Evaluation of EPI-Based Distortion Correction Techniques for Cardiac Diffusion Tensor Imaging

Area of Science:

• Medical Imaging
• Biophysics
• Cardiovascular Research

Background:

• Cardiac diffusion tensor imaging (cDTI) is essential for assessing cardiac microstructure.
• Echo planar imaging (EPI) readouts used in cDTI are susceptible to significant image distortions.
• Existing distortion correction techniques, such as FSL TOPUP and TORTOISE DR-BUDDI, are primarily used in neurological applications.

Purpose of the Study:

• To evaluate and compare the suitability of FSL TOPUP and TORTOISE DR-BUDDI for correcting EPI distortions in cDTI.
• To determine which distortion correction technique offers the best performance for cDTI applications.

Main Methods:

• cDTI data were acquired from 10 healthy volunteers using blip-up (BU) and blip-down (BD) EPI readouts.
• Datasets were processed using both TOPUP and DR-BUDDI distortion correction software.
• Image quality was assessed using geometric fidelity metrics (Dice Similarity Coefficient, Hausdorff Distance) and quantitative parametric maps (mean diffusivity, fractional anisotropy, helix angle pitch), along with uncertainty measurements.

Main Results:

• Both TOPUP and DR-BUDDI significantly improved geometric fidelity (DSC and AHDepi) compared to uncorrected BU and BD data.
• Distortion correction led to decreased fractional anisotropy (FA) and slightly altered mean diffusivity (MD).
• Uncertainty measurements (dMD, dFA, de1) increased following distortion correction, with TOPUP showing slightly better results than DR-BUDDI.

Conclusions:

• EPI distortion correction techniques enhance geometric accuracy in cDTI but decrease the precision of uncertainty measurements.
• TOPUP demonstrates superior performance compared to DR-BUDDI for cDTI applications, offering a better balance between geometric correction and data precision.
• Further optimization of distortion correction methods is warranted for robust cDTI analysis.