Table 5.

Classification results per group using the unenhanced T1-weighted images, features determined with the unenhanced 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	0.7458	0.5898	0.7298	0.6612	0.5524	0.4983	0.6442	0.5713
2	0.8442	0.7429	0.7616	0.7524	0.6100	0.5958	0.5700	0.5829
3	0.8607	0.7665	0.7492	0.7577	0.5458	0.5633	0.4958	0.5296
4	0.8818	0.8004	0.7843	0.7922	0.6319	0.5925	0.6150	0.6038
5	0.9200	0.8349	0.8441	0.8396	0.6387	0.7350	0.5425	0.6388
6	0.9059	0.8369	0.8157	0.8261	0.6505	0.6158	0.5775	0.5967
7	0.9139	0.8557	0.8220	0.8385	0.6300	0.5783	0.5550	0.5667
8	0.9202	0.8706	0.8212	0.8454	0.5605	0.5033	0.5733	0.5383
9	0.9605	0.9253	0.8747	0.8995	0.5644	0.5358	0.5450	0.5404
10	0.9567	0.9200	0.8645	0.8917	0.5609	0.5083	0.5892	0.5488
11	0.9578	0.9147	0.8788	0.8964	0.6421	0.5542	0.6083	0.5813
12	0.9543	0.9127	0.8708	0.8913	0.6372	0.5742	0.6233	0.5988
13	0.9621	0.9220	0.8825	0.9019	0.6177	0.5842	0.6058	0.5950
14	0.9663	0.9276	0.8990	0.9130	0.6205	0.5792	0.6058	0.5925
15	0.9559	0.9145	0.8743	0.8940	0.6111	0.5867	0.6000	0.5933