Phantom-based field maps for gradient nonlinearity correction in diffusion imaging.

Area of Science:

• Medical Imaging
• Physics
• Biomedical Engineering

Background:

• Gradient coils in magnetic resonance imaging (MRI) generate nonlinear fields.
• These nonlinearities introduce systematic errors in diffusion imaging parameters, affecting reliability across different sites and scanners.
• Existing methods for correcting these errors may be impractical for many research sites.

Purpose of the Study:

• To implement and validate a simple empirical field-mapping procedure for correcting gradient field nonlinearities in diffusion MRI.
• To assess the accuracy and precision of this method in reproducing manufacturer-provided field maps.
• To quantify the reduction in spatially varying errors for diffusion parameters.

Main Methods:

• An empirical field-mapping procedure was developed using a large phantom and a solid harmonic approximation of coil fields.
• The accuracy was evaluated by comparing the generated field maps to manufacturer gold standards.
• The method's effectiveness in reducing errors in quantitative diffusion imaging was tested on an isotropic phantom.

Main Results:

• Median B value error was reduced from 33-41% to 0-4% at 100 mm from isocenter.
• On-axis spatial variation in estimated mean diffusivity decreased from 2.2%-4.1% to 0.5%-1.6% within 60 mm of isocenter.
• Estimated fractional anisotropy errors were substantially reduced, e.g., 72% reduction in the frequency encoding direction.

Conclusions:

• The developed empirical field-mapping procedure is accurate and precise for correcting gradient field nonlinearities.
• This simple method effectively reduces spatially varying errors in quantitative diffusion imaging, enhancing data reliability.
• The approach offers a practical solution for sites lacking specialized engineering support to improve diffusion MRI accuracy.