Cortical microstructural abnormalities in dementia with Lewy bodies and their associations with Alzheimer's disease copathologies

Abstract

Dementia with Lewy bodies (DLB) frequently coexists with Alzheimer's disease pathology, yet the pattern of cortical microstructural injury and its relationship with amyloid, tau, and cerebrovascular pathologies remains unclear. We applied neurite orientation dispersion and density imaging (NODDI) to assess cortical microstructural integrity in 57 individuals within the DLB spectrum and 57 age- and sex-matched cognitively unimpaired controls by quantifying mean diffusivity (MD), tissue-weighted neurite density index (tNDI), orientation dispersion index (ODI), and free water fraction (FWF). Amyloid and tau levels were measured using PiB and Flortaucipir PET imaging. Compared to controls, DLB exhibited increased MD and FWF, reduced tNDI across multiple regions, and focal ODI reductions in the occipital cortex. Structural equation modeling revealed that APOE genotype influenced amyloid levels, which elevated tau, leading to microstructural injury. These findings highlight the role of AD pathology in DLB neurodegeneration, advocating for multi-target therapeutic approaches addressing both AD and DLB-specific pathologies.

Fig. 1. Topographic distributions of GM microstructural injury in DLB spectrum relative to matched CU individuals.

Pairwise group differences in (a) regional MD, (b) tNDI, (c) ODI, and (d) FWF were examined using conditional logistic regression models to account for the 1:1 matching between groups. T-statistics from the significant comparisons after adjusting for FDR are depicted on both volumetric templates and 3D glass brain renderings. In the DLB spectrum group, significantly elevated MD and FWF was accompanied by decreased tNDI across widespread GM regions, while reductions in ODI were more focal and preferentially localized to posterior-occipital brain regions. CU cognitively unimpaired, DLB dementia with Lewy bodies, FDR false discovery rate, GM gray matter, MD mean diffusivity. ODI orientation dispersion index, tNDI tissue-weighted neurite density index, FWF free water fraction.

Fig. 2. Structural equation modeling of pathological cascades associated with microstructural injury in DLB.

The diagram shows significant direct effects (p < 0.05) between age, APOE ε4 genotype, AD biomarkers (amyloid-β and tau), and composite measures of gray matter microstructure (MD, tNDI, ODI, and FWF) in the DLB spectrum. Pathways are color-coded by their origin: age (green), APOE (purple), amyloid-β (orange), and tau (red). Line thickness is proportional to the magnitude of the standardized coefficient, with values and significance levels shown (***p < 0.001, **p < 0.01, *p < 0.05). Variables were transformed for analysis: amyloid-β and tau were log-transformed, age was measured in decades, and mean diffusivity was scaled by 100. The model reveals both direct pathways (e.g., tau → MD) and indirect pathways (e.g., APOE → amyloid-β → tau) contributing to microstructural changes. To interpret the SEM findings within the context of anatomical patterns, the topography of each composite GM microstructure is provided next to the DTI and NODDI nodes (FDR q < 0.05 from pairwise group comparisons between DLB spectrum and CU controls). AD Alzheimer’s disease, APOE apolipoprotein E, CU cognitively unimpaired, DLB dementia with Lewy bodies, FWF free water fraction, FDR false discovery rate, GM gray matter, MD mean diffusivity, ODI orientation dispersion index, tNDI tissue-weighted neurite density index, SEM structural equation modeling.

Fig. 3. Regional associations between tau and GM microstructure integrity in DLB spectrum.

A T-statistics map showing significant associations between tau and MD after FDR correction. Yellow-red colors indicate positive associations. B T-statistics map showing significant associations between tau and tNDI after FDR correction. Blue colors indicate negative associations. C Scatter plot showing the positive relationship between log tau SUVR and predicted MD values in the middle temporal gyrus, with a 95% confidence interval shown in brown shading. D Scatter plot showing the negative relationship between log tau SUVR and predicted tNDI values in the middle temporal gyrus, with a 95% confidence interval shown in blue shading. All regression models were adjusted for age, APOE genotype, amyloid-β, gray matter volume (expressed as a % of TIV), and WMH (expressed as a % of TIV). APOE apolipoprotein E, DLB dementia with Lewy bodies, FDR false discovery rate, GM gray matter, MD mean diffusivity, SUVR standardized uptake value ratio, tNDI tissue-weighted neurite density index, TIV total intracranial volume, WMH white matter hyperintensities.

Fig. 4. Associations between gray matter microstructure metrices and clinical measures in dementia with Lewy bodies.

Scatterplots rerpesent age-adjusted correlations between composite gray matter microstructural metrics and clinical measures in the DLB spectrum group. A Associations with motor severity (B) Associations with global cognitive impairment. CDR-SB Clinical Dementia Rating-Sum of Boxes, DLB Dementia with Lewy Bodies, GM Gray Matter, MD Mean Diffusivity, tNDI tissue-weighted Neurite Density Index, ODI Orientation Dispersion Index, FWF Free Water Fraction, ROI Region of Interest, UPDRS-III Unified Parkinson’s Disease Rating Scale Part III.

Fig. 5. Graphical overview of the study design and key analyses.

The study design involved fitting diffusion-weighted imaging datasets with DTI and NODDI models to obtain whole-brain maps of MD, tNDI, ODI, and FWF. Participants underwent PET imaging of PiB and Flortaucipir to quantify amyloid-β and tau deposition, respectively. Regional gray matter volumes were derived from structural MRI using SPM12. WMHs were calculated from FLAIR using an automated segmentation technique. The voxelwise MD, tNDI, ODI, and FWF parameter maps were parcellated using the MCALT atlas to derive bilateral median values reflecting regional GM microstructural integrity. We employed conditional logistic models to compare regional MD, tNDI, ODI, and FWF between the DLB spectrum and CU groups. Subsequently, statistically significant ROIs from the group-wise comparisons were used as composite outcome measures in SEMs to determine their multivariate associations with age, WMH, APOE genotype, amyloid-β, and tau. In exploratory analyses, partial Pearson’s correlations were used to determine the degree to which composites of MD, tNDI, ODI, and FWF are related to disease severity and core features in the DLB spectrum. Supplemental regression analyses were used to explore the topographical associations of tau and amyloid PET meta-ROIs with regionally specific MD, tNDI, ODI, and FWF, irrespective of group differences. These models were adjusted for age, APOE genotype, gray matter volumes (expressed as the % of TIV), and WMH (expressed as the % of TIV). AD Alzheimer’s disease, APOE apolipoprotein E, CDR-SB clinical dementia rating-sum of boxes, CU cognitively unimpaired, DLB dementia with Lewy bodies, DTI diffusion tensor imaging, FDR false discovery rate, FLAIR fluid-attenuated inversion recovery, FWF free water fraction, GM gray matter, TIV total intracranial volume, MCALT Mayo clinic adult lifespan template, MD mean diffusivity, NODDI neurite orientation dispersion and density imaging, ODI orientation dispersion index, PET positron emission tomography, Amyloid-β Pittsburgh compound B, RBD REM sleep behavior disorder, ROIs regions of interest, SEM structural equation modeling, tNDI tissue-weighted neurite density index, UPDRS unified Parkinson’s disease rating scale, WMH white matter hyperintensity.