Myelin water imaging from accelerated 3D-GRASE acquisitions using subspace constrained reconstruction

Riwaj Byanju1, Stefan Klein1, Alexandra Cristobal-Huerta1

  • 1Department of Radiology and Nuclear Medicine, Erasmus MC, Rotterdam, The Netherlands.

PubMed
Abstract

Insights

This study explored subspace constrained reconstruction for faster brain MRI scans. While it improved myelin water fraction (MWF) and intra-/extra-cellular water (IET2) mapping, it did not surpass current methods for MWF but offered acceleration for IET2.

Area of Science:

  • Neuroimaging
  • Quantitative MRI
  • Biomarker Development

Background:

  • Quantitative MRI markers like myelin water fraction (MWF) and intra-/extra-cellular water T2 (IET2) are crucial for diagnosing brain disorders.
  • Current methods for acquiring these markers involve lengthy scan times due to multi-echo spin-echo sequences.
  • Accelerated imaging techniques, such as undersampled 3D-GRASE with parallel imaging, are needed to reduce scan duration.

Purpose of the Study:

  • To investigate the potential of subspace constrained reconstruction (SCR) for further accelerating 3D-GRAdient Echo and Spin Echo (3D-GRASE) MRI scans.
  • To assess the feasibility of using SCR to generate artifact-free myelin water fraction (MWF) and intra-/extra-cellular water T2 (IET2) maps at higher acceleration factors.
  • To evaluate the performance of SCR compared to state-of-the-art and reference multi-spin-echo techniques for quantitative MRI marker mapping.

Main Methods:

  • Two novel undersampling techniques for 3D-GRASE acquisition were developed, exploiting redundancy across echoes.
  • Retrospective undersampling of fully sampled phantom and in-vivo data was performed to test the proposed techniques.
  • Comparisons included MWF and IET2 mapping against a reference multi-spin-echo technique and prospective undersampled in-vivo acquisitions.

Main Results:

  • Retrospective analysis showed worse Root Mean Square Deviation (RMSD) for MWF with proposed techniques compared to state-of-the-art.
  • For IET2, retrospective RMSD was similar or slightly improved using the proposed techniques.
  • Prospective scans revealed that undersampling artifacts degraded MWF maps, but IET2 maps remained accurate within 10 ms of the reference.

Conclusions:

  • Exploiting echo redundancy with SCR does not yield additional acceleration benefits beyond the current state-of-the-art for myelin water fraction (MWF) mapping.
  • However, SCR enables acceleration beyond the state-of-the-art for intra-/extra-cellular water T2 (IET2) mapping.
  • These findings highlight the potential of SCR for specific quantitative MRI biomarkers, improving scan efficiency.