Parallel spectroscopic imaging reconstruction with arbitrary trajectories using k-space sparse matrices.

Area of Science:

• Magnetic Resonance Imaging
• Spectroscopic Imaging
• Medical Physics
• Image Reconstruction

Background:

• Parallel imaging reconstruction accelerates magnetic resonance spectroscopic imaging (MRSI) by reducing scan times.
• Current methods like SENSE, SMASH, and GRAPPA are effective for Cartesian k-space data.
• Reconstructing undersampled MRSI data with arbitrary k-space trajectories using image-domain methods is computationally intensive.

Purpose of the Study:

• To introduce a new k-space domain-based parallel spectroscopic imaging reconstruction algorithm.
• To apply this algorithm to MRSI with arbitrary k-space trajectories, specifically spiral trajectories.
• To reduce memory requirements and computing times for MRSI reconstruction.

Main Methods:

• Developed a k-space domain-based parallel MRSI reconstruction algorithm utilizing k-space sparse matrices.
• Applied the algorithm to MRSI data acquired with spiral k-space trajectories.
• Evaluated performance using both phantom and in vivo studies at various undersampling factors.

Main Results:

• The novel algorithm successfully reconstructed undersampled MRSI data with arbitrary k-space trajectories.
• Achieved significant reductions in memory requirements and computing times compared to existing methods.
• Reconstructed spectroscopic images closely matched those obtained with fully sampled data.

Conclusions:

• The proposed k-space domain reconstruction method offers an efficient solution for parallel MRSI with arbitrary trajectories.
• This approach enables faster MRSI acquisition without compromising image quality.
• The technique holds promise for improving clinical MRSI applications by reducing scan duration and computational load.