User-guided compressed sensing for magnetic resonance angiography.

Area of Science:

• Medical Imaging
• Biophysics
• Signal Processing

Background:

• Compressed sensing (CS) magnetic resonance imaging (MRI) accelerates image acquisition by reconstructing images from undersampled k-space data.
• A key requirement for CS-MRI is the sparse representation of the target image in a specific transform domain.
• MR angiograms exhibit inherent sparsity and can be further sparsified using finite-difference transforms, making them suitable for CS-MRI applications.

Purpose of the Study:

• To address the issue of low-contrast vessels disappearing in compressed sensing MRI at high undersampling ratios.
• To propose and evaluate a user-guided CS-MRI method that enhances the visibility of low-contrast vessels within a region of interest (ROI).

Main Methods:

• Development of a user-guided CS-MRI technique to specifically preserve contrast in selected regions.
• Utilizing the inherent sparsity of MR angiograms, potentially enhanced by finite-difference transforms.
• Comparison of the proposed method against conventional CS-MRI reconstruction techniques using simulations.

Main Results:

• The proposed user-guided CS-MRI method effectively mitigates the underestimation of sparse coefficients by L1 minimization.
• Low-contrast vessels, crucial for clinical diagnosis, are well-maintained and visible at high undersampling ratios.
• Simulations demonstrate superior local image quality in the region of interest compared to conventional CS-MRI.

Conclusions:

• User-guided CS-MRI offers a significant improvement for visualizing low-contrast vessels in MR angiography.
• This technique enhances diagnostic confidence by preventing the loss of clinically relevant vascular structures.
• The method provides a valuable tool for optimizing CS-MRI applications where subtle vascular details are critical.