Magnetic susceptibility source separation (χ-separation) in quantitative susceptibility mapping

Area of Science:

• Biomedical Imaging
• Neuroimaging
• Magnetic Resonance Imaging (MRI)

Background:

• Conventional quantitative susceptibility mapping (QSM) provides voxel-averaged magnetic susceptibility, failing to resolve co-localized paramagnetic iron and diamagnetic myelin due to phase cancellation.
• This limitation obscures the accurate quantification of crucial biomarkers within tissues.

Purpose of the Study:

• To review and categorize methodologies for chi-separation (χ-separation), a technique designed to disentangle sub-voxel magnetic sources.
• To evaluate the potential of these methods for improving the specificity of iron and myelin quantification in neuroimaging.

Main Methods:

• Categorization of χ-separation techniques into multi-sequence approaches (using R2 or R2' relaxation parameters) and gradient-recalled echo (GRE)-only methods.
• Evaluation of deep learning frameworks (e.g., χ-sepnet) for synthesizing missing relaxation parameters and enabling direct source mapping.
• Analysis of acquisition protocol simplification trends and biophysical challenges, including relaxometric constant determination and static dephasing assumptions.

Main Results:

• χ-separation significantly improves specificity for quantifying iron and myelin compared to conventional QSM.
• A trend towards simplified acquisition protocols is observed, but challenges in determining accurate relaxometric constants persist.
• The absence of a definitive in vivo gold standard hinders absolute accuracy assessment of current χ-separation methods.

Conclusions:

• χ-separation offers a promising solution to overcome phase cancellation limitations in QSM for improved biomarker quantification.
• Future research must focus on anisotropy-aware modeling and robust ground-truth validation to ensure the clinical viability of χ-separation techniques.