Accelerated MRI with CIRcular Cartesian UnderSampling (CIRCUS): a variable density Cartesian sampling strategy for...

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Medical Imaging
• Signal Processing

Background:

• Accelerated MRI acquisition using compressed sensing (CS) and parallel imaging (PI) is crucial for efficient data acquisition.
• Image quality in accelerated MRI is highly dependent on the undersampling scheme design.
• Conventional undersampling methods include pseudo-random, radial, and spiral trajectories, each with limitations.

Purpose of the Study:

• To propose and evaluate a novel undersampling pattern for 3D Cartesian MRI acquisition.
• To enhance image reconstruction accuracy in CS and PI by optimizing undersampling strategies.
• To develop an undersampling method that combines the benefits of random and radial/spiral sampling.

Main Methods:

• A novel undersampling pattern was designed for the ky-kz plane in 3D Cartesian acquisition.
• The proposed method circularly selects sampling points, integrating features of random and radial/spiral sampling.
• Retrospective and prospective undersampling were used to evaluate the proposed scheme.

Main Results:

• The proposed undersampling scheme outperformed conventional random and radial/spiral sampling methods in 3D Cartesian acquisition.
• Performance was comparable to advanced variable-density Poisson-Disk sampling (vPDS).
• The scheme demonstrated interleaving flexibility for dynamic imaging and favorable preliminary results in prospective implementation.

Conclusions:

• The novel undersampling patterns offer desirable randomization and radial/spiral trajectory properties for 3D Cartesian MRI.
• The method provides easy implementation and flexible sampling capabilities.
• High accuracy in image reconstruction with CS and PI is achieved.