Through-Time GRAPPA MRI Calibration for High Frame Rate Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic resonance imaging (MRI) applications face challenges in achieving high resolution and high frame rates simultaneously, particularly in cardiac imaging, where acquiring fully-sampled calibration data sets is time-consuming and requires multiple breath holds, leading to inconsistent images and prolonged data acquisition times.
Innovation Solution
The through-time generalized auto-calibrating partially parallel acquisition (GRAPPA) method acquires calibration data with a varying gradient perpendicular to the non-Cartesian encoded plane, treating different groups of lines as unique calibration frames to reduce the number of fully-sampled data sets needed, allowing for high frame rates and resolution without sacrificing image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If fully-sampled calibration data sets are acquired to enable under-sampling image data, then calibration accuracy is improved, but data acquisition time increases significantly
Solution Approach 1:
The patent applies partial action by acquiring only a reduced set of calibration data points rather than fully-sampled data. Specifically, it acquires calibration data at fewer phase encoding steps (e.g., every fourth or eighth line) and uses iterative reconstruction algorithms to recover the missing information, thereby reducing calibration time while maintaining sufficient accuracy for the application.
Solution Approach 2:
The patent changes the sampling parameters by using different acceleration factors (R=2, R=4, R=8) and adjusting the number of calibration lines acquired. It also modifies the reconstruction parameters through iterative SENSE-GRAPPA algorithms with varying regularization terms to achieve optimal balance between calibration speed and accuracy.
2Volume of moving object
If multiple breath holds are used to acquire multiple slices of the heart, then complete cardiac coverage is achieved, but image consistency deteriorates due to varying breath hold techniques
Solution Approach 1:
The patent segments the cardiac imaging into multiple thin slices acquired sequentially during a single breath hold, rather than requiring multiple breath holds for different slice groups. This segmentation approach allows complete cardiac coverage while maintaining consistent breathing conditions throughout the acquisition.
Solution Approach 2:
The patent uses periodic radial k-space sampling during the single breath hold to systematically cover multiple slices over time. The radial trajectories are periodically repeated with different phase encoding offsets, ensuring complete coverage while the patient maintains a single consistent breath-hold position throughout.
3Productivity
If under-sampling is increased to improve frame rates, then temporal resolution is improved, but image resolution deteriorates
Solution Approach 1:
The patent performs preliminary calibration data acquisition and iterative reconstruction before final image generation. By pre-computing the SENSE-GRAPPA sensitivity maps and reconstruction kernels from the acquired calibration data, it enables rapid reconstruction of high-resolution images from under-sampled data without sacrificing temporal resolution.
Solution Approach 2:
The patent introduces iterative reconstruction algorithms as an intermediary process between raw under-sampled data and final images. These algorithms act as a mediator that recovers high-resolution information from the under-sampled data by incorporating physical constraints and prior knowledge about the imaging system and object being imaged.
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 enables the acquisition of high-resolution images at high frame rates, such as a whole-heart stack of images at 48 ms/cardiac phase during a single breath hold, reducing data acquisition time and improving image consistency by using additional sensitivity information from gradient-encoded data points.
Implementation Method 1
Some magnetic resonance imaging (MRI) applications desire both high resolution and high frame rates
Data Source
AI summary
Example apparatus and methods control a magnetic resonance imaging (MRI) apparatus to acquire, from an object to be imaged, throughout a period of time, a partitioned non-Cartesian fully-sampled calibration data set. Different groups of lines in the calibration data set are acquired at different points in time under different gradient encoding conditions that yield phase encoding in the direction perpendicular to the non-Cartesian encoded plane. The MRI apparatus is controlled to acquire an under-sampled non-Cartesian data set from the object to be imaged and to reconstruct an image from the under-sampled data set based, at least in part, on a through-time GRAPPA calibration. A GRAPPA weight set can be computed from data in different groups of lines in the calibration data set because different groups of lines can be treated as unique calibration time frames due to phase encoding produced by the different gradient encoding conditions.