Accelerating MR Imaging Liver Steatosis Measurement Using Combined Compressed Sensing and Parallel Imaging: A...

Area of Science:

• Magnetic Resonance Imaging
• Medical Imaging Physics
• Quantitative MRI

Background:

• Liver fat fraction quantification is crucial for diagnosing and monitoring non-alcoholic fatty liver disease.
• Conventional MRI methods require long breath-hold periods, limiting patient comfort and increasing motion artifacts.
• Accelerated MRI techniques aim to reduce scan times without compromising diagnostic accuracy.

Purpose of the Study:

• To evaluate the accuracy and reliability of accelerated hepatic fat fraction and R2* relaxation rate measurements.
• To compare accelerated MRI reconstructions with conventional fully sampled acquisitions using Bland-Altman analysis.
• To assess the impact of acceleration factors on image quality and quantitative accuracy.

Main Methods:

• Prospective acquisition of undersampled data at various acceleration ratios (2.6x to 4.8x) alongside fully sampled data.
• Reconstruction of fat fraction maps using combined compressed sensing and parallel imaging (CS-PI).
• Bland-Altman analysis for limits of agreement, inter- and intrarater variability, and image quality assessment.

Main Results:

• Accelerated acquisitions showed high agreement with fully sampled data, with 95% limits of agreement for fat fractions ranging from 1.1% to 1.5%.
• Acceleration up to 3.8x did not significantly degrade fat fraction measurements compared to intra- and interrater variability.
• Image quality remained acceptable up to 3.8x acceleration, with minimal artifacts, while higher acceleration (4.8x) introduced substantial artifacts.

Conclusions:

• Prospective undersampling combined with CS-PI reconstruction effectively accelerates liver fat fraction measurements.
• The technique is reliable and accurate up to a 3.8x acceleration factor.
• This acceleration significantly reduces breath-hold duration from 17.7 to 4.7 seconds, improving clinical feasibility.