Biomarkers.

Area of Science:

• Neurology
• Neuroimaging
• Biomarkers

Background:

• Alzheimer's disease (AD) is heterogeneous and often co-occurs with cerebral small vessel disease (SVD).
• Previous research identified AD subtypes using imaging in AD-focused cohorts.
• This study defines imaging subtypes in distinct dementia cohorts with varying SVD burdens.

Purpose of the Study:

• To define imaging-derived subtypes in low-SVD and high-SVD dementia cohorts.
• To investigate if amyloid or free-water predicts vascular lesions (WMH, PVS) at baseline and over time within these subtypes.

Main Methods:

• Analyzed 262 individuals across two cohorts: TRIAD (low-SVD) and MITNEC-C6 (high-SVD).
• Quantified white matter hyperintensities (WMH) and enlarged perivascular spaces (PVS) using deep learning segmentation (segCSVD).
• Identified disease subtypes using the SuStaIn algorithm with markers including amyloid SUVR, WMH, PVS, and DTI-derived free-water.

Main Results:

• Identified 'vascular-first' and 'mixed' subtypes in both cohorts; 'amyloid-first' subtype found in the low-SVD cohort.
• Amyloid correlated with WMH volume in 'mixed' subtypes; free-water associated with WMH and PVS volumes across subtypes.
• Baseline amyloid predicted faster WMH growth in 'vascular-first' subtypes; free-water predicted faster WMH growth in 'mixed' subtypes.

Conclusions:

• The 'mixed' subtype, common in clinical settings, shows amyloid associated with baseline WMH.
• Free-water levels, not amyloid, predict future WMH growth in the 'mixed' subtype.
• Amyloid or free-water did not predict WMH or PVS growth in the 'amyloid-first' subtype.