Table 4.

Classification results per group using the unenhanced T1-weighted images, features determined with the contrast-enhanced T1-weighted images. AUC: area under the receiver operator characteristic curve. Sens.: sensitivity. Spec.: specificity. Acc.: accuracy.

Number of	Training data				Independent test data
features	AUC	Sens.	Spec.	Acc.	AUC	Sens.	Spec.	Acc.
1	0.7061	0.6027	0.6914	0.6479	0.6308	0.4458	0.7517	0.5988
2	0.7108	0.6188	0.6802	0.6501	0.6187	0.4033	0.7775	0.5904
3	0.8027	0.6759	0.7418	0.7095	0.5387	0.4267	0.6133	0.5200
4	0.8091	0.6876	0.7508	0.7198	0.5378	0.3742	0.5733	0.4738
5	0.8571	0.7469	0.7976	0.7728	0.5649	0.5875	0.4817	0.5346
6	0.8829	0.7549	0.8043	0.7801	0.6099	0.5825	0.5650	0.5738
7	0.8810	0.7604	0.8041	0.7827	0.5956	0.5942	0.5533	0.5738
8	0.8780	0.7614	0.7908	0.7764	0.6246	0.5833	0.5842	0.5838
9	0.8754	0.7688	0.7898	0.7795	0.6254	0.5875	0.5642	0.5758
10	0.8784	0.7700	0.8037	0.7872	0.5856	0.5392	0.5675	0.5533
11	0.8729	0.7716	0.7869	0.7794	0.6158	0.5450	0.5858	0.5654
12	0.8602	0.7527	0.7806	0.7669	0.5826	0.5025	0.5650	0.5338
13	0.8737	0.7633	0.7982	0.7811	0.5865	0.5300	0.5692	0.5496
14	0.8664	0.7612	0.7988	0.7804	0.5594	0.4575	0.5717	0.5146
15	0.8718	0.7667	0.8051	0.7863	0.5542	0.4475	0.5850	0.5163