Characterizing the Origin of the Arterial Spin Labelling Signal in MRI Using a Multiecho Acquisition Approach

Abstract

Arterial spin labelling (ASL) can noninvasively isolate the MR signal from arterial blood water that has flowed into the brain. In gray matter, the labelled bolus is dispersed within three main compartments during image acquisition: the intravascular compartment; intracellular tissue space; and the extracellular tissue space. Changes in the relative volumes of the extracellular and intracellular tissue space are thought to occur in many pathologic conditions such as stroke and brain tumors. Accurate measurement of the distribution of the ASL signal within these three compartments will yield better understanding of the time course of blood delivery and exchange, and may have particular application in animal models of disease to investigate the relationship between the source of the ASL signal and pathology. In this study, we sample the transverse relaxation of the ASL perfusion weighted and control images acquired with and without vascular crusher gradients at a range of postlabelling delays and tagging durations, to estimate the tricompartmental distribution of labelled water in the rat cortex. Our results provide evidence for rapid exchange of labelled blood water into the intracellular space relative to the transit time through the vascular bed, and provide a more solid foundation for cerebral blood flow quantification using ASL techniques.

Measuring Biexponential Transverse Relaxation of the ASL Signal at 9.4 T to Estimate Arterial Oxygen Saturation and the Time of Exchange of Labeled Blood Water into Cortical Brain Tissue

“…In this study, we find that the slow T2 component of the ASL signal closely matches the T2 of the control signal which suggests that: (1) the signal is not weighted toward the EC space and (2) the labeled spins rapidly equilibrate between the IC and EC compartments on the timescale of the TIs. The reason that the ASL signal does not appear to be weighted to the EC space in this work may be because the EC and IC T2 values are similar at 9.4 T or it may reflect the differences in the labeling technique used and the associated magnetization transfer effects (FAIR in this work and CASL in Wells et al 18 ). This finding may also reflect incomplete suppression of the IV signal using diffusion gradients with a b value of 40 s/mm 2 in our previous study.…”

“…[10][11][12]18 The mean DM IV / (DM IV þ DM EV ) ratio as measured from the biexponential transverse decay curve was 0.39 ± 0.04, which is significantly greater (P ¼ 0.0002) than the ratio as estimated from the diffusion sensitized measurements which may reflect the difficulty of separating the microvascular and capillary signal from the EV compartment by applying diffusion gradients on one direction as described by Kim and Kim. 12 (Part ii) Multiple Inversion Time Acquisitions (n ¼ 9) The mean ( ± s.e.m.)…”

“…22 Thus, it becomes feasible to separate the IV and EV distribution of the ASL signal by sampling its transverse decay as we have shown in this work. In our previous study, 18 we observed a consistently higher T2 app DM (with the inclusion of vascular crusher gradients with a b value of 40 s/ mm 2 ) in comparison with the T2 of the control signal. We interpreted this as evidence that the ASL signal was weighted toward the EC space (EV EC tissue space) relative to the IC (EV IC tissue space).…”

“…22 It should be noted that the accuracy of the oxygen saturation measurements reported in this work is dependent on the accuracy and in vivo applicability of the data reported in Lee et al 22 In future studies, the accuracy of the technique may benefit from integration of a calibration protocol to measure the T2 and oxygen saturation of systemic arterial blood, such as an 'arterial version' of the TRUST method described in Lu and Ge. 41 Previously, we investigated the time course of delivery and exchange of labeled blood water in the rat brain by measuring the T2 of the CASL signal with and without diffusion weighting over a range of tagging durations and post-labeling delay times at 2.35 T. 18 A multiecho CASL sequence was implemented with global 1801 refocusing pulses to prevent underestimation of the T2 app DM because of fast moving spins in the vessels. Our measurements suggested that the T2 of labeled water in the vasculature (i.e., arterioles/capillaries) was similar to that of the tissue.…”

“…However, each study made assumptions that may limit the accuracy of the physiologic interpretation of their data. In our previous study at 2.35 T, 18 we observed that the T2 of labeled water in the vessels was similar to that of the tissue. Consequently, we were not able to estimate the compartmentation of the ASL signal in the IV and EV spaces based on measurements of the transverse decay with any practical precision.…”