Parenchymal CSF fraction is a measure of brain glymphatic clearance and positively associated with amyloid beta deposition on PET

Abstract

Introduction: Mapping of microscopic changes in the perivascular space (PVS) of the cerebral cortex, beyond magnetic resonance-visible PVS in white matter, may enhance our ability to diagnose Alzheimer's disease (AD) early.

Methods: We used the cerebrospinal fluid (CSF) water fraction (CSFF), a magnetic resonance imaging-based biomarker, to characterize brain parenchymal CSF water, reflecting microscopic PVS in parenchyma. We measured CSFF and amyloid beta (Aβ) using ¹¹ C Pittsburgh compound B positron emission tomography to investigate their relationship at both the subject and voxel levels.

Results: Our research has demonstrated a positive correlation between the parenchymal CSFF, a non-invasive imaging biomarker indicative of parenchymal glymphatic clearance, and Aβ deposition, observed at both individual and voxel-based assessments in the posterior cingulate cortex.

Discussion: This study shows that an increased parenchymal CSFF is associated with Aβ deposition, suggesting that CSFF could serve as a biomarker for brain glymphatic clearance, which can be used to detect early fluid changes in PVS predisposing individuals to the development of AD.

Highlights: Cerebrospinal fluid fraction (CSFF) could be a biomarker of parenchymal perivascular space. CSFF is positively associated with amyloid beta (Aβ) deposition at subject level. CSFF in an Aβ+ region is higher than in an Aβ- region in the posterior cingulate cortex. Correspondence is found between Aβ deposition and glymphatic clearance deficits measured by CSFF.

Keywords: Alzheimer's disease; aging effect; amyloid beta; glymphatic clearance; magnetic resonance imaging; magnetic resonance transverse relaxation time relaxometry; parenchymal cerebrospinal fluid fraction; positron emission tomography.

FIGURE 1

CSF fraction, intra‐extracellular water fraction, and myelin water fraction water maps derived from three‐water model using multi‐echo MR FAST‐T2 data. The maps are from a 72‐year‐old male subject with MCI/AD: (A) CSFF for long T2 component (T2: 200–2000 ms) of multi‐echo T2 signal; (B) IEWF for intermediate T2 component (T2: 20–200 ms); and (C) MWF for short T2 component (T2: 5–20 ms). AD, Alzheimer's disease; CSF, cerebrospinal fluid; CSFF, cerebrospinal fluid water fraction; IEWF, intra‐extracellular water fraction; MCI, mild cognitive impairment; MR, magnetic resonance; MWF, myelin water fraction.

FIGURE 2

The partial regression plots from multivariable analyses between beta‐amyloid deposition measured by PiB PET SUVR in ADmask, CSFF, and age: (A) partial regression plots between PiB PET SUVR and CSFF in ADmask; (B) partial regression plots between PiB PET SUVR and age. Black dots represent CN subjects and red triangles indicate MCI/AD subjects. We see that PiB SUVR positively associates with CSFF, and there is no association with age. AD, Alzheimer's disease; ADmask, Alzheimer's disease cortical mask; CN, cognitively normal; CSFF, cerebrospinal fluid water fraction; MCI, mild cognitive impairment; PET, positron emission tomography; PiB, Pittsburgh compound B; SUVR, standardized uptake value ratio.

FIGURE 3

The relationship between ADmask CSFF and age by diagnostic group: (A) the regression plot between CSFF and age by group; (B) boxplot of CSFF by diagnosis. (A) Shows that CSFF increases with age in both MCI/AD and CN subjects with positive correlation coefficients. There is an elevation of CSFF for MCI/AD compared to CN controlled for age (P < 0.01). (B) Shows significant increase in CSFF in MCI/AD compared to CN subjects. AD, Alzheimer's disease; ADmask, Alzheimer's disease cortical mask; CN, cognitively normal; CSFF, cerebrospinal fluid water fraction; MCI, mild cognitive impairment; PET, positron emission tomography; PiB, Pittsburgh compound B; SUVR, standardized uptake value ratio.

FIGURE 4

Example of Aβ+ and Aβ‐ regions in PCC; (A–C) are Aβ+ and Aβ– masks on T1w sagittal, coronal, and axial views, respectively; and (D–F) overlays of the masks on PiB PET SUVR map. The red region is for Aβ+ voxels and the blue region is for Aβ– voxels. Please note that the masks were dilated 1 voxel from the outside to enhance the visibility. Aβ, amyloid beta; PCC, posterior cingulate cortex; PET, positron emission tomography; PiB, Pittsburgh compound B; SUVR, standardized uptake value ratio.

FIGURE 5

The results of CSFF values in Aβ+ and Aβ– subregions of PCC by diagnostic group. (A) and (B) are plots and comparisons of CSFF in PCC Aβ+ and Aβ– subregions for CN subjects. (C) and (D) are plots of CSFF change between in PCC Aβ+ and Aβ– subregions for MCI/AD subjects. In (A) and (C), each point represents a subject. In (B) and (D), each pair of connected points represents the change of CSFF for one subject. We see in CN group, CSFF in Aβ+ regions are significantly higher than that for Aβ– regions. Aβ, amyloid beta; AD, Alzheimer's disease; CN, cognitively normal; CSFF, cerebrospinal fluid water fraction; MCI, mild cognitive impairment; PCC, posterior cingulate cortex; PET, positron emission tomography; PiB, Pittsburgh compound B; SUVR, standardized uptake value ratio.

FIGURE 6

The effect of Aβ+ threshold of PiB PET SUVR on CSFF in PCC. In the CN group, CSFF is different between Aβ+ and Aβ– regions in PCC with variable thresholds from 1.55 to 1.9. The CSFF differences on all thresholds are significantly greater than zero. We also observe an increasing CSFF in the Aβ+ region as the PiB SUVR threshold increases. Aβ, amyloid beta; AD, Alzheimer's disease; CN, cognitively normal; CSFF, cerebrospinal fluid water fraction; MCI, mild cognitive impairment; PCC, posterior cingulate cortex; PET, positron emission tomography; PiB, Pittsburgh compound B; SUVR, standardized uptake value ratio.

FIGURE 7

An illustration of perivascular space (PVS), intracellular space, and extracellular space shows the different microenvironments of water components. The fluid in PVS is CSF‐like water, which is freely movable with long MR T2 time. The water in extracellular space is constrained by the extracellular matrix, which decreases the MR T2 time of the water molecule (modified from Figure 1 in Lau et al. ⁴⁹ ). CSF, cerebrospinal fluid; MR, magnetic resonance.