The Diagnostic Value of the Combined 3D Pseudo-Continuous Arterial Spin Labeling and Diffusion Kurtosis Imaging in Patients With Binswanger’s Disease

BACKGROUND AND PURPOSE: The clinical diagnosis of Binswanger’s disease (BD), a chronic progressive form of subcortical vascular dementia, remains challenging. 3D pseudo-continuous arterial-spin-labeling (pcASL) and diffusion kurtosis imaging (DKI) can quantitatively reveal the microcirculation changes and heterogeneity of white matter (WM), respectively. We thus aimed to determine the diagnostic value of the combined 3D-pcASL and DKI in BD. MATERIALS AND METHODS: A total of 35 patients with BD and 33 healthy controls underwent 3D-ASL and DKI experiments. The perfusion parameter of cerebral blood flow (CBF), diffusion parameters of fractional anisotropy (FA), mean/axial/radial diffusivity (MD/Da/Dr), and kurtosis parameters of anisotropy fraction of kurtosis (FAk) and mean/axial/radial kurtosis MK/Ka/Kr were obtained to quantitatively measure the parametric distributions of functional brain subregions. One-way analysis of variance and post hoc t-test were applied to explore the different distributions of DKI/ASL-derived parameters among brain subregions of BD. In addition, all region-specific DKI/ASL parameters were separately analyzed in Pearson correlation analysis to investigate the relationship with Mini-Mental State Examination (MMSE), a typical clinical scale for cognitive function assessment in patients with BD. RESULTS: FA/FAk/MK/Ka/Kr was significantly declined in all WM hyperintensities (WMHs) of BD compared with healthy controls, while the corresponding MD/Da/Dr was significantly increased (all p < 0.005). In addition, significant changes, similar to the WMHs of patients with BD, were also observed in almost all DKI parameters in WM normal areas and genu/splenium of the corpus callosum (GCC/SCC) in BD (p < 0.005). Finally, CBF was significantly reduced in all of the above regions we measured in patients with BD (p < 0.005). For patients with BD, MMSE showed a negative correlation with MD/Da in thalamus (r = −0.42/−0.58; p < 0.05), and a positive correlation with CBF in PWM/TWM (r = 0.49/0.39; p < 0.05). Using receiver operating characteristic (ROC) analysis, FA/FAk/Kr in GCC, CBF/FA/Dr/FAk in SCC, MD/Da/Ka in thalamus, and the combined FA/MD/Dr/CBF in TWM showed high accuracy [area under the curves (AUCs) 0.957/0.946/0.942/0.986] in distinguishing patients with BD from healthy controls. CONCLUSION: We found that combined DKI and 3D-ASL are helpful in diagnosing patients with BD, especially with FA, MD, Dr, and CBF in the temporal WM region. Additionally, the kurtosis parameters of DKI can sensitively monitor the potentially damaged WM areas in patients with BD patients, adding complementary clinical value.