A Comparison of Tissue Property Values Estimated Using Conventional Cardiac MRF and MT-Cardiac MRF

Area of Science:

• Biomedical Imaging
• Quantitative MRI
• Cardiovascular Magnetic Resonance

Background:

• Cardiac Magnetic Resonance Fingerprinting (cMRF) enables simultaneous T1 and T2 quantification.
• Conventional cMRF reconstruction often neglects Magnetization Transfer (MT) effects.
• MT processes can influence the accuracy of T1 and T2 measurements in myocardial tissue.

Purpose of the Study:

• To evaluate the impact of incorporating MT modeling into cMRF reconstruction.
• To assess the accuracy of T1 and T2 quantification with and without MT modeling.
• To determine if MT-modeled cMRF can improve the detection of myocardial scar.

Main Methods:

• Developed an MT-cMRF dictionary modeling free and bound water pools and bound pool fraction (BPF).
• Validated MT-cMRF in simulations and phantoms for T1f, T2f, and BPF accuracy.
• Acquired cMRF data in healthy subjects and patients with suspected myocardial scar at 1.5T.
• Utilized Linear Discriminant Analysis (LDA) to differentiate scar from remote myocardium.

Main Results:

• MT-cMRF demonstrated high accuracy in simulations (RMS error 2.4%) and phantoms (R²=0.99).
• MT-cMRF yielded significantly different T1f and T2f values in healthy myocardium compared to conventional cMRF.
• MT-cMRF achieved higher LDA accuracy (82.4% for T1f/T2f, 86.1% for T2f/BPF) in scar discrimination than conventional cMRF (61.8%).

Conclusions:

• Incorporating MT modeling into cMRF significantly improves T1f and T2f quantification.
• MT-cMRF enables the measurement of BPF, providing additional tissue characterization.
• Native MT-cMRF parameters (T1f, T2f, BPF) effectively distinguish scar from healthy myocardium.