Stable and efficient retrospective 4D-MRI using non-uniformly distributed quasi-random numbers

Area of Science:

• Medical Imaging
• Magnetic Resonance Imaging
• Biophysics

Background:

• Acquiring 4D abdominal MRI during free breathing is challenging due to motion artifacts.
• Conventional 2D and 3D MRI techniques often require long scan times or compromise image quality.

Purpose of the Study:

• To develop a robust and reliable 3D Cartesian imaging technique for fast and flexible retrospective 4D abdominal MRI during free breathing.
• To improve image quality and temporal resolution for visualizing organ and tumor motion.

Main Methods:

• Incorporation of non-uniform quasi-random (NU-QR) reordering of phase encoding lines into 3D Cartesian acquisition.
• Utilizing respiratory self-gating and SPIRiT-reconstruction for 4D-MRI data reconstruction.
• Quantitative evaluation of image quality (NMI) versus scan times, comparing NU-QR with conventional sampling schemes.

Main Results:

• NU-QR sampling demonstrated more efficient undersampling patterns for parallel imaging compared to linear and uniform quasi-random sampling.
• NU-QR sampling achieved the highest median NMI values across all scan times, indicating superior image quality.
• The method enabled reconstruction of motion-artifact-reduced 4D datasets with isotropic 2.1 mm resolution in 3 minutes, capturing tumor displacements up to 46 mm.

Conclusions:

• NU-QR sampling in 3D Cartesian MRI significantly enhances image quality and efficiency for 4D abdominal imaging during free breathing.
• This technique allows for stable 4D-MRI with high temporal and spatial resolution, crucial for visualizing organ or tumor motion.
• Further research is needed to explore clinical applicability, particularly for radiotherapy planning.