Modeling the Properties of White Matter Tracts Using Diffusion Tensor Imaging to Characterize Patterns of Injury in Aging and Neurodegenerative Disease

Abstract

Diffusion tensor imaging (DTI) is a relatively novel magnetic resonance-based imaging methodology that can provide valuable insight into the microstructure of white matter tracts of the brain. In this paper, we evaluated the reliability and reproducibility of deriving a semi-automated pseudo-atlas DTI tractography method vs. standard atlas-based analysis alternatives, for use in clinical cohorts with neurodegeneration and ventriculomegaly. We showed that the semi-automated pseudo-atlas DTI tractography method was reliable and reproducible across different cohorts, generating 97.7% of all tracts. However, DTI metrics obtained from both methods were significantly different across the majority of cohorts and white matter tracts (p < 0.001). Despite this, we showed that both methods produced patterns of white matter injury that are consistent with findings reported in the literature and with DTI profiles generated from these methodologies. Scatter plots comparing DTI metrics obtained from each methodology showed that the pseudo-atlas method produced metrics that implied a more preserved neural structure compared to its counterpart. When comparing DTI metrics against a measure of ventriculomegaly (i.e., Evans’ Index), we showed that the standard atlas-based method was able to detect decreasing white matter integrity with increasing ventriculomegaly, while in contrast, metrics obtained using the pseudo-atlas method were sensitive for stretch or compression in the posterior limb of the internal capsule. Additionally, both methods were able to show an increase in white matter disruption with increasing ventriculomegaly, with the pseudo-atlas method showing less variability and more specificity to changes in white matter tracts near to the ventricles. In this study, we found that there was no true gold-standard for DTI methodologies or atlases. Whilst there was no congruence between absolute values from DTI metrics, differing DTI methodologies were still valid but must be appreciated to be variably sensitive to different changes within white matter injury occurring concurrently. By combining both atlas and pseudo-atlas based methodologies with DTI profiles, it was possible to navigate past such challenges to describe white matter injury changes in the context of confounders, such as neurodegenerative disease and ventricular enlargement, with transparency and consistency.

Alcohol Use Disorder and Dementia: A Review

“… 207 – 211 Gross white matter volume remains relatively stable across adulthood; 201 , 212 – 214 however, appropriate imaging modalities (e.g., fluid-attenuated inversion recovery, diffusion tensor imaging) demonstrate more hyperintense inclusions (i.e., white matter hyperintensities [WMH]), 215 – 218 and microstructural compromise in older relative to younger individuals. 219 – 222 …”

“…[207][208][209][210][211] Gross white matter volume remains relatively stable across adulthood; 201,[212][213][214] however, appropriate imaging modalities (e.g., fluid-attenuated inversion recovery, diffusion tensor imaging) demonstrate more hyperintense inclusions (i.e., white matter hyperintensities [WMH]), [215][216][217][218] and microstructural compromise in older relative to younger individuals. [219][220][221][222] Cross-sectional neuroimaging reports support AUD-related volume shrinkage in specific brain structures, including frontal, temporal, and parietal cortices; diencephalon; brain stem; and cerebellum. [223][224][225][226][227][228][229] In contrast to results of postmortem analyses of neuronal numbers, neuroimaging studies describe significant volume deficits in people with AUD, relative to healthy controls, in hippocampus and basal ganglia (i.e., caudate, putamen, nucleus accumbens) that may be accounted for by white matter compromise.…”

Structural volumetric and Periodic Table DTI patterns in Complex Normal Pressure Hydrocephalus—Toward the principles of a translational taxonomy

“…The challenges of utilizing DTI to study the spectrum of NPH encompass several technical factors, biological considerations and scientific/clinical concerns. They can be summarized as the following conceptual list of problems: ( Keong et al, 2016 ; Kok et al, 2022 )–(i) “the problem of the scanner”—DTI measures are dependent on machine-specific/technical specifications for scanning acquisition and the effect of such variations between sites may be hard to quantify/correct for, (ii) “the problem of gold standard”—DTI output is dependent upon processing software techniques for which there are varying advantages and disadvantages but no single, unifying standard, (iii) “the problem of DTI methodology”—DTI metrics are subject to biological confounders such as multiple pathophysiological processes or crossing fibers occurring within the sample/region-of-interest, (iv) “the problem of cohorts”—DTI results can be inconsistent both within and across patient groups within the same disease process, as well as over time, (v) “the problem of consistency of interpretation”—DTI results can appear contradictory across the full panel of DTI measures even within known functional neuroanatomical groupings, (vi) “the problem of lack of samples”—there are insufficient published DTI samples of clinical cohorts who can represent distinct milestones within the spectrum of reversible to irreversible injury and (vii) “the problem of lack of comparators”—unlike structural imaging measures, individual DTI metrics are not clinically comparable across sites, leading to a lack of baseline reference values to be used in common.…”