Accelerated noncontrast-enhanced pulmonary vein MRA with distributed compressed sensing

Area of Science:

• Medical Imaging
• Cardiovascular Imaging
• Magnetic Resonance Imaging

Background:

• Pulmonary vein magnetic resonance angiography (PV MRA) is crucial for cardiac imaging.
• Accelerating free-breathing, electrocardiogram (ECG)-triggered noncontrast PV MRA is essential for clinical efficiency.
• Compressed sensing (CS) offers potential for accelerating MRI acquisition.

Purpose of the Study:

• To evaluate the effectiveness of distributed compressed sensing (CS) in accelerating free-breathing, ECG-triggered noncontrast PV MRA.
• To compare distributed CS with coil-by-coil CS for PV MRA reconstruction.

Main Methods:

• Retrospective undersampling of fully sampled PV MRA data from seven healthy subjects (factors of 2-10).
• Reconstruction of undersampled data using distributed CS and coil-by-coil CS methods.
• Image quality assessment by blinded readers, artifact evaluation, PV diameter measurements, and mean square error (MSE) calculation.

Main Results:

• Distributed CS and coil-by-coil CS achieved image quality comparable to fully sampled data up to 6-fold and 4-fold undersampling, respectively.
• No significant differences in artifacts or PV diameters were observed between CS methods and reference scans.
• Both methods demonstrated comparable MSE at similar undersampling rates, with significant differences across all rates.

Conclusions:

• Distributed CS enables high undersampling factors (up to 6-fold) for noncontrast PV MRA.
• The sparsity of PV MRA data and joint coil image sparsity support accelerated imaging with distributed CS.
• Distributed CS is a promising technique for accelerating free-breathing PV MRA without compromising image quality.