Machine learning for differentiation of lipid-poor adrenal adenoma and subclinical pheochromocytoma based on multiphase CT imaging radiomics

Table 3 Comparison of diagnostic efficacy and ROC curves for the six classification models on seven groups datasets

Groups of CT dataset		KNN	LR	DT	RF	SVM	MLP
NCCT	Test AUC	0.919	0.979	0.835	0.967	0.979	0.981
	Val. AUC	0.891	0.974	0.891	0.964	0.949	0.979
	P-value	0.622	0.871	0.396	0.929	0.408	0.937
APCT	Test AUC	0.923	0.981	0.878	0.957	0.988	0.964
	Val. AUC	0.978	0.96	0.953	0.97	0.973	0.96
	P-value	0.194	0.626	0.189	0.757	0.614	0.927
VPCT	Test AUC	0.904	0.922	0.779	0.913	0.912	0.901
	Val. AUC	0.933	0.968	0.909	0.969	0.972	0.98
	P-value	0.599	0.352	0.075	0.292	0.209	0.127
DPCT	Test AUC	0.921	0.968	0.706	0.971	0.974	0.955
	Val. AUC	0.887	0.975	0.875	0.98	0.975	0.952
	P-value	0.588	0.813	0.036	0.727	0.965	0.933
TKCT	Test AUC	0.893	0.959	0.866	0.947	0.963	0.972
	Val. AUC	0.979	1	0.945	1	0.999	1
	P-value	0.019	0.019	0.126	0.013	0.028	0.032
TNCT	Test AUC	0.883	0.958	0.898	0.976	0.96	0.964
	Val. AUC	0.909	0.959	0.935	0.961	0.957	0.934
	P-value	0.541	0.98	0.318	0.533	0.912	0.355
MMCT	Test AUC	0.857	0.925	0.844	0.941	0.931	0.947
	Val. AUC	0.926	0.97	0.916	0.96	0.969	0.955
	P-value	0.028	0.046	0.049	0.425	0.101	0.731

ISSN: 1471-2342