Arterial spin labeling signal in the CSF: Implications for partial volume correction and blood–CSF barrier characterization. (9th November 2022)

• Record Type: Journal Article
• Authors: Petitclerc, Léonie; Hirschler, Lydiane; Örzsik, Balázs; Asllani, Iris; van Osch, Matthias J. P.
• Is Part Of: NMR in biomedicine. Volume 36:Number 3(2023)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-11-09
• Subjects: arterial spin labeling -- blood–CSF barrier -- brain clearance -- glymphatics -- neurofluids -- partial volume correction; Nuclear magnetic resonance -- Periodicals; Magnetic Resonance Spectroscopy -- Periodicals; 574
• DOI: 10.1002/nbm.4852
• Languages: English
• ISSNs: 0952-3480

Abstract

Abstract : For better quantification of perfusion with arterial spin labeling (ASL), partial volume correction (PVC) is used to disentangle the signals from gray matter (GM) and white matter within any voxel. Based on physiological considerations, PVC algorithms typically assume zero signal in the cerebrospinal fluid (CSF). Recent measurements, however, have shown that CSF‐ASL signal can exceed 10% of GM signal, even when using recommended ASL labeling parameters. CSF signal is expected to particularly affect PVC results in the choroid plexus. This study aims to measure the impact of CSF signal on PVC perfusion measurements, and to investigate the potential use of PVC to retrieve pure CSF‐ASL signal for blood–CSF barrier characterization. In vivo imaging included six pCASL sequences with variable label duration and post‐labeling delay (PLD), and an eight‐echo 3D‐GRASE readout. A dataset was simulated to estimate the effect of CSF‐PVC with known ground‐truth parameters. Differences between the results of CSF‐PVC and non‐CSF‐PVC were estimated for regions of interest (ROIs) based on GM probability, and a separate ROI isolating the choroid plexus. In vivo, the suitability of PVC‐CSF signal as an estimate of pure CSF was investigated by comparing its time course with the long‐TE CSF signal. Results from both simulation and in vivo data indicated that including the CSF signal in PVC improves quantification of GM CBF by approximately 10%. In simulated data, this improvement was greater for multi‐PLD (model fitting) quantification than for single PLD (~1–5% difference). In the choroid plexus, the difference between CSF‐PVC and non‐CSF‐PVC was much larger, averaging around 30%. Long‐TE (pure) CSF signal could not be estimated from PVC CSF signal as it followed a different time course, indicating the presence of residual macrovascular signal in the PVC. The inclusion of CSF adds value to PVC for more accurate measurements of GM perfusion, and especially for quantification of perfusion in the choroid plexus and study of the glymphatic system. Abstract : ASL signal in the CSF is non‐zero, and including this in partial volume correction leads to better estimates of GM CBF. Differences in GM signal between CSF‐PVC and non‐CSF‐PVC averaged 10% over all the GM, and 30–40% in the choroid plexus. CSF‐PVC is especially recommended for choroid plexus perfusion measurements.