Brain glymphatic fluid mapping in Alzheimer's disease: a human MRI and PET study

Glymphatic system has been identified as a fluid exchange network in brain parenchymal for removal of toxic metabolites in Alzheimer's disease. However, a clinically feasible, accurate, and non-invasive imaging technique for mapping global glymphatic fluid distribution throughout the entire brain and monitoring its dysfunction in Alzheimer's disease is currently lacking. This cross-sectional retrospective study aims to compare three MRI-based measures of the glymphatic system structural alterations, a novel multi-echo T2 relaxometry-based parenchymal CSF (pCSF) mapping, T2 weighted-based segmented perivascular space burden, and diffusion tensor imaging-based free-water mapping. We evaluated their cross-correlation and investigated their respective association with PET measured beta-amyloid deposition using ¹¹C-PiB PET. A total of 29 subjects (18 Female, age: 70.07 ± 8.73 years old), among them, 16 were cognitively normal and 13 were mild cognitive impairment or Alzheimer's disease, underwent both MRI and ¹¹C-PiB PET scans. Parenchymal CSF mapping, diffusion-based free-water mapping and perivascular spaces burden were generated from their respective MRI. Age and sex effects and group differences of these biomarkers were evaluated. The associations among these measures and Aβ deposition on ¹¹C-PiB PET were analysed and compared. Our analysis demonstrated moderate correlations between pCSF, diffusion-based free-water mapping and perivascular space burden, suggesting these biomarkers capture overlapping yet distinct aspects of glymphatic fluid distribution. Importantly, linear regression analyses revealed that pCSF exhibited a significantly stronger positive association with beta-amyloid deposition (white matter: t = 3.536, P = 0.002, R ² = 0.446; Alzheimer's disease-related meta regions: t = 4.510, P < 0.001, R ² = 0.541), compared with diffusion-based free-water mapping in white matter (t = 0.843, P = 0.407, R ² = 0.191) and perivascular space burden (t = 0.422, P = 0.677, R ² = 0.174), after adjusting for age and sex. This study identifies pCSF mapping, derived from multi-echo T2 data with clinically feasible 5 mins scan FAST-T2 sequence, as a potential non-invasive imaging biomarker for assessing glymphatic dysfunction in Alzheimer's disease. The superior sensitivity of pCSF mapping arises from its direct quantification of glymphatic fluid at the water molecular level based on free CSF-specific T2 relaxation properties. These results suggest that pCSF could be useful in the monitoring of Alzheimer's disease progression as beta-amyloid accumulating and predicting response to anti-amyloid therapies, potentially leading to better diagnostic strategies and therapeutic interventions of Alzheimer's disease.