Parallel MRI Acceleration Optimization via Non-Integer k-Space Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current parallel MRI techniques face challenges in achieving optimal acceleration while maintaining signal-to-noise ratio (SNR), particularly in three-dimensional imaging, where acceleration methods often result in reduced SNR and aliasing issues due to non-integer k-space increments and foldover effects.
Innovation Solution
A method for three-dimensional parallel MRI that determines optimal in-plane acceleration factors along phase-encoding directions to generate a k-space sampling pattern, allowing for optimized image reconstruction with minimized noise amplification and SNR loss, using coil sensitivity maps and parallel image reconstruction techniques like SENSE.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If acceleration methods are used to reduce acquisition time, then productivity is improved, but signal-to-noise ratio deteriorates
Solution Approach 1:
The patent changes the acceleration parameter from integer values to non-integer values optimized for specific anatomical regions. By calculating region-specific acceleration factors based on coil sensitivity maps and anatomical variability, the system achieves higher overall acceleration while maintaining acceptable SNR in critical areas. This resolves the contradiction by allowing accelerated acquisition without uniform SNR degradation across the entire image.
Solution Approach 2:
The patent applies different acceleration factors to different regions of the image based on local coil sensitivity and anatomical importance. Critical regions with poor coil sensitivity or high anatomical variability are assigned lower acceleration factors to preserve SNR, while regions with good sensitivity tolerate higher acceleration. This local optimization resolves the contradiction between overall speed improvement and local SNR maintenance.
2Productivity
If higher acceleration factors are used, then productivity is improved, but noise amplification increases
Solution Approach 1:
The patent optimizes acceleration factors to non-integer values that balance noise amplification (g-factor) against acquisition time reduction. By using coil sensitivity maps to calculate region-specific acceleration factors, the system identifies the optimal point where further acceleration would cause unacceptable noise amplification in critical regions. This resolves the contradiction by preventing excessive noise amplification while still achieving high overall acceleration.
3Ease of operation
If integer acceleration factors are used, then ease of operation is improved, but manufacturing precision deteriorates
Solution Approach 1:
The patent extends acceleration factors from integer values to continuous non-integer values, allowing precise optimization for each anatomical region and coil configuration. This continuous parameter optimization improves image reconstruction accuracy by matching the acceleration factor to the actual coil sensitivity distribution, rather than forcing integer values that may be suboptimal. The system maintains ease of operation through automated calculation of these optimized factors.
Solution Approach 2:
The patent performs preliminary calculation of region-specific acceleration factors using coil sensitivity maps before the actual imaging acquisition. This preliminary optimization step determines the precise acceleration factors that will minimize noise amplification and maximize image quality for the specific patient anatomy and coil configuration. This resolves the contradiction by preparing optimized parameters in advance, making the complex non-integer acceleration transparent to the operator.
Data Source
AI summary
A method for three-dimensional parallel magnetic resonance imaging (MRI) using an MRI system is provided. The method includes determining in-plane acceleration factors that optimize a selected criterion, such as an image quality criterion defined by maximal noise amplification in a reconstructed image. The estimated in-plane acceleration factors are used to establish a k-space sampling pattern, which is used to acquire k-space data. An image is reconstructed from the acquired k-space data using a parallel image reconstruction technique.