ensemble	0.761	0.757	0.765	0.768	0.754	0.843 (0.783, 0.900)

PPV positive predictive value, NPV negative predictive value, AUC area under the curve, CI confidence interval