Feasibility of arterial spin labeling on a 1T open MRI scanner

To determine the clinical feasibility of arterial spin labeling (ASL) on a 1T open bore scanner. First, the optimal postlabeling delay (PLD) at 1T was determined (n = 5), with and without vascular crushing. Second, the effect of different labeling approaches (pseudo-continuous ASL [pCASL] vs. pulsed ASL [PASL]), background suppression (BSup) and readout options (GRASE vs. EPI) was investigated (n = 9). Each effect was quantified by calculating the signal-to-noise ratio (SNR), convergence, and number of significant gray matter (GM) voxels in the ASL images. Finally, an example of an obese volunteer who could not have been scanned in a cylindrical scanner is presented. The optimal PLDs were found to be 1300 msec for pCASL with and without vascular crushing. pCASL labeling outperformed PASL labeling in terms of convergence, anatomical correspondence between GM and perfusion maps, and SNR (P < 0.05). BSup appeared to have no additional value on the convergence, anatomical GM correspondence, and SNR (P > 0.05). EPI readout yielded a slightly better convergence, while the SNR of the GRASE readout was higher (P < 0.05). ASL on 1T is clinically feasible using state-of-the-art sequences that were primarily developed for higher field strengths