Non-Equispaced System Matrix Acquisition for Magnetic Particle Imaging Based on Lissajous Node Points.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Physics

Background:

• Magnetic Particle Imaging (MPI) is an emerging preclinical and clinical imaging technology.
• Current MPI reconstruction methods often use system matrices based on equispaced sampling grids.
• Lissajous trajectories, commonly used in MPI, result in inhomogeneous field-of-view coverage, with denser sampling at the edges than the center.

Purpose of the Study:

• To introduce a novel, non-equispaced sampling grid for Magnetic Particle Imaging reconstruction.
• To leverage the inherent inhomogeneity of Lissajous trajectories for more efficient MPI data acquisition.
• To improve the accuracy and efficiency of MPI image reconstruction.

Main Methods:

• Developed a sampling grid using node points from Lissajous trajectories to match the inhomogeneous coverage.
• Implemented a tailored polynomial interpolation for reconstructing the MPI signal from the non-equispaced data.
• Validated the approach using simulated and measured Magnetic Particle Imaging data.

Main Results:

• Demonstrated the first use of a trajectory-related non-equispaced grid for MPI image reconstruction.
• Showed that the number of sampling positions can be reduced significantly.
• Maintained constant spatial resolution despite the reduction in sampling points.

Conclusions:

• The proposed non-equispaced sampling grid, derived from Lissajous trajectories, offers an efficient alternative for MPI reconstruction.
• This method optimizes data acquisition by accounting for the inherent sampling inhomogeneity.
• The findings suggest potential for faster and more resource-efficient MPI systems.