Accelerated free-breathing liver fat and R 2 * quantification using multi-echo stack-of-radial MRI with...

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Medical Imaging Analysis
• Liver Disease Quantification

Background:

• Liver fat quantification using MRI is crucial for diagnosing and monitoring nonalcoholic fatty liver disease (NAFLD).
• Traditional MRI methods often require breath-holding, which can be challenging for patients and lead to motion artifacts.
• Accelerated MRI techniques are needed to improve efficiency and reduce patient burden.

Purpose of the Study:

• To enhance image quality and minimize quantification biases for free-breathing liver proton density fat fraction (PDFF) and T2* mapping.
• To accelerate liver PDFF and T2* quantification using radial k-space undersampling.
• To validate a novel free-breathing multi-echo stack-of-radial MRI method with compressed sensing.

Main Methods:

• Developed a free-breathing multi-echo stack-of-radial MRI sequence incorporating compressed sensing and multidimensional regularization.
• Validated the method in motion phantoms and 11 subjects (6 with NAFLD) against reference breath-hold Cartesian acquisitions.
• Reconstructed images, PDFF, and T2* maps using varying radial view undersampling factors and compared results using Bland-Altman analysis.

Main Results:

• The proposed method achieved PDFF quantification biases of [-1.0%; -5.8%, 3.8%] and T2* biases of [-0.5; -33.6, 32.7] s⁻¹ at an optimal sampling factor of 0.25.
• Demonstrated significantly lower coefficient of variation (p < 0.001), indicating superior image quality compared to other methods.
• Reduced effective acquisition time to 59-64 seconds from 153-171 seconds for baseline protocols.

Conclusions:

• The developed method enables accelerated free-breathing liver PDFF and T2* mapping.
• This technique offers reduced biases and variations in quantification.
• Potential for improved diagnostic accuracy and patient comfort in liver disease assessment.