Highly accelerated vessel-selective arterial spin labeling angiography using sparsity and smoothness constraints

Area of Science:

• Medical Imaging
• Radiology
• Biomedical Engineering

Background:

• Dynamic arterial spin labeling (ASL) angiography is crucial for visualizing blood vessels.
• Current ASL angiography methods may face limitations in scan time or image quality when aiming for vessel selectivity.

Purpose of the Study:

• To demonstrate that vessel-selective dynamic ASL angiography can be achieved without increasing scan time or compromising image quality compared to conventional ASL angiography.
• To validate the use of compressed-sensing reconstruction for improving ASL angiography.

Main Methods:

• Simulations on numerical phantoms and in vivo data acquisition from healthy volunteers were performed.
• A compressed-sensing framework with enforced image domain sparsity and temporal smoothness was utilized.
• Data were acquired using a 2D golden-angle radial trajectory and reconstructed at acceleration factors up to R=34.

Main Results:

• Relative sparsity was identified as a key factor for reconstruction fidelity in vessel-encoded angiograms.
• Image quality differences between vessel-selective and nonselective angiography were negligible at high acceleration factors (R=34).
• Reconstruction quality was robust and not highly sensitive to parameter tuning.

Conclusions:

• Increased relative sparsity in vessel-selective angiograms allows for higher acceleration factors.
• This approach enables the acquisition of vessel-selective information without any scan-time penalty.
• The findings support the use of this technique for efficient and high-quality ASL angiography.