On the feasibility of myelin water imaging at low-field MRI using multi-echo GRASE

Motivation: Low-field MRI may offer benefits for myelin water fraction (MWF) mapping due to the longer T2 of water and more homogeneous B0 and B1+ fields compared to higher field strengths. Goal(s): To investigate a multi-echo GRASE sequence for MWF mapping at 0.55T. Approach: Four healthy volunteers were scanned using a multi-echo GRASE sequence with GRAPPA x2 and 2x2 acceleration. Images were reconstructed using both conventional GRAPPA and a deep learning-based method. Resulting MWF maps were compared visually and using a Bland-Altman analysis. Results: Using GRAPPA 2x2 and deep-learning reconstruction, good quality MWF maps in agreement with GRAPPA x2 could be obtained in 6:41$$$\,$$$minutes. Impact: By leveraging deep learning-based reconstructions, myelin water imaging is feasible at 0.55T with a multi-component decomposition of multi-echo T2 data acquired in approximately 6 minutes at 2.5 mm isotropic resolution, encouraging further neuroimaging research at low field strengths.