MRI Field of View Extension via Nonlinear Gradient Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Magnetic Resonance Imaging (MRI) techniques, such as B0 homogenization using gradient enhancement (HUGE), face limitations in extending the field of view (FOV) due to the need for optimal readout gradient determination for each side of the subject, leading to increased acquisition time and truncation of anatomical regions, especially in large body habitus patients, which is critical for accurate patient-specific models in hybrid imaging applications like PET and radiation therapy planning.
Innovation Solution
The method involves determining linear field gradients associated with the MR scanner's gradient coil, acquiring a k-space dataset using multiple readout gradient amplitudes, and employing an iterative reconstruction process, like the Kaczmarz algorithm, to generate an extended FOV image that accounts for gradient distortions, allowing for full body modeling and PET reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If the HUGE technique is used to extend FOV to 60 cm, then the field of view is improved, but the acquisition time increases substantially due to needing to determine optimal readout gradient for each side at each bed position
Solution Approach 1:
The patent measures gradient nonlinearity once at each bed position before actual imaging, and uses this pre-measured information in the reconstruction process. This preliminary measurement eliminates the need to repeatedly determine optimal readout gradients for each side during acquisition, thus extending FOV without proportionally increasing acquisition time
Solution Approach 2:
The patent incorporates measured gradient nonlinearity data into the image reconstruction process as feedback. By using this feedback information to correct for distortions during reconstruction, the system can achieve extended FOV with accurate anatomical representation without requiring multiple gradient optimization measurements
2Manufacturing precision
If the typical gradient linearity region of 50 cm is used, then the system operates within linear gradient constraints, but anatomical regions such as arms and large body habitus areas are truncated
Solution Approach 1:
The patent converts the harmful effect of gradient nonlinearity (which causes distortion) into a beneficial measurement. By characterizing the nonlinearity and incorporating it into the reconstruction process, the system can image beyond the traditional 50 cm linear region while maintaining geometric accuracy. The previously problematic nonlinear region becomes usable imaging space
Solution Approach 2:
The patent changes the approach from requiring linear gradients to accepting and correcting for nonlinear gradients. By measuring the actual gradient field and using this information in reconstruction, the system expands the usable imaging area beyond the traditional linear gradient limit while maintaining image accuracy
3Area of stationary object
If multiple readout gradient amplitudes are acquired to extend FOV, then anatomical coverage is improved, but the complexity of the reconstruction process increases
Solution Approach 1:
The patent introduces gradient nonlinearity measurements as an intermediary element between the raw k-space data and the final image reconstruction. This intermediary information acts as a correction factor that simplifies the reconstruction process by providing known distortion characteristics, making the extended FOV reconstruction more tractable than if no such information were available
Data Source
AI summary
A method for increasing field of view (FOV) in magnetic resonance imaging includes determining linear field gradients of associated with a gradient coil of a magnetic resonance (MR) scanner and using the MR scanner to acquire a k-space dataset representative of a patient using a plurality of readout gradient amplitudes. An extended FOV image is generated based on the k-space dataset using an iterative reconstruction process to solve a forward model that incorporates a measurement of gradient distortion as a deviation from the linear field gradients.