The impact of post-processing on spinal cord diffusion tensor imaging.

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Radiology

Background:

• Diffusion Tensor Imaging (DTI) is crucial for assessing brain and spinal cord microstructure.
• Spinal cord DTI quality lags behind brain DTI due to small size and artifacts like eddy currents and cardiac pulsation.
• Existing improvements often rely on cardiac gating and specialized acquisition, with retrospective correction methods less established.

Purpose of the Study:

• To develop and validate an improved post-processing pipeline specifically for spinal cord DTI data.
• To enhance the quality and reliability of spinal cord DTI metrics, particularly fractional anisotropy (FA) maps.
• To enable better differentiation of grey and white matter in the spinal cord.

Main Methods:

• Compared 3D-affine and slice-wise registration for eddy current and motion correction.
• Introduced a novel robust tensor fitting method to mitigate whole-volume outliers.
• Evaluated the combined pipeline's impact on image quality metrics.

Main Results:

• The combination of slice-wise motion correction, eddy current correction, and robust tensor fitting provided optimal results.
• Achieved approximately 30% increase in Contrast-to-Noise Ratio (CNR) for FA maps.
• Reduced intra-subject variation in FA maps by 18%.

Conclusions:

• The developed post-processing pipeline significantly enhances spinal cord DTI quality without increasing scan time.
• Improved FA maps facilitate clearer distinction between spinal cord grey and white matter.
• The pipeline is compatible with various multi-directional DTI acquisition schemes.