Quasi-Random Single-Point Imaging Using Low-Discrepancy $k$ -Space Sampling.

Area of Science:

• Medical Imaging
• Magnetic Resonance Imaging (MRI)
• Image Reconstruction

Background:

• Ultra-short echo-time (USE) sequences are crucial for imaging short relaxation time spin systems in MRI.
• Existing USE sequences face challenges with image blurring and artifacts from chemical shifts and magnetic susceptibility.
• There is a clinical need for faster, artifact-free MRI techniques.

Purpose of the Study:

• To present a novel concept for spherical quasi-random single-point imaging.
• To demonstrate the method's capability for high acceleration and metal artifact suppression.
• To enable faster MRI acquisition times for volumetric imaging.

Main Methods:

• Utilized quasi-random sampling in k-space based on a low-discrepancy sequence.
• Combined undersampling with non-linear optimization reconstruction techniques, specifically compressed sensing (CS).
• Employed spherical trajectory for data acquisition.

Main Results:

• The low-discrepancy trajectory exhibited ideal noise-like undersampling properties for CS reconstruction.
• Achieved denoised images with excellent reduction in metal artifacts.
• Demonstrated eightfold undersampling, enabling volume acquisition in minutes.

Conclusions:

• Spherical quasi-random single-point imaging offers a highly accelerateable MRI approach.
• The method effectively suppresses artifacts and reduces blurring.
• This technique holds promise for significantly reducing MRI scan times in clinical applications.