In vivomicrostructural heterogeneity of white matter lesions in Alzheimer’s disease using tissue compositional analysis of diffusion MRI data

Abstract

White matter hyperintensities (WMH) are commonly observed in elderly individuals, and are typically more prevalent in Alzheimer's disease subjects than in healthy subjects. These lesions can be identified on fluid attenuated inversion recovery (FLAIR) MRI, on which they are hyperintense compared to their surroundings. These MRI-visible lesions appear homogeneously hyperintense despite known heterogeneity in their pathological underpinnings, and are commonly regarded as surrogate markers of small vessel disease in in vivo studies.Consequently, the extent to which these lesions contribute to Alzheimer's disease remains unclear, likely due to the somewhat limited way in which these lesions are assessed in vivo.Diffusion MRI is sensitive to white matter microstructure, and might thus be used to investigate microstructural changes within WMH. In this study, we applied a method called single-shell 3-tissue constrained spherical deconvolution, which models white matter microstructure while also accounting for other tissue compartments, to investigate WMH in vivo. Diffusion MRI data and FLAIR images were obtained from Alzheimer's disease (n = 48) and healthy elderly control (n = 94) subjects from the Australian Imaging, Biomarkers and Lifestyle study of ageing. WMH were automatically segmented and classified as periventricular or deep lesions from FLAIR images based on their continuity with the lateral ventricles, and the 3-tissue profile of different classes of WMH was characterised by three metrics, which together characterised the relative tissue profile in terms of the white matter-, grey matter-, and fluid-like characteristics of the diffusion signal. Our findings revealed that periventricular and deep lesion classes could be distinguished from one another, and from normal-appearing white matter based on their 3-tissue profile, with substantially higher free water content in periventricular lesions than deep. Given the higher lesion load of periventricular lesions in Alzheimer's disease patients, the 3-tissue profile of these WMH could be interpreted as reflecting the more deleterious pathological underpinnings that are associated with disease.However, when alternatively classifying lesion sub-regions in terms of distance contours from the ventricles to account for potential heterogeneity within confluent lesions, we found that the highest fluid content was present in lesion areas most proximal to the ventricles, which were common to both Alzheimer's disease subjects and healthy controls. We argue that whatever classification scheme is used when investigating WMH, failure to account for heterogeneity within lesions may result in classification-scheme dependent conclusions. Future studies of WMH in Alzheimer's Disease would benefit from inclusion of microstructural information when characterising lesions.

Test-retest reliability and long-term stability of 3-tissue constrained spherical deconvolution methods for analyzing diffusion MRI data

“…Standard neuroimaging techniques do not provide quantifiable data on brain microstructural architecture, however this study has demonstrated a reproducible and reliable method for obtaining whole brain maps with quantifiable estimates of brain microstructure. This measure is stable enough to be used in longitudinal studies lasting at least up to three months; and provides information on a voxel-or region-wise basis for analysis of subcortical structures, lesions, or developing brains (Dhollander et al, 2017;Mito et al, 2018;Mito et al, 2019;Bastiani et al, 2019). Related microstructural analysis of free water signal fractions has been performed in the context of Parkinson's disease (Ofori et al, 2015;Burciu et al, 2016), Schizophrenia (Pasternak et al, 2012;Mandl et al, 2015), and concussion (Pasternak et al, 2014).…”

“…Pietsch et al, (2019) applied two different WM response functions representing mature and immature WM in a developing adolescent cohort to observe WM maturation. Mito et al, (2019) proposed to apply a statistical framework of compositional data analysis to analyze the full 3-tissue composition of WM-, GM-and CSF-like signal fractions directly to study microstructure in white matter lesions, following the initial suggestion of moving towards such WM/GM/CSF-like diffusion signal fraction interpretation by Dhollander et al (2017). In Aerts et al, (2019), this idea was furthermore used for the purpose of disentangling WM FODs representing infiltrated WM tracts in the presence of gliomas, so as to enable more reliable within-tumor tractography.…”

“…Finally, we upsampled the preprocessed data to 1.3⨉1.3⨉1.3mm 3 isotropic voxels (Greenspan, 2008;Kuklisova-Murgasova et al, 2012;Bastiani et al, 2019). These preprocessing steps are largely similar to those used in other recently published works (Bastiani et al, 2019;Pietsch et al, 2019;Mito et al, 2019;Aerts et al, 2019). Brain masks were obtained for all subjects by performing a recursive application of the Brain Extraction Tool (Avants et al, 2014).…”

“…The CSF-like compartment can in this context also be interpreted as a free-water (FW) compartment (Dhollander et al, 2017). Finally, each subject's three tissue compartments were then normalised to sum to 1 on a voxel-wise basis, resulting in the final 3-tissue signal fraction maps (Mito et al, 2019); the metrics for which we performed the test-retest analyses in this work.…”

“…Resolving these different compartments using either 3-tissue CSD method (i.e., MSMT-CSD or SS3T-CSD) holds value beyond improving WM tractography: it can also serve as a proxy for the evaluation of brain microstructure and tissue composition (Dhollander et al, 2017;Mito et al, 2018;Mito et al 2019). By interrogating brain voxels for diffusion signal patterns that look 'like' compositions of the diffusion signals represented by the WM/GM/CSF response functions, it might be possible to gain quantifiable information about microstructure ( Figure 1).…”