Evaluation of non-Gaussian diffusion in cardiac MRI

Area of Science:

• Cardiovascular MRI
• Diffusion MRI Physics

Background:

• Cardiac diffusion MRI commonly uses the Gaussian diffusion tensor model.
• Biological tissue diffusion is often non-Gaussian due to microstructural constraints.
• Non-Gaussian models may offer improved quantification of diffusion in cardiac tissue.

Purpose of the Study:

• Investigate non-Gaussian diffusion characteristics in healthy and hypertrophic rat hearts.
• Compare the performance of Gaussian and non-Gaussian diffusion models in detecting cardiac hypertrophy.

Main Methods:

• Ex vivo imaging of 13 rat hearts (healthy, sham, hypertrophic) using diffusion-weighted MRI.
• Acquisition of images at high b-values (up to 10,000 s/mm²).
• Model fitting and selection using the Akaike information criterion.

Main Results:

• The Gaussian diffusion tensor model performed best at low b-values (<2000 s/mm²).
• A non-Gaussian beta distribution model provided a better fit at high b-values.
• Hypertrophic hearts exhibited >20% higher diffusion kurtosis and skewness in specific directions compared to healthy hearts.

Conclusions:

• Non-Gaussian diffusion models demonstrate superior sensitivity for detecting cardiac hypertrophy.
• Diffusion kurtosis is a potential biomarker for characterizing cardiac disease and remodeling.
• Findings suggest advanced diffusion modeling can enhance cardiac MRI analysis.