Experimental considerations for fast kurtosis imaging.

Area of Science:

• Medical Imaging
• Neuroscience
• Biophysics

Background:

• Clinical application of kurtosis imaging is limited by lengthy acquisition times and complex postprocessing.
• Recent advancements allow estimation of mean kurtosis tensor (W¯) and mean diffusivity (D¯) using the 1-3-9 protocol with 13 diffusion-weighted MRI acquisitions.
• Challenges remain in addressing noise and nonideal diffusion encoding in these protocols.

Purpose of the Study:

• To analyze the impact of noise and nonideal diffusion encoding on kurtosis imaging.
• To propose a novel correction strategy for improved parameter estimation.
• To introduce an enhanced 1-9-9 protocol offering increased robustness and minimal additional scan time.

Main Methods:

• Acquisition of 1-3-9 and 1-9-9 diffusion MRI data in rat and human brains.
• Comparison of D¯, fractional anisotropy (FA), and W¯ estimates with traditional methods using extensive diffusion kurtosis imaging datasets.
• Utilizing simulations to assess the influence of noise and deviations in diffusion encoding, and to evaluate the correction strategy's performance.
• Determining optimal b-values through simulations and experimental data analysis.

Main Results:

• Accuracy and precision of D¯ and W¯ estimations are comparable to nonlinear least squares methods.
• The 1-9-9 protocol demonstrates improved accuracy and precision.
• The proposed compensation strategy significantly enhances parameter estimation from nonideal datasets.
• The 1-9-9 protocol provides a fast estimation of FA without increasing computation time.

Conclusions:

• The developed framework provides a robust and concise method for estimating multiple diffusion metrics.
• The 1-9-9 protocol is easily implementable in clinical settings.
• This advancement facilitates more efficient and reliable diffusion kurtosis imaging.