Advances in Quantitative Magnetic Resonance Imaging of Myelin

Myelin water imaging (MWI) is a quantitative magnetic resonance imaging (MRI) technique generally regarded as the most rigorous approach for non-invasive, in-vivo measurement of myelin content.

Although MWI has proven valuable for the study of development, aging, disease, injury, genetics, and fundamental biology in the central nervous system, the power of its insights hinge on accurate characterization of normative values. To that end, we used MWI data from 100 adults (age 20-78) to create an optimized, unbiased myelin atlas and characterize how myelin content changes throughout the adult life span; an invaluable, openly available reference for future studies.

In practice, lengthy acquisition times have limited the utility of MWI and often lead to alternative approaches being used to acquire surrogates for MWI. To compare the traditional multi-echo T2 relaxation and alternative steady-state MWI approaches, we created multivariate brain and spinal cord atlases and found an approximately linear relationship between myelin estimates, which broke down in the presence of unique relaxation times (spinal cord, tissue affected by disease pathology). This work will improve retrospective interpretation, and guide future design, of MWI studies.

Next, we addressed lengthy MWI acquisition times using conventional compressed sensing before ultimately introducing the Constrained, Adaptive, Low-dimensional, Intrinsically Precise Reconstruction (CALIPR) framework. Drastically improved reconstruction performance allowed whole-brain MWI to be acquired using a previously unattainable sequence (fully sampled acquisition time 2h:57m:20s) in only 7m:26s with CALIPR (acceleration factor 23.9, 4.2% of the dataset). Reproducibility experiments demonstrated excellent precision, and CALIPR provided markedly increased sensitivity to demyelinating disease pathology (the hallmark application for myelin imaging). We implemented CALIPR for MWI of brain and spinal cord, and for two of the three largest MRI manufacturers. The CALIPR framework provides increased acceleration, precision, and sensitivity for MWI, and could be similarly transformative for other quantitative MRI applications.

Finally, we implemented CALIPR on an ultra-low field (0.064T) portable, point-of-care MRI scanner to acquire accurate, quantitative T2 mapping data in <10 minutes. In combination with the accessible, low-cost imaging enabled by this platform, this work could help revolutionize the care of neurological disorders by enabling frequent, quantitative assessment of subtle tissue changes.