Motion-Resolved Cardiovascular MRI Reconstruction for Free-Breathing Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cardiovascular magnetic resonance imaging (CMR) systems face challenges in reconstructing images with motion artifacts due to respiratory and cardiac movements, leading to degraded image quality and clinical utility.
Innovation Solution
A computer-implemented method involving k-space data sorting into bins corresponding to cardiac and respiratory phases, with iterative computation of soft participation weights and weighted optimization to resolve motion artifacts, using an expectation maximization (EM) framework with ADMM iterations for convergence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If k-space data is sorted into bins corresponding to cardiac and respiratory phases, then motion artifacts are reduced, but computational complexity increases
Solution Approach 1:
The patent segments k-space readouts into multiple bins based on cardiac and respiratory phases. Each bin corresponds to a specific phase combination, allowing separate image reconstruction for each phase. This segmentation enables motion artifact reduction by isolating data from different motion states, while the systematic binning approach manages computational complexity through organized data structure.
Solution Approach 2:
The patent employs iterative expectation maximization (EM) algorithm that dynamically updates participation weights for each bin across multiple iterations. The algorithm adapts the weighting scheme based on previous iteration results, allowing the system to dynamically optimize the balance between motion artifact reduction and computational efficiency until convergence criteria are met.
2Measurement precision
If iterative computation of soft participation weights is performed, then binning inaccuracies are mitigated, but processing time increases
Solution Approach 1:
The iterative EM algorithm implements feedback by using image estimates from previous iterations to update participation weights, which then improve the next image estimate. This feedback loop progressively refines binning accuracy by incorporating information from all bins weighted by their participation probabilities, converging to an optimal solution that mitigates binning inaccuracies while providing a stopping criterion to manage processing time.
3Ease of operation
If volumetric CMR is performed under free-breathing conditions, then patient comfort is improved, but motion artifacts increase
Solution Approach 1:
The patent converts the harmful effect of free-breathing motion into a benefit by using the motion itself as a classification feature. Instead of trying to eliminate motion, the system sorts k-space readouts into bins based on their cardiac and respiratory phase, treating the motion patterns as useful information for organizing data. This allows free-breathing acquisition to maintain patient comfort while the binning process recovers image quality by separating data from different motion states.
Data Source
AI summary
An example computer-implemented method for image reconstruction, includes: receiving k-space data from a magnetic resonance imaging (MRI) machine, the k-space data comprising a plurality of k-space readouts; sorting the k-space readouts into a set of bins comprising binned k-space data, each of the set of bins corresponding to a respective phase of a respiratory cycle; iteratively performing the steps of: (i) computing a soft participation weight for each k-space readout, where the soft participation weight reflects a posterior probability that each k-space readout belongs to each of the set of bins; and (ii) updating an image estimate by solving a weighted optimization problem; determining a convergence criterion is reached; and outputting a motion-resolved volumetric MRI image when the convergence criterion is reached.