Nonlinear Gradient Field Estimation in Diffusion MRI Tensor Simulation.

Area of Science:

• Medical Imaging
• Biophysics
• Applied Mathematics

Background:

• Gradient nonlinearities in MRI cause spatial distortion and affect diffusion sensitization in diffusion-weighted MRI.
• Accurate correction of gradient nonlinearities is crucial due to advances in scanner performance.
• Current estimation methods often rely on phantom calibration, which is not always feasible, especially for retrospective data.

Purpose of the Study:

• To derive and validate a novel mathematical approach for estimating the complete gradient nonlinear field (L(r)) in diffusion MRI.
• To assess the accuracy of the estimated field and its impact on diffusion tensor metrics.
• To provide a method for gradient field estimation independent of calibration scans.

Main Methods:

• Developed a quadratic minimization problem to estimate the gradient nonlinear field L(r) from corrupted diffusion MRI signals.
• Evaluated the method in two scenarios: with the true diffusion tensor known and with the diffusion tensor being estimated.
• Assessed the estimated field using diffusion tensor metrics: mean diffusivity (MD), fractional anisotropy (FA), and principal eigenvector (V1).

Main Results:

• The mathematical model was shown to be stable and not ill-posed through tensor simulations.
• When the true diffusion tensor was known, the estimated L(r) field closely matched the true field (determinant change near zero).
• The estimated L(r) corrected diffusion metrics showed near-zero median differences compared to true values, with results dependent on the level of nonlinearity corruption.

Conclusions:

• A novel mathematical approach effectively estimates gradient nonlinear fields in diffusion MRI without requiring additional calibration scans.
• The proposed method accurately corrects diffusion tensor imaging metrics, enhancing the reliability of retrospective and prospective data analysis.
• This work offers a significant advancement for improving the quality and interpretability of diffusion MRI data affected by gradient nonlinearities.