SPECT Attenuation Compensation via Deep Learning Scatter Projections
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current SPECT imaging methods require a CT scan for attenuation compensation, leading to increased radiation dose, high costs, and potential misalignment issues, which hinder accurate diagnosis, especially in SPECT systems without a CT component.
Innovation Solution
A physics- and deep learning-based method (PDLAC) generates attenuation maps from scatter-energy window projections, allowing for attenuation compensation in SPECT imaging without a separate transmission scan, using a trained model to combine these projections with photopeak-energy window projections for accurate image reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a CT scan is used for attenuation compensation in SPECT imaging, then the accuracy of attenuation compensation is improved, but the radiation dose increases
Solution Approach 1:
The patent extracts the attenuation map generation function from the CT scanner and implements it using scatter-energy window projections from the SPECT scanner itself. This separates the attenuation compensation function from the CT component, allowing SPECT systems without CT to perform attenuation compensation using only their existing SPECT data.
Solution Approach 2:
The patent creates a virtual attenuation map that replicates the function of a CT-based attenuation map without requiring actual CT imaging. The deep learning model generates this virtual map from scatter projections, effectively copying the essential information needed for attenuation compensation without the harmful radiation exposure of a CT scan.
2Measurement precision
If a CT scanner is included in the SPECT system, then attenuation compensation can be performed, but the device complexity and cost increase
Solution Approach 1:
The patent makes the SPECT scanner multi-functional by enabling it to perform both SPECT imaging and attenuation map generation using its existing scatter-energy window projection capability. This eliminates the need for a separate CT scanner, as the SPECT system universally handles both diagnostic imaging and attenuation compensation functions.
Solution Approach 2:
The SPECT system serves itself by generating its own attenuation map from its own scatter projections without requiring an external CT scanner. The deep learning model processes the scatter data produced during normal SPECT acquisition to create the attenuation map needed for image reconstruction.
3Measurement precision
If a CT scan is performed for attenuation compensation, then accurate attenuation maps can be generated, but misalignment between SPECT and CT scans may occur
Solution Approach 1:
The patent merges the attenuation map generation process with the SPECT data acquisition process. Since both the scatter projections and the photopeak data come from the same SPECT scanner and are acquired simultaneously during the same scan, they are inherently aligned in space and time, eliminating misalignment issues between separate SPECT and CT scans.
4Measurement precision
If a CT scanner is added to perform attenuation compensation, then diagnostic accuracy is improved, but patient convenience deteriorates
Solution Approach 1:
The patent enables continuous acquisition of both scatter-energy window projections and photopeak-energy window projections during a single SPECT scan without interruption. This continuous simultaneous acquisition eliminates the need for separate CT scanning sessions, maintaining patient comfort while achieving accurate attenuation compensation.
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
The PDLAC method achieves statistically noninferior performance to CT-based attenuation compensation methods, improving the accuracy of detecting myocardial perfusion defects while reducing radiation dose and costs, and enabling diagnostic cardiac SPECT at remote locations.
Implementation Method 1
receive a scatter-energy window projection of a first subject to be examined; execute the model with the scatter-energy window projection of the first subject as an input, wherein the model generates an attenuation map
Implementation Method 2
receive a photopeak-energy window projection of the first subject to be examined; perform attenuation compensation on the photopeak-energy window projection
Data Source
AI summary
A system for single-photon emission computed tomography (SPECT) is provided. The system includes a computer device comprises at least one processor in communication with at least one memory device. The at least one processor is programmed to: a) store a model trained to generate an attenuation map of a subject being examined; b) receive a scatter-energy window projection of a first subject to be examined; c) execute the model with the scatter-energy window projection of the first subject as an input, wherein the model generates an attenuation map; d) receive a photopeak-energy window projection of the first subject to be examined; and e) perform attenuation compensation on the photopeak-energy window projection using the generated attenuation map.