Brain protein burden is related to intravoxel incoherent motion: PET-MR imaging study.
Authors
Affiliations (7)
Affiliations (7)
- Department of Neurology and Neurological Sciences, Stanford University School of Medicine, Stanford, CA, United States.
- Department of Radiology, Stanford University School of Medicine, Stanford, CA, United States.
- Wu Tsai Neuroscience Institute, Stanford University, Stanford, CA, United States.
- Phil and Penny Knight Initiative for Brain Resilience, Stanford University, Stanford, CA, United States.
- Biosciences Division, SRI International, Menlo Park, CA, United States.
- Department of Psychology, Palo Alto University, Palo Alto, CA, United States.
- Department of Neurosurgery, Stanford University School of Medicine, Stanford, CA, United States.
Abstract
Dysfunction in brain protein clearance mechanisms is thought to contribute to many neurodegenerative diseases, yet non-invasive assessment of these mechanisms in humans remains challenging. This study is the first to examine whether intravoxel incoherent motion (IVIM) diffusion MRI metrics, measures of water diffusion and fluid dynamics, are associated with pathological protein accumulation and cognition in aging individuals, and hence whether they serve as a proxy for brain waste clearance function. We analyzed data from 94 participants (<i>n =</i> 45 <i>β</i>-amyloid positive) who underwent simultaneous PET/MRI scans to calculate three key IVIM metrics: D (true diffusion coefficient), D* (pseudo-diffusion coefficient reflecting perfusion), and <i>f</i> (perfusion fraction) within 98 regions of interest. A machine learning model was trained to identify the most informative IVIM features for predicting <i>β</i>-amyloid (Aβ) status. Selected features were then evaluated for correlations with protein burden (Aβ and tau) and cognitive performance. The model identified a subset of 25 key features that effectively predicted Aβ status, achieving a predictive accuracy of 80.0% on unseen data. Regions with important IVIM features aligned with previously identified Aβ-affected regions and showed significant correlations with Aβ burden (r = 0.53, <i>p <</i> 0.0001) and tau burden (r = 0.61, <i>p <</i> 0.0001). A significant negative correlation was observed between IVIM features and cognitive decline (r = -0.60, <i>p <</i> 0.0001). When stratified by Aβ status, this correlation remained significant only in the Aβ-positive group (r = -0.61, <i>p <</i> 0.0001), but not in the Aβ-negative group. IVIM-derived metrics (D, D*, and <i>f</i>), which measure water diffusion and perfusion dynamics in the brain, may be valuable non-invasive biomarkers of protein accumulation and associated cognitive decline in the aging human brain.