Susceptibility Source Separation for Quantification of Remyelination in Multiple Sclerosis Lesions from multi Gradient-Echo (mGRE) Data

Motivation: Susceptibility source separation (SSS) was proposed to estimate myelin and iron content of the tissues from gradient-echo data. However, its quantification accuracy is limited by the model assumptions. Goal(s): This study aims to improve SSS by accounting for changes in water content. Approach: Twelve MS patients with 40 chronic lesions scanned six months and five years after time of enhancement. SSS with and without water modeling were compared for their ability to detect changes in myelin. Results: Demyelination of MS lesions was only detected in SSS when modeling water content. Impact: Understanding about the mechanism and timing of the remyelination in multiple sclerosis (MS) is incomplete. Therefore, noninvasive quantification of myelin is fast and accurately is an important clinical need. Susceptibility source separation from gradient-echo carries potential to satisfy this need.