Habitat analysis of iron deposition in the basal ganglia for diagnosing cognitive impairment in chronic kidney disease: evidence from a case-control study

Table 3 Predictive performance of various models on train and test sets

Models	Sets	AUC	95 CI%	ACC	SEN	SPE	NPV	PPV
CNL-L	Training	0.51	0.40–0.63	0.49	0.47	0.51	0.45	0.53
	Test	0.46	0.29–0.64	0.52	0.46	0.60	0.48	0.58
CNL-R	Training	0.58	0.47–0.69	0.58	0.55	0.62	0.54	0.63
	Test	0.43	0.25–0.61	0.41	0.38	0.45	0.38	0.45
PUT-L	Training	0.55	0.43–0.66	0.47	0.42	0.53	0.44	0.51
	Test	0.57	0.40–0.75	0.55	0.54	0.55	0.50	0.59
PUT-R	Training	0.59	0.48–0.71	0.57	0.53	0.62	0.53	0.62
	Test	0.57	0.40–0.75	0.57	0.58	0.55	0.52	0.61
GPL-L	Training	0.44	0.33–0.56	0.47	0.47	0.47	0.43	0.51
	Test	0.49	0.31–0.67	0.52	0.42	0.65	0.48	0.59
GPL-R	Train	0.54	0.42–0.65	0.50	0.51	0.49	0.46	0.54
	Test	0.51	0.33–0.69	0.52	0.46	0.60	0.48	0.58
Basal ganglia	Training	0.93	0.87–0.98	0.89	0.87	0.92	0.86	0.92
	Test	0.77	0.63–0.91	0.71	0.67	0.75	0.65	0.76
Two habitats	Training	0.96	0.92-1.00	0.93	0.98	0.87	0.98	0.90
	Test	0.93	0.84-1.00	0.86	0.96	0.75	0.94	0.82
Three habitats	Training	0.94	0.90–0.99	0.88	0.91	0.85	0.89	0.88
	Test	0.69	0.53–0.85	0.66	0.58	0.75	0.60	0.74
Four habitats	Training	0.91	0.85–0.97	0.87	0.87	0.87	0.85	0.89
	Test	0.62	0.45–0.79	0.57	0.34	0.80	0.52	0.69
Five habitats	Training	0.98	0.86-1.00	0.92	0.93	0.92	0.92	0.93
	Test	0.80	0.65–0.95	0.77	0.75	0.80	0.73	0.82
Clinical	Training	0.83	0.75–0.91	0.80	0.83	0.78	0.81	0.80
	Test	0.79	0.65–0.93	0.75	0.70	0.79	0.76	0.74
Combined	Training	0.97	0.95-1.00	0.87	0.87	0.88	0.86	0.88
	Test	0.91	0.83–0.99	0.80	0.90	0.71	0.89	0.72

AUC: Area Under the Curve; ACC: Accuracy; CNL-L: Left Caudate Nucleus; CNL-R: Right Caudate Nucleus; CI: Confidence Interval; GPL-L: Left Globus Pallidus; GPL-R: Right Globus Pallidus; NPV: Negative Predictive Value; PPV: Positive Predictive Value; PUT-L: Left Putamen; PUT-R: Right Putamen; SEN: Sensitivity; SPE: Specificity

ISSN: 1471-2342