Correction for fast pseudo-diffusive fluid motion contaminations in diffusion tensor imaging.

Area of Science:

• Neuroimaging
• Biophysics
• Medical Physics

Background:

• Diffusion tensor imaging (DTI) measures water diffusion in brain tissue to infer white matter structure.
• Signal contamination from blood perfusion and cerebrospinal fluid (CSF) movement can affect DTI accuracy.
• Intravoxel incoherent motion (IVIM) modeling offers a potential method to differentiate diffusion from perfusion/CSF flow.

Purpose of the Study:

• To quantify the impact of fast pseudo-diffusion (perfusion/CSF) on DTI metrics in healthy brain tissue.
• To evaluate the effectiveness of an IVIM-based model in correcting for these contaminations.
• To compare DTI metrics derived from different signal modeling approaches.

Main Methods:

• Prospective study involving five healthy volunteers at 3 Tesla.
• Diffusion-weighted imaging acquired with multiple b-values and diffusion directions.
• DTI metrics (FA, MD) computed using two b-value (2b), monoexponential (mono), and IVIM (bi) models.
• Region of interest analysis in white matter, gray matter, and ventricles.

Main Results:

• Significant differences in mean diffusivity (MD) and fractional anisotropy (FA) were observed across all evaluated regions based on the modeling method.
• The IVIM model yielded lower MD values compared to 2b and monoexponential models.
• The nucleus caudatus and lateral ventricles showed the largest MD deviations, while the nucleus caudatus and putamen showed significant FA deviations.

Conclusions:

• Fast pseudo-diffusion significantly affects DTI metric quantification in the brain.
• The proposed IVIM-based method effectively corrects for signal contaminations from CSF or perfusion.
• Accurate DTI measurements require accounting for intravoxel incoherent motion, especially in regions with high CSF or perfusion influence.