A parallel spatial and Bloch manifold regularized iterative reconstruction method for MR Fingerprinting.

Area of Science:

• Medical Imaging
• Biophysics

Background:

• Magnetic Resonance Fingerprinting (MRF) reconstructs tissue maps from undersampled images.
• Current MRF methods utilize MR physics (Bloch equations) and regularization but rarely combine multiple types.
• Existing methods struggle with noise, especially with short scan sequences.

Purpose of the Study:

• To develop and evaluate a novel MRF reconstruction framework.
• To integrate Bloch manifold, spatial, and low-rank regularizations into a joint optimization algorithm.
• To improve the robustness and accuracy of MRF tissue map reconstruction.

Main Methods:

• Proposed a new family of MRF reconstruction methods.
• Incorporated Bloch manifold regularization with spatial and low-rank priors.
• Developed a joint optimization algorithm to effectively combine these regularizations.

Main Results:

• Demonstrated significant improvements in estimated tissue map quality.
• Validation performed on digital phantoms, NIST phantoms, and volunteer scans.
• The proposed methods show enhanced robustness to noise, particularly for shorter sequences.

Conclusions:

• Jointly applying Bloch manifold, spatial, and low-rank regularizations enhances MRF reconstruction.
• The novel methods offer superior performance compared to existing techniques.
• This approach advances the accuracy and reliability of MRF for tissue characterization.