Robust, fast and accurate mapping of diffusional mean kurtosis

Area of Science:

• Biomedical Imaging
• Diffusion MRI Physics
• Quantitative MRI

Background:

• Diffusional kurtosis imaging (DKI) measures non-Gaussian diffusion, vital for neurological disease assessment.
• Current DKI methods face challenges in robust, fast, and accurate kurtosis estimation from clinical data.
• Limitations exist in conventional DKI regarding maximum b-value and acquisition time.

Purpose of the Study:

• To develop an accurate, fast, and robust method for estimating mean kurtosis using a sub-diffusion framework.
• To overcome the b-value limitations of conventional DKI.
• To enable clinically feasible kurtosis mapping with reduced acquisition times.

Main Methods:

• Developed a novel kurtosis estimation approach based on the sub-diffusion mathematical framework.
• Proposed a fast and robust fitting procedure using two diffusion times for sub-diffusion model parameter estimation.
• Evaluated the sub-diffusion-based kurtosis mapping method using simulations and human brain Connectome 1.0 data.

Main Results:

• The sub-diffusion framework extends DKI, overcoming b-value limitations and allowing kurtosis/diffusivity computation from sub-diffusion model parameters.
• The new fitting procedure enables fast and robust parameter estimation without increasing acquisition time.
• Exquisite tissue contrast was achieved with diffusion data acquired in minutes.

Conclusions:

• The proposed sub-diffusion-based kurtosis mapping offers a robust, fast, and accurate method for mean kurtosis estimation.
• This approach is compatible with clinically feasible diffusion-weighted magnetic resonance imaging acquisition times.
• The findings suggest significant potential for improved clinical diagnosis and monitoring of neurological conditions.