Table 2.

Comparison of the classification performance of the proposed 3DAN method with other methods for AD diagnosis in Strategies 1 and 2 in Table 1

Method	Training: In‐house, Testing: ADNI				Training: ADNI, Testing: In‐house
	ACC	SEN	SPE	AUC	ACC	SEN	SPE	AUC
3DAN	0.861	0.881	0.846	0.912	0.870	0.789	0.961	0.913
ResNet	0.853	0.863	0.846	0.907	0.860	0.759	0.974	0.910
VBM	0.712	0.947	0.538	0.907	0.821	0.667	0.996	0.908
ROI‐AAL	0.720	0.947	0.551	0.885	0.811	0.651	0.991	0.888
ROI‐BNA	0.744	0.960	0.584	0.901	0.813	0.651	0.996	0.894

Abbreviations: ACC = accuracy; SEN = sensitivity; SPE = specificity; AUC = area under the curve of the receiver operating characteristic; BNA = Brainnetome Atlas; AAL = anatomical automatic labeling; ROI = region of interest; VBM = voxel‐based morphometric.