MRI Fast Spin Echo K-Space Sampling Density Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current MRI techniques face challenges in reducing acquisition time while maintaining image quality, particularly with non-Cartesian sampling methods that can lead to artifacts and reduced Signal-to-Noise Ratio (SNR), necessitating a method to efficiently sample k-space data.
Innovation Solution
The proposed method involves using a fast spin echo (FSE) sequence with specific encoding gradients that include steady phases and transition phases, and employing image reconstruction techniques such as compressed sensing, parallel imaging, or partial Fourier reconstruction to undersample k-space data, optimizing the distribution of sampling density in the k-space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If non-Cartesian sampling methods are used to reduce acquisition time, then acquisition time is reduced, but image quality deteriorates due to artifacts and reduced SNR
Solution Approach 1:
The patent applies local quality by using variable sampling density in different regions of k-space. Specifically, the center region of k-space is oversampled while the peripheral region is undersampled. This localized differentiation allows the method to reduce overall acquisition time through undersampling while maintaining image quality by preserving sufficient data in the critical center region that contains low-frequency information essential for image contrast and structure.
2Reliability
If RF pulse sequences are transmitted for long time exposure, then image data can be adequately acquired, but physical damage may occur to the subject
Solution Approach 1:
The patent employs self-service by using the inherent properties of the MR signal and k-space structure to achieve adequate image data acquisition in reduced time. Through optimized undersampling strategies and advanced reconstruction algorithms that exploit signal coherence and redundancy, the system obtains sufficient diagnostic information without requiring prolonged RF exposure, thereby avoiding physical damage to the subject while maintaining reliable image data acquisition.
3Loss of time
If k-space data is undersampled to reduce acquisition time, then acquisition time is reduced, but SNR and image quality deteriorate
Solution Approach 1:
The patent applies parameter changes by systematically varying the sampling density parameter across different k-space regions. The sampling rate is adjusted as a function of spatial frequency, with higher sampling rates in the center region and lower rates in the periphery. This parameter optimization allows the method to reduce overall data acquisition while maintaining sufficient SNR in critical regions, thereby improving the time-SNR trade-off compared to uniform undersampling.
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 reduces acquisition time while minimizing artifacts and maintaining image quality by efficiently sampling k-space data, particularly in the center region of k-space, thereby improving the Signal-to-Noise Ratio (SNR) and detail restoration in reconstructed images.
Implementation Method 1
Magnetic Resonance Imaging (MRI) is a widely used medical technique which produces images of a region of interest (ROI) by exploiting a powerful magnetic field
Implementation Method 2
The gradient magnet may be used to generate magnet field gradients
Implementation Method 3
The RF coil may be used to transmit RF signals to or receive MR signals from the ROI
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
A system (100) and method for magnetic resonance imaging. The method includes generating a main magnetic field through a region of interest (ROI), applying a slice selection gradient to an slice of the ROI, applying a plurality of RF pulses to the slice to generate a plurality of echoes, applying a first encoding gradient and a second encoding gradient on the echoes, and generating MR images based on the echoes.