SC-GROG followed by L+S reconstruction with multiple sparsity constraints for accelerated Golden-angle-radial DCE-MRI

Area of Science:

• Medical Imaging
• Magnetic Resonance Imaging
• Image Reconstruction

Background:

• Golden-angle-radial Sparse Parallel MRI (GRASP) accelerates free-breathing DCE-MRI.
• GRASP limitations include temporal averaging and long reconstruction times.
• Existing GRASP methods struggle with dynamic contrast and computational burden.

Purpose of the Study:

• To develop an improved reconstruction technique for free-breathing DCE-MRI.
• To address GRASP's limitations in temporal resolution and dynamic contrast.
• To reduce the computational burden of DCE-MRI reconstruction.

Main Methods:

• Proposed a novel method combining SC-GROG gridding with low-rank plus sparse (L+S) reconstruction.
• Utilized multiple sparsity constraints and Monotone FISTA with variable acceleration (MFISTA-VA) for optimization.
• Tested on 3T free-breathing in-vivo DCE-MRI datasets.

Main Results:

• The proposed method significantly improved reconstruction time and dynamic contrast over GRASP and XD-GRASP.
• MFISTA-VA integration resulted in faster convergence compared to conventional L+S reconstruction.
• Evaluated using convergence error, arterial signal intensity, and reconstruction time.

Conclusions:

• The novel SC-GROG gridding and L+S reconstruction method enhances free-breathing DCE-MRI.
• This approach offers superior temporal resolution and dynamic contrast with reduced reconstruction time.
• The MFISTA-VA optimization further improves reconstruction efficiency.