Free-running time-resolved first-pass myocardial perfusion using a multi-scale dynamics decomposition: CMR-MOTUS

Area of Science:

• Cardiovascular Magnetic Resonance Imaging
• Medical Image Reconstruction
• Biomedical Engineering

Background:

• First-pass myocardial perfusion imaging is crucial for diagnosing heart conditions.
• Accurate quantification requires high spatiotemporal resolution, challenging due to cardiac and respiratory motion.
• Existing methods struggle to precisely capture contrast agent inflow dynamics.

Purpose of the Study:

• To introduce CMR-MOTUS, a novel framework for reconstructing time-resolved, free-running first-pass myocardial perfusion.
• To jointly estimate high-quality motion fields and contrast-varying images for improved accuracy.
• To overcome limitations of static reference images in perfusion imaging.

Main Methods:

• CMR-MOTUS extends the MR-MOTUS framework using low-rank plus sparse decomposition.
• It integrates a contrast-varying reference image to capture blood flow and contrast inflow.
• The joint reconstruction alternates between solving for contrast changes and motion fields, eliminating the need for a pre-acquisition static reference.

Main Results:

• Simulations demonstrated superior image similarity and motion field accuracy compared to existing methods.
• In vivo tests showed effective capture of cardiac and respiratory motion dynamics.
• Resulting cine images exhibited sharper features than those from state-of-the-art techniques.

Conclusions:

• CMR-MOTUS significantly advances first-pass myocardial perfusion reconstruction by modeling motion and contrast dynamics.
• The framework supports a data-efficient, free-running workflow with high-quality motion field correlation.
• This approach holds potential for enhancing cardiac MRI diagnostic value, pending further clinical validation.