Accelerated 3D myelin water imaging using joint spatio-temporal reconstruction.

Area of Science:

• Magnetic Resonance Imaging
• Neuroimaging
• Biomedical Engineering

Background:

• Myelin water imaging (MWI) is crucial for assessing white matter integrity.
• Traditional 3D multi-echo gradient echo (mGRE) acquisition for MWI is time-consuming.
• Accelerating MWI acquisition is essential for clinical feasibility and broader research applications.

Purpose of the Study:

• To develop and evaluate a method for accelerating 3D mGRE acquisition for MWI.
• To combine joint parallel imaging (JPI) and joint deep learning (JDL) for enhanced reconstruction.
• To achieve high acceleration factors without compromising image quality.

Main Methods:

• A multistep reconstruction process integrating JPI and JDL was implemented.
• JPI estimated missing k-space data, followed by JDL for artifact reduction.
• Variable splitting optimization with spatiotemporal denoiser, data consistency, and weighted average blocks was utilized.
• The method was evaluated using 2D Cartesian uniform undersampling for each echo.

Main Results:

• The combined JPI and JDL method demonstrated acceptable MWI quality.
• Quantitative values were improved compared to using JPI or JDL individually.
• Reconstructions showed high similarity to fully sampled MWI, evidenced by low normalized mean-square error and high-frequency error norm.

Conclusions:

• The joint spatiotemporal reconstruction approach effectively combines JPI and JDL.
• High acceleration factors are achievable for 3D mGRE-based MWI.
• This method offers a promising solution for faster and high-fidelity MWI acquisition.