Reduction of Motion Artifacts in the Recovery of Undersampled DCE MR Images Using Data Binning and L+S Decomposition

Area of Science:

• Medical Imaging
• Magnetic Resonance Imaging (MRI)
• Image Reconstruction

Background:

• Motion artifacts, such as blurring and ghosting, are significant challenges in MRI, particularly in Dynamic Contrast Enhancement (DCE) MRI.
• Distinguishing motion artifacts from rapid contrast agent changes in DCE-MRI is complex.

Purpose of the Study:

• To develop and validate a novel technique for reducing motion artifacts in DCE-MRI.
• To improve the quality of DCE-MRI images acquired during free breathing.

Main Methods:

• A new technique combining data binning with a low rank plus sparse (L+S) reconstruction method was developed.
• Golden-angle radial sampling was used for continuous k-space data acquisition, enabling extraction of respiratory signals for motion state binning.
• Compressed sensing-based L+S matrix decomposition reconstructed motion-sorted DCE MR images.

Main Results:

• The proposed technique demonstrated visually sharper recovered images compared to conventional L+S decomposition.
• Quantitative analysis showed better structural similarity index in images reconstructed with the new method.
• The technique effectively reduced motion artifacts in free-breathing 3D liver and abdominal DCE-MRI datasets.

Conclusions:

• Data binning as a preprocessing step significantly enhances MR image quality in free-breathing DCE-MRI using L+S decomposition.
• Data binning effectively resolves respiratory motion by sorting data into distinct respiratory phases (bins).
• The method successfully differentiates respiratory motion from contrast agent variations, leading to superior image quality.