Image-Space Gridding for Nonrigid Motion-Corrected MR Image Reconstruction

Area of Science:

• Medical Imaging
• Biophysics
• Computational Science

Background:

• Free-breathing cardiac MRI faces challenges due to patient motion during data acquisition.
• Respiratory motion causes image blurring and artifacts, particularly affecting coronary artery visualization.
• Accurate motion correction is crucial for diagnostic quality in cardiac MR imaging.

Purpose of the Study:

• To develop and validate a model-based approach for nonrigid motion correction in free-breathing cardiac MRI.
• To address challenges in motion representation and estimation for improved image reconstruction.
• To enhance the diagnostic utility of cardiac MR imaging by reducing motion-induced artifacts.

Main Methods:

• Utilized image-space gridding via the nonuniform fast Fourier transform (NUFFT) for nonrigid motion representation.
• Introduced nonrigid SENSE operators to integrate motion into the multi-coil MR acquisition model.
• Employed low-resolution 3D image-based navigators (iNAVs) and high-resolution 3D self-navigating iNAVs (self-iNAVs) for motion estimation.
• Reconstructed 3D self-iNAVs using data from multiple heartbeats within the same respiratory phase for nonrigid motion estimation.

Main Results:

• The proposed method successfully estimated nonrigid respiratory motion.
• Enhanced sharpness was observed in coronary arteries.
• Image quality in non-cardiac regions showed significant improvement.
• Performance surpassed that of translational motion-corrected reconstruction.

Conclusions:

• The developed model-based approach effectively corrects nonrigid motion in free-breathing cardiac MRI.
• The method offers a promising solution for improving image quality and diagnostic accuracy.
• Further validation in clinical settings is warranted.