White matter microstructural metrics are sensitively associated with clinical staging in Alzheimer's disease

Abstract

Introduction: White matter microstructure may be abnormal along the Alzheimer's disease (AD) continuum.

Methods: Diffusion magnetic resonance imaging (dMRI) data from the Alzheimer's Disease Neuroimaging Initiative (ADNI, n = 627), Baltimore Longitudinal Study of Aging (BLSA, n = 684), and Vanderbilt Memory & Aging Project (VMAP, n = 296) cohorts were free-water (FW) corrected and conventional, and FW-corrected microstructural metrics were quantified within 48 white matter tracts. Microstructural values were subsequently harmonized using the Longitudinal ComBat technique and inputted as independent variables to predict diagnosis (cognitively unimpaired [CU], mild cognitive impairment [MCI], AD). Models were adjusted for age, sex, race/ethnicity, education, apolipoprotein E (APOE) ε4 carrier status, and APOE ε2 carrier status.

Results: Conventional dMRI metrics were associated globally with diagnostic status; following FW correction, the FW metric itself exhibited global associations with diagnostic status, but intracellular metric associations were diminished.

Discussion: White matter microstructure is altered along the AD continuum. FW correction may provide further understanding of the white matter neurodegenerative process in AD.

Highlights: Longitudinal ComBat successfully harmonized large-scale diffusion magnetic resonance imaging (dMRI) metrics.Conventional dMRI metrics were globally sensitive to diagnostic status.Free-water (FW) correction mitigated intracellular associations with diagnostic status.The FW metric itself was globally sensitive to diagnostic status. Multivariate conventional and FW-corrected models may provide complementary information.

Keywords: diagnosis; diffusion MRI; free‐water; harmonization; white matter.

FIGURE 1

Tractography templates used in the present study. This study leveraged 48 well‐established tractography templates of the association (A), limbic (B), projection (C), occipital transcallosal (TC) (D), parietal TC (E), motor TC, and prefrontal TC tracts.

FIGURE 2

Demonstration of free‐water (FW) correction and Longitudinal ComBat harmonization in covariate‐matched participants. Propensity score matching was conducted to find covariate‐matched participants in the Alzheimer's Disease Neuroimaging Initiative (ADNI), Baltimore Longitudinal Study of Aging (BLSA), and Vanderbilt Memory and Aging Project (VMAP) cohorts. It is notable that the propensity score matching was conducted on a single protocol for each cohort and participants were matched to balance age, sex, education race/ethnicity, diagnosis, apolipoprotein E (APOE) ε4 carrier status, and APOE ε2 carrier status. Conventional dMRI measures are shown for the cingulum tract template (A), and distributions of the raw, conventional measures in addition to the associations with age are shown (B). The dMRI scans were subsequently FW corrected and Longitudinal ComBat harmonization was conducted on the microstructural values. Distributions of the harmonized, FW‐corrected measures in addition to the associations with age are shown (C).

FIGURE 3

Diagnostic differences in white matter microstructure. The strengths (i.e., absolute z‐values) are shown for the relationship between diagnostic category and conventional (A, top row) and free‐water (FW) corrected (A, second row) white matter microstructure. Limbic tract diagnostic differences for the conventional (B) and FW‐corrected (C) measures are illustrated, which show prominent differences in the fornix and cingulum. FW for the cingulum, fornix, and inferior longitudinal fasciculus (ILF) are shown (D).

FIGURE 4

Multivariate logistic regression on diagnostic category. The top 10 neuroimaging features and prediction probabilities from a recursive feature elimination analysis are shown for the conventional (A) and free‐water (FW)–corrected (B) multivariate regression analyses.