Accelerated PROPELLER MRI with Compressed Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
PROPELLER magnetic resonance imaging (MRI) acquisition methods have a relatively long scan time, making existing imaging acceleration methods inadequate for applications where motion is prevalent and difficult to control, such as fetal imaging or imaging of subjects with Parkinson's disease, as they are sensitive to motion and require additional setup time.
Innovation Solution
An accelerated PROPELLER pulse sequence with an undersampling pattern in k-space is used, reducing the number of views and blades, combined with compressed sensing and Non-uniform Fast Fourier Transform (NUFFT) for phase and motion corrections to reconstruct images efficiently, thereby reducing scan time and artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If PROPELLER sampling scheme is used, then motion sensitivity is improved, but scan time increases
Solution Approach 1:
The patent applies partial sampling in k-space by acquiring only a subset of the usual PROPELLER blades and views. Specifically, it samples every other blade or view (factor of 2 undersampling) while maintaining the rotational PROPELLER trajectory, thereby reducing scan time by approximately 50% while preserving sufficient motion robustness through the remaining sampled data points
Solution Approach 2:
The patent changes the sampling parameters by introducing an undersampling factor R that reduces the number of acquired blades and views. This parameter modification allows the system to operate with fewer measurements (e.g., sampling every other blade instead of all blades) while using compressed sensing reconstruction to maintain image quality and motion correction capabilities
2Loss of time
If imaging acceleration is applied, then scan time is reduced, but image quality deteriorates
Solution Approach 1:
The patent introduces compressed sensing as an intermediary reconstruction technique that bridges the gap between undersampled data and high-quality images. The compressed sensing algorithm uses sparsity constraints and iterative optimization to reconstruct images from the reduced set of k-space samples, thereby maintaining diagnostic image quality despite the accelerated acquisition
Solution Approach 2:
The patent employs iterative compressed sensing reconstruction with feedback mechanisms that refine the image estimate progressively. The reconstruction process uses the acquired undersampled data to generate initial images, evaluates sparsity in transformed domains (such as wavelet or gradient domains), and iteratively adjusts the reconstruction to maximize consistency with the measured data while enforcing sparsity constraints, thereby recovering fine details and reducing artifacts
3Object-affected harmful factors
If PROPELLER acquisition is used, then motion artifact resistance is improved, but scan time increases
Solution Approach 1:
The patent applies partial sampling to the PROPELLER sequence by acquiring only a subset of blades and views (e.g., every other blade) while maintaining the rotational sampling pattern. This partial acquisition reduces scan time by approximately 50% while the remaining sampled data, when processed with compressed sensing, preserves the motion artifact resistance characteristic of full PROPELLER imaging
Solution Approach 2:
The patent modifies the acquisition parameters by introducing an undersampling factor that reduces the number of blades and views acquired. Despite this reduction, the rotational PROPELLER trajectory is maintained in the sampled data, and compressed sensing reconstruction recovers the motion-corrected image, thereby preserving motion artifact resistance with reduced scan time
Data Source
AI summary
A magnetic resonance (MR) imaging acceleration method is provided. The method includes applying, by an MR system, a pulse sequence having a k-space trajectory of a plurality of blades being rotated in k-space, each blade including a plurality of views, wherein the k-space trajectory has an undersampling pattern in the k-space. The method also includes receiving k-space data of a subject acquired by the pulse sequence, reconstructing MR images of the subject based on the k-space data using compressed sensing, and outputting the reconstructed images.