Pilot study of MRI white matter tissue properties in Alzheimer’s, vascular and mixed dementias

Abstract

AbstractBackgroundAlzheimer’s disease (AD) with cerebrovascular disease is known as ‘mixed’ dementia (MixD).[Wang_2012][Langa_2004] MixD can have heterogeneous clinical/imaging presentations. This makes it difficult to distinguish MixD from AD or vascular dementia (VaD) using structural markers such as atrophy or white matter hyperintensity (WMH) volumes. [Suri_2014] We explored whether WM tissue properties on MRI, represented by R2* and diffusion‐tensor (DTI) images, could distinguish MixD from AD or VaD.MethodN=17 participants (cross‐sectional; 7 MixD/5 Subcortical VaD/5 AD; Sex: 11M/6F; Age: 75±8yrs) were scanned on a 3T Philips Achieva. WMHs were segmented on 3D‐Fluid Attenuated Inversion Recovery images. T1‐weighted MP‐RAGE images were segmented into the grey/white‐matters using SPM12. These outputs were combined to construct WMH and normal‐appearing WM (NAWM) masks. DTI images were processed using FSL. R2* images were computed using in‐house software. For each participant, the average R2*, fractional anisotropy (FA) and mean diffusivity (MDf) values were calculated within the WMH and NAWM masks.ResultAverage WMH volumes were (mean±SD) AD: 5191±4693mm3, MD: 34680±17059mm3 (p<.05 compared to AD), SVaD: 20896±14920mm3 (p>.05 compared to MixD or AD). A linear model was used to predict the measured R2*, FA or MDf values from the diagnosis subtypes, adjusting for age and sex. R2* results: Pairwise t‐tests revealed significantly lower R2* values within the WMHs compared to NAWM (all subtypes p<0.0005). MixD had significantly lower WMH R2* values compared to AD (p=0.01) or VaD (p=0.02) subgroups. DTI results: Pairwise t‐tests revealed significantly higher MDf values within the WMHs compared to NAWM (all subtypes p<0.009). FA values were significantly lower within the WMHs compared to NAWM for the MixD (p=0.0002) and VaD (p=0.03) but not the AD (p=0.09) subtype.ConclusionOur MixD cohort was characterized by potentially disrupted fiber integrity (represented by decreased FA) and increased water content (represented by lower R2*) within the WMH areas. These abnormalities likely represent etiologies caused by both neurodegenerative and cerebrovascular factors. Future studies that incorporate measures of neurodegeneration or neuroinflammation, such as biofluid markers, may help to further characterize WM tissue abnormalities in MixD compared to those found in ‘pure’ AD or VaD.