Rapid whole-brain T₂* and susceptibility mapping using 3D multiple overlapping-echo detachment acquisition and...

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Medical Physics
• Neuroimaging

Background:

• Quantitative susceptibility mapping (QSM) and T2* relaxometry provide valuable insights into brain tissue properties.
• Conventional MRI techniques for whole-brain T2* and QSM mapping are often limited by long acquisition times and motion sensitivity.
• Developing rapid and motion-robust imaging methods is crucial for clinical translation and broader application.

Purpose of the Study:

• To develop and validate a novel 3D multiple overlapping-echo detachment (3D-MOLED) imaging technique.
• To establish data generation and reconstruction strategies for rapid whole-brain T2* and QSM.
• To evaluate the performance of 3D-MOLED against conventional 3D gradient-recalled echo (3D-GRE) methods.

Main Methods:

• Extended MOLED encoding to a 3D multi-shot acquisition combined with dual-echo blip-reversed EPI trains.
• Employed a deep learning-based missing modality synthesis for generating co-registered multi-parametric templates for Bloch simulations.
• Utilized a pseudo-3D Bloch simulation to accelerate synthetic data generation for network training and evaluated in healthy and clinical cohorts.

Main Results:

• 3D-MOLED demonstrated significant improvements in scan speed and motion robustness compared to 3D-GRE.
• Over 70% of scans in both healthy and clinical cohorts achieved good image quality.
• The developed deep learning and simulation framework enabled efficient generation of high-quality training data for accurate quantitative mapping.

Conclusions:

• 3D-MOLED enables simultaneous whole-brain T2* and QSM mapping at 1mm isotropic resolution within 50 seconds.
• The technique offers superior motion robustness compared to conventional 3D-GRE.
• 3D-MOLED represents a significant advancement for rapid and reliable quantitative MRI of the brain.