PET Attenuation Map Alignment for MR Coil Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In PET/MR imaging systems, the inability to generate suitable attenuation correction maps for hardware objects like MR coils and patient tables leads to significant errors in PET image reconstruction due to unaccounted photon attenuation, as MR imaging does not measure electron density and conventional CT-based correction assumes fixed object positions.
Innovation Solution
The system registers attenuation correction maps for hardware objects and the patient using intrinsic knowledge, CT or transmission PET imaging, and employs a trained model to directly convert MR images to pseudo-CT images, optimizing the alignment and orientation of these maps with PET data to ensure consistent attenuation correction across all view angles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If MR imaging is used to generate attenuation correction maps, then patient body attenuation can be corrected, but hardware objects like MR coils and patient tables cannot be visualized and thus cannot be used for accurate attenuation correction
Solution Approach 1:
The patent segments the attenuation correction process into two parts: MR imaging for patient body attenuation correction, and separate CT-based or transmission PET-based attenuation maps for hardware objects. This segmentation allows each modality to contribute where it is most effective, resolving the contradiction between MR's ability to image soft tissue and its inability to visualize hardware objects.
Solution Approach 2:
The patent introduces transmission PET imaging or CT imaging as an intermediary to capture hardware object attenuation information that MR imaging cannot detect. This intermediary modality bridges the information gap, providing the missing hardware object data needed for complete attenuation correction.
2Measurement precision
If pre-generated attenuation correction maps for hardware objects are used, then attenuation correction can be performed, but spatial position discrepancies between actual and assumed hardware object locations reduce image quality
Solution Approach 1:
The patent transforms the static, fixed-position assumption of hardware objects into a dynamic registration process. By using image registration techniques that allow for spatial transformation and alignment based on actual imaging data, the system adapts to the true positions of hardware objects, resolving the spatial misalignment issue.
Solution Approach 2:
The patent implements a feedback mechanism where the actual PET data and MR images are used to evaluate and refine the alignment of attenuation correction maps. The registration process continuously adjusts the positioning of hardware object attenuation maps based on the consistency of attenuation-corrected PET data across different view angles, ensuring optimal spatial alignment.
3Device complexity
If hardware objects are ignored during attenuation correction, then the correction process is simpler, but significant errors occur in reconstructed PET images due to unaccounted photon attenuation
Solution Approach 1:
The patent segments the attenuation correction into patient-specific correction (using MR images) and hardware-specific correction (using pre-generated maps). This segmentation allows the system to handle hardware objects systematically without overwhelming complexity, while still achieving accurate PET image reconstruction by combining both correction components.
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 improves the accuracy of PET image reconstruction by accounting for hardware object attenuation, reducing errors and enhancing image quality by ensuring consistent attenuation correction across all view angles.
Implementation Method 1
employs a trained model to directly convert MR images to pseudo-CT images
Implementation Method 2
These intervening objects attenuate the photons based on their respective electron densities
Data Source
AI summary
Systems and methods include acquisition of magnetic resonance data of a subject disposed in a first position, acquisition of positron emission tomography data of imaging hardware and of the subject disposed substantially in the first position, generation of a subject attenuation correction map of the subject based on the magnetic resonance data, determination of an imaging hardware attenuation correction map associated with the imaging hardware, determination of a target location and orientation of the imaging hardware attenuation correction map with respect to the positron emission tomography data and based on the positron emission tomography data and on the subject attenuation correction map, and application of attenuation correction to the positron emission tomography data based on the imaging hardware attenuation correction map in the target location and orientation and the subject attenuation correction map to generate attenuation-corrected positron emission tomography data.