Correction of high-order phase variation effects in dynamic field monitoring

TL;DR

This paper presents a correction algorithm for high-order phase variations in field monitoring, improving image quality in MRI by reducing phase-related artifacts caused by distant probes.

Contribution

The work introduces a novel three-component correction algorithm that effectively reduces nonlinear phase errors in field monitoring for MRI, enhancing image reconstruction accuracy.

Findings

Significant artifact reduction in diffusion-weighted images

Improved image quality with phase correction

Stepwise fitting yields best correction results

Abstract

Purpose: Field monitoring measures field perturbations, which can be accounted for during image reconstructions. In certain field monitoring environments, significant phase deviations can arise far from isocenter due to the finite extent of the gradient and/or main magnet. This can degrade the accuracy of field dynamics when field probes are placed near or outside the diameter spherical volume of the gradient coils and/or main magnet, leading to corrupted image quality. The objective of this work was to develop a correction algorithm that reduces errors from highly nonlinear phase variations at distant field probes in field dynamic fits. Methods: The algorithm is split into three components. Component one fits phase coefficients one spatial order at a time, while the second implements a weighted least squares solution based on probe distance. After initial fitting, component three calculates phase residuals and removes the phase for distant probes before re-fitting. Two healthy volunteers were scanned on a head-only 7T MRI using diffusion-weighted single-shot spiral and EPI sequences and field monitoring was performed. Images were reconstructed with and without phase coefficient correction and compared qualitatively. Results: The algorithm was able to correct corrupted field dynamics, resulting in image quality improvements. Significant artefact reduction was observed when correcting higher order fits, especially for diffusion weighted images. Stepwise fitting provided the most correction benefit, which was marginally improved when adding weighted least squares and phase residual corrections. Conclusion: The proposed algorithm can mitigate effects of phase errors in field monitoring, providing improved reliability of field dynamic characterization.