The association between temporal-lobe tensor-based morphometry and plasma amyloid-β in mild cognitive impairment.

BACKGROUND

Alzheimer's disease (AD) is characterized by amyloid-β (Aβ) accumulation and brain atrophy; however, the assocation between plasma Aβ biomarkers and regional neurodegeneration remains unclear.

OBJECTIVE

We investigated whether plasma Aβ42, Aβ40, and the Aβ42/40 ratio are associated with temporal lobe atrophy measured using tensor-based morphometry (TBM) in cognitively healthy controls (HC) and participants with mild cognitive impairment (MCI).

METHODS

We analyzed longitudinal MRI and plasma biomarkers data from 29 participants from ADNI (HC = 14, MCI = 15) with imaging and blood samples available at baseline, 24 months, and 48 months. TBM Jacobian maps were summarized within temporal lobe regions of interest (ROIs). Associations between plasma Aβ measures and TBM-derived atrophy were examined with linear mixed-effects models, adjusting for age, sex, and APOE ε4 status, with false-discovery-rate correction.

RESULTS

Participants with MCI showed greater temporal lobe atrophy compared with HC people, with significantly lower TBM values at follow-up. Plasma Aβ42, Aβ40, and Aβ42/40 levels showed no consistent or robust differences between diagnostic groups. After covariate adjustment and FDR correction, no plasma Aβ-TBM associations were significant at baseline or 24 months. At 48 months, positive associations were identified between Aβ42 and temporal lobe atrophy (measure 2) in HC participants (β = 0.70, p = 0.046) and between Aβ40 and measure 2 in participants with MCI (β = 0.60, p = 0.036). In contrast, a negative association was observed between the Aβ42/40 and temporal lobe atrophy (measure 2) in MCI group (β = -0.53, p = 0.049).

CONCLUSIONS

TBM captured greater temporal lobe atrophy in participants with MCI compared with HC. Plasma amyloids showed only limited and inconsistent associations with temporal lobe atrophy over time. These findings suggest that plasma Aβ measures alone may not reliably reflect longitudinal regional neurodegeneration in early AD.