Comparison of 123I-MIBG myocardial scintigraphy, brain perfusion SPECT, and voxel-based MRI morphometry for distinguishing between dementia with Lewy bodies and Alzheimer's disease

Objective This study aimed to compare the diagnostic value of I-123-metaiodobenzylguanidine (MIBG) myocardial scintigraphy, N-isopropyl-p[I-123] iodoamphetamine (IMP) brain perfusion single-photon emission computed tomography (SPECT), and brain magnetic resonance imaging (MRI) voxel-based morphometry (VBM) for the differentiation of dementia with Lewy bodies (DLB) and Alzheimer's disease (AD). Methods Thirty-five and 34 patients with probable DLB and probable AD, respectively, were enrolled. All patients underwent I-123-MIBG myocardial scintigraphy, I-123-IMP brain perfusion SPECT, and brain MRI. For I-123-MIBG imaging, we calculated early and delayed heart-to-mediastinum (H/M) uptake ratios. Three-dimensional stereotactic surface projections (3D-SSP) were used to analyze the results of I-123-IMP SPECT. VBM with statistical parametric mapping 8 plus diffeomorphic anatomical registration using exponentiated Lie algebra (DARTEL) was used to analyze the brain MRI data. Results The area under the receiver operating characteristic curves (AUC) for discriminating DLB and AD was highest (0.882) for the delayed H/M ratio on I-123-MIBG scintigraphy. AUC for z-score measurement in the occipital lobe was 0.818 and that for the extent of gray matter (GM) atrophy in the whole brain was 0.788. AUC for the combination of 3D-SSP and VBM analysis was 0.836. The respective sensitivities and specificities for distinguishing DLB from AD were 97.1 and 100 % for the delayed H/M ratio using I-123-MIBG scintigraphy; 88.6 and 73.5 % for the occipital lobe z-score using 3D-SSP analysis; 85.7 and 64.7 % for the extent of whole brain GM atrophy using voxel-based MRI morphometry; and 91.4 and 76.5 % for the combination of 3D-SSP analysis and VBM. Conclusions I-123-MIBG myocardial scintigraphy was superior to brain perfusion SPECT and brain MRI using an advanced statistical technique to differentiate DLB and AD.