Diffusion MRI Distortion Correction Using Orthogonal Adjustment Measurements
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for correcting image distortions in diffusion-weighted magnetic resonance images, such as those using echoplanar imaging, require multiple adjustment measurements across various diffusion directions, leading to lengthy acquisition times and inaccuracies due to eddy currents and movement artifacts, especially when calculating parameters like scaling, shear, and translation.
Innovation Solution
A method that uses two adjustment measurements with specific diffusion weightings in three orthogonal directions to calculate correction parameters, assuming linear superposition of eddy current fields, allowing for the de-skewing of diffusion-weighted MR images with reduced measurement time and improved precision, and incorporates movement detection and signal intensity thresholding to enhance robustness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple adjustment measurements are acquired for every diffusion direction to correct eddy current distortions, then the precision of distortion correction improves, but the acquisition time becomes intolerably long
Solution Approach 1:
The patent segments the diffusion directions into a small set of representative directions (e.g., 5-200 directions) for which adjustment measurements are actually acquired. Instead of measuring all possible directions, the method selects a subset that captures the essential distortion characteristics, thereby reducing acquisition time while maintaining sufficient correction precision.
Solution Approach 2:
The patent performs preliminary acquisition of adjustment measurements with low diffusion weighting (b=150 s/mm²) before the actual diffusion-weighted imaging. These preliminary measurements are used to determine distortion parameters (scaling M, shear S, translation T) that are then extrapolated to correct images with higher b-values (e.g., b=1000 s/mm²), avoiding the need to acquire separate adjustment measurements for each diffusion direction and weighting.
2Loss of time
If adjustment measurements with low diffusion weighting (b=150 s/mm²) are used to determine distortion parameters, then the acquisition time is reduced, but the distortions are not strongly pronounced making precise determination of parameters difficult
Solution Approach 1:
The patent uses feedback by acquiring multiple adjustment measurements and iteratively refining the distortion parameter determination. The system acquires adjustment measurements, determines distortion parameters, and uses these parameters to correct the actual diffusion-weighted images. The method incorporates feedback loops where the correction results inform subsequent measurements or parameter refinements, ensuring accurate parameter determination even with low diffusion weighting.
Solution Approach 2:
The patent changes the diffusion weighting parameter (b-value) strategically. Adjustment measurements are acquired at low b-values (150 s/mm²) to minimize distortion and acquisition time, while the actual diagnostic images are acquired at higher b-values (1000 s/mm²). The distortion parameters determined at low b-values are then extrapolated to the high b-value regime, leveraging the parameter relationship between different diffusion weightings.
3Measurement precision
If multiple adjustment measurements are acquired for each diffusion direction and weighting, then the correction accuracy improves, but the signal-to-noise ratio in high b-value images becomes extremely low making parameter determination difficult
Solution Approach 1:
The patent extracts the distortion correction information from separate adjustment measurements with low diffusion weighting, rather than trying to extract it from the actual high b-value diffusion-weighted images. By taking out the distortion measurement task from the diagnostic imaging process and performing it separately at optimized low b-values, the method preserves the signal-to-noise ratio in the diagnostic images while still obtaining accurate distortion parameters for correction.
4Measurement precision
If adjustment measurements are acquired with inverted polarity to correct distortions, then the distortion parameters can be determined, but two images must be acquired for each diffusion direction and weighting increasing acquisition time
Solution Approach 1:
The patent merges the acquisition of adjustment measurements with the actual diffusion-weighted imaging sequence. Instead of acquiring separate adjustment measurements with inverted polarity for each diffusion direction, the method combines distortion correction measurements and diagnostic imaging into a unified sequence, acquiring both types of data simultaneously or in an integrated manner, thereby halving the acquisition time requirement.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach significantly reduces acquisition time and improves the accuracy of diffusion-weighted image correction by minimizing the need for multiple adjustment measurements and accounting for movement, resulting in more precise and robust parameter determination.
Implementation Method 1
The eddy current fields that are generated by the diffusion gradients, however, lead to image distortions whose appearance depends both on the amplitude of the gradients (i.e. the diffusion weighting) and on their direction.
Data Source
AI summary
In a method and magnetic resonance (MR) system for correction of image distortions that occur in acquisitions of diffusion-weighted MR images of an examination subject a first adjustment measurement with a first diffusion weighting is implemented, a second adjustment measurement with a second diffusion weighting is implemented and correction parameters to de-skew diffusion-weighted MR images are automatically calculated in a computer on the basis of the two adjustment measurements. One of the two adjustment measurements is implemented with a predetermined diffusion weighting in three orthogonal diffusion directions, and correction parameters are determined for the three orthogonal diffusion directions.