Myelin water fraction mapping with joint inversion of gradient-echo and spin-echo data

Accurate estimation of brain myelin-water content from multi-echo data is challenging due to the inherent ill-posedness of the inversion problem. In this study, we propose a novel method for myelin-water imaging that jointly utilizes gradient-echo and spin-echo imaging data to enhance the accuracy of myelin-water estimation.

Multi-echo gradient-echo and spin-echo data were simulated and acquired in vivo. The simulations are based on a parameterized myelin and free water signal model, which is also used for the inversion by means of nonlinear local-search optimization. Single inversions of the individual datasets as well as joint inversion of the combined datasets were performed on simulated and real data. While single inversions estimate either the $$ or $$ relaxation spectrum, the joint inversion estimates both spectra simultaneously.

Simulation results show that the accuracy of myelin-water imaging improves when jointly inverting gradient-echo and spin-echo synthetic data. In vivo experiments show that the joint inversion of both datasets leads to sharper and more distinct myelin-water images as compared to the individual inversions.

Our method addresses the ill-posed nature of the myelin-water inversion problem by leveraging complementary information from multi-echo gradient-echo and multi-echo spin-echo imaging sequences, thus improving the reliability of myelin-water quantification.