Joint T1 and T2 Mapping With Tiny Dictionaries and Subspace-Constrained Reconstruction

Area of Science:

• Medical Imaging
• Biophysics
• Computational Science

Background:

• Accurate T1-T2 mapping in magnetic resonance imaging (MRI) is crucial for quantitative diagnostics.
• Traditional dictionary-based methods require large dictionaries, limiting computational efficiency.
• Existing methods struggle to balance dictionary size with representation accuracy.

Purpose of the Study:

• To develop a novel method for adaptively generating small dictionaries for joint T1-T2 mapping in MRI.
• To improve the efficiency of T1-T2 mapping without compromising accuracy.
• To reduce the dictionary size required for accurate parameter mapping.

Main Methods:

• Approximating the Bloch-response manifold using piece-wise linear functions.
• Adaptively refining the sampling grid based on locally-linear approximation error.
• Utilizing a 2D radially sampled Inversion-Recovery Hybrid-State Free Precession sequence for data acquisition.

Main Results:

• Adaptive dictionaries were generated with varying error tolerances and compared to heuristic dictionaries.
• Tiny dictionaries were successfully employed for T1-T2 mapping in phantom and in vivo studies.
• Reconstruction and parameter mapping were performed efficiently within a subspace.

Conclusions:

• The proposed adaptive dictionary method significantly reduces dictionary size (by one to two orders of magnitude) compared to heuristic approaches.
• Excellent agreement was observed between the adaptive mapping technique and template matching using heuristic dictionaries.
• This approach enables efficient and accurate T1-T2 mapping in MRI.