Three-dimensional through-time radial GRAPPA for renal MR angiography

Area of Science:

• Magnetic Resonance Imaging
• Medical Imaging Techniques
• Radiology

Background:

• Fast imaging techniques are crucial for high temporal and spatial resolution in contrast-enhanced time-resolved MR angiography (trMRA).
• Non-Cartesian parallel imaging methods are essential for accelerating these demanding MR exams.
• Generalized Autocalibrating Partially Parallel Acquisition (GRAPPA) is a key technique for parallel imaging reconstruction.

Purpose of the Study:

• To evaluate the 3D through-time radial GRAPPA method for reconstructing highly accelerated stack-of-stars data in time-resolved renal MR angiography.
• To explore different GRAPPA weight calibration schemes for optimizing acquisition time and reconstruction accuracy.
• To assess the feasibility of achieving high acceleration factors for dynamic renal imaging.

Main Methods:

• The study employed a 3D through-time radial GRAPPA approach for non-Cartesian GRAPPA weight calibration.
• Different GRAPPA weight calibration schemes were investigated through simulations.
• The optimized calibration schemes were applied to reconstruct undersampled stack-of-stars data for renal trMRA.

Main Results:

• Simulations identified an optimal 3D calibration strategy combining central partitions with extensive temporal repetitions.
• Renal trMRA data were successfully reconstructed with an in-plane acceleration factor up to 12.6.
• The best reconstruction, achieving a 16.4% root mean squared error, used 8 partitions and 16 repetitions.

Conclusions:

• 3D through-time radial GRAPPA effectively reconstructs highly accelerated non-Cartesian MR data.
• In-plane radial undersampling allows trMRA acquisition with <4s/frame temporal footprint.
• This method achieves high spatial resolution (approx. 1.5 mm x 1.5 mm x 3 mm) for dynamic renal angiography.