Enhancing cardiac MRI reliability at 3 T using motion-adaptive B₀ shimming.

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Medical Physics
• Cardiovascular Imaging

Background:

• High-field MRI (≥3T) faces challenges in cardiac imaging due to B0-field inhomogeneities at the heart-lung interface.
• Conventional hardware-based shimming methods often fail to account for dynamic thoracic motion, resulting in off-resonance artifacts.
• Accurate B0-field mapping is crucial for mitigating these artifacts and improving image quality in cardiac MRI.

Purpose of the Study:

• To characterize B0-field fluctuations in the heart induced by cardiac and respiratory motion at 3T.
• To develop and evaluate a deep learning-based motion-adaptive B0 shimming pipeline.
• To assess the effectiveness of the proposed shimming approach in improving B0 homogeneity and T2* mapping accuracy.

Main Methods:

• Implemented a motion-resolved B0 mapping sequence at 3T to quantify motion-induced B0 variations.
• Developed a motion-adaptive shimming framework validated via numerical simulations and human imaging studies.
• Assessed B0-field homogeneity and T2* mapping accuracy using standard and motion-adaptive shimming across different breath-hold states.

Main Results:

• Respiratory motion significantly impacted myocardial B0 fields (p < 0.01), while cardiac motion had a minimal effect (p = 0.49).
• Motion-adaptive B0 shimming significantly improved field uniformity in both inspiratory and expiratory states compared to conventional shimming (p < 0.05).
• Observed corresponding improvements in myocardial T2* map homogeneity, with a reduced coefficient of variation (p < 0.01).

Conclusions:

• The developed motion-adaptive B0 shimming effectively compensates for respiration-induced B0 fluctuations.
• This approach enhances B0 field homogeneity and reduces off-resonance artifacts in high-field cardiac MRI.
• The strategy improves the robustness and reproducibility of T2* mapping, enabling more reliable cardiac MRI.