Pre-treatment Perivascular Diffusivity Predicts Amyloid Clearance Rate in Lecanemab-Treated Patients: A Retrospective
Rafael Imbett-Martinez1, Jeremy N Ford2, Nelson H Gil2
1From the Division of Neuroradiology, Department of Radiology, Massachusetts General Hospital, Boston, MA, USA. rmartinez34@mgh.harvard.edu.
Background And Purpose:
To determine whether perivascular diffusivity, as interrogated by the ALPS-index, predicts amyloid clearance rates in patients undergoing lecanemab therapy.
Materials And Methods:
Patients were included if they received lecanemab for ≥6 months, had at least two florbetaben PET scans, and underwent baseline diffusion MRI as part of the anti-amyloid protocol. Of 45 eligible participants, 23 had a DTI sequence (≥20 diffusion-encoding directions) and were included in the analysis; DTI was not acquired in the remaining 22. Diffusivities along orthogonal fiber axes were extracted to compute the ALPS-index. The outcome variable was amyloid clearance rate, defined as monthly change in centiloid units between baseline and follow-up PET. Simple and multiple linear regression models examined associations between baseline ALPS-index and amyloid clearance rate, with and without adjustment for baseline centiloid.
Results:
Baseline ALPS-index predicted amyloid clearance rate (β = -6.53, P = .03, R2 = 0.20). After adjusting for baseline centiloid, both ALPS-index (β = -4.57, P = .03) and baseline centiloid (β = -0.0287, P < .001) remained independent predictors; the adjusted model explained a substantially greater proportion of the variance (R2 = 0.63). Higher baseline ALPS-index and greater amyloid burden were each independently associated with faster centiloid decline.
Conclusions:
In this retrospective cohort, a higher baseline ALPS-index was associated with faster amyloid clearance, suggesting that pre-treatment perivascular diffusivity may help explain inter-individual variability in response to lecanemab. Larger prospective cohorts are needed to establish whether the baseline ALPS-index can serve as a predictor of treatment response.