Improved diffusion parameter estimation by incorporating T₂ relaxation properties into the DKI-FWE model.

Area of Science:

• Medical Imaging
• Biophysics
• Neuroimaging

Background:

• Diffusion kurtosis imaging (DKI) and its free water elimination (FWE) variant provide tissue-specific diffusion information.
• Ill-conditioned parameter estimation in DKI-FWE necessitates advanced techniques, potentially introducing bias.
• Free water partial volume effects can compromise diffusion metric accuracy.

Purpose of the Study:

• Introduce the T2-DKI-FWE model to improve parameter estimation conditioning in diffusion imaging.
• Exploit T2 relaxation properties of tissue and free water compartments.
• Provide a new potential biomarker: the T2 relaxation time of tissue.

Main Methods:

• Developed the T2-DKI-FWE model incorporating T2 relaxation properties.
• Estimated T2 of tissue as an unknown parameter, fixing T2 of free water to a literature value.
• Analyzed error propagation from erroneous T2 of free water assumptions.
• Utilized Cramér-Rao lower bound for conditioning assessment.
• Validated on simulated and real diffusion MRI datasets.

Main Results:

• The T2-DKI-FWE model demonstrates improved conditioning compared to DKI-FWE.
• Errors from inaccurate T2 of free water assumptions were found to be minimal across various diffusion metrics.
• The model enhances the identifiability of diffusion parameters.
• Precise and accurate diffusion parameter estimates were achieved, reducing free water partial volume effects.

Conclusions:

• Incorporating T2 relaxation properties into DKI-FWE enhances model fitting conditioning.
• The T2-DKI-FWE model enables accurate diffusion parameter estimation using standard nonlinear least squares.
• This approach requires only a minimal acquisition scheme with at least two echo times.