Fast and accurate T₂ mapping using Bloch simulations and low-rank plus sparse matrix decomposition.

Area of Science:

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

Background:

• T2 relaxation time mapping is crucial for MRI-based diagnosis and prognosis.
• Current T2 mapping methods require lengthy scan times, limiting clinical application.

Purpose of the Study:

• To develop and evaluate accelerated T2 mapping techniques using low-rank plus sparse (L+S) signal decomposition.
• To reconstruct accurate T2 maps from highly undersampled multi-echo spin-echo (MESE) MRI data.

Main Methods:

• Introduced two L+S algorithms: a standard iterative approach and SPArse and fixed RanK (SPARK).
• SPARK utilizes a pre-calculated fixed rank for the low-rank component, assuming it captures most MESE information.
• Algorithms were validated on in vivo brain and calf data with acceleration factors from 2x to 6x.

Main Results:

• Accelerated T2 maps reconstructed with L+S and SPARK demonstrated improved accuracy compared to fully sampled ground truth maps.
• Both methods outperformed standard GRAPPA acceleration in accuracy.

Conclusions:

• The SPARK technique provides accurate T2 maps with enhanced robustness to reconstruction parameter selection.
• SPARK's suitability for various applications facilitates the integration of quantitative T2 information into clinical practice.