Microstructural Bias in the Assessment of Periventricular Flow as Revealed in Postmortem Brains

Background Diffusion tensor imaging (DTI) analysis along the perivascular space (ALPS) has been proposed to assess global glymphatic activity in several neurologic diseases. However, it is unclear whether this measurement is biased by fiber microstructure. Purpose To determine the impact of fiber microstructure on DTI-ALPS in postmortem brains with or without Alzheimer disease (AD) and to evaluate the corrected ALPS in living participants. Materials and Methods From June 2021 to August 2023, amyloid β (Aβ)-positive and Aβ-negative postmortem human brains underwent DTI-ALPS. DTI-ALPS and microstructural asymmetries in the postmortem samples were compared between pathologic groups with use of two-way analysis of variance and correlated with age with use of Pearson correlation. The corrected ALPS index, where the age-dependent and pathologic abnormality-related microstructural effect was removed, was applied to in vivo data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) and was compared among diagnostic groups with use of analysis of covariance controlling for clinical covariates, with a family-wise error-corrected P < .05 considered indicative of a statistically significant difference. Results Eighteen donors with or without AD (mean age, 69 years ± 13 SD; 15 men) and 110 ADNI participants (mean age, 76 years ± 6.8; 65 women) were evaluated. Microstructural asymmetry of the projection fibers was higher in Aβ-negative postmortem specimens than Aβ-positive (Aβ-negative mean, 0.98; Aβ-positive mean, 0.95; P = .045), while asymmetry of the association fibers correlated negatively with age (r = -0.48; P = .044). In the ADNI data, the uncorrected ALPS correlated positively with the Preclinical Alzheimer's Cognitive Composite (PACC) score (r = 0.30; P = .001) and negatively with florbetapir (AV45) PET (r = -0.27; P = .005). For the corrected ALPS, the correlation with AV45 PET was not statistically significant (r = -0.18; P = .059), and the correlation with the PACC score (r = 0.26; P = .006) decreased (Steiger test z = 2.99; P = .003). Conclusion Microstructural asymmetry observed in postmortem DTI-ALPS varied with Aβ pathologic characteristics and age, and after the minimization of microstructural bias, the corrected ALPS correlated less with AD markers in vivo. © RSNA, 2025 Supplemental material is available for this article. See also the editorial by Rovira and Pareto in this issue.