PET/CT Attenuation Correction With Low-Dose Image Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing PET/CT systems face challenges in spatial misalignment between PET and CT images, leading to artifacts and increased radiation exposure, particularly in total-body scanning, which poses risks to patients, especially children.
Innovation Solution
A dual dose reduction strategy using low-dose PET and CT images for attenuation correction, employing a deep learning model to generate standard-dose PET images by performing multi-scale feature extraction, adaptive spatial alignment, and scale-invariant feature extraction to fuse and upsample images, reducing both PET and CT doses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If CT-based attenuation correction is used, then quantitative analysis accuracy is improved, but spatial misalignment artifacts and radiation exposure increase
Solution Approach 1:
The patent changes the parameter of CT dose by using low-dose CT images for attenuation correction while maintaining quantitative analysis accuracy through deep learning-based correction. The system reduces CT dose parameters while compensating for potential quality loss through neural network reconstruction.
Solution Approach 2:
The patent creates a virtual copy of standard-dose CT images from low-dose images using deep learning models. The neural network learns to reconstruct high-quality attenuation correction images from low-dose inputs, effectively copying the appearance and quality of standard-dose images without the radiation burden.
2Area of stationary object
If total-body scanning is performed, then imaging coverage is improved, but radiation exposure increases
Solution Approach 1:
The patent segments the total-body scanning process into two independent low-dose scans (PET and CT) that can be performed separately and combined through deep learning. This segmentation allows each scan to use reduced dose while the neural network reconstructs the complete low-dose total-body images, reducing overall radiation exposure compared to traditional high-dose sequential scanning.
3Object-affected harmful factors
If low-dose CT images are used for attenuation correction, then radiation exposure is reduced, but image quality deteriorates
Solution Approach 1:
The patent implements feedback mechanisms where the deep learning model continuously refines the reconstruction by comparing low-dose CT images with corresponding PET images and standard-dose reference data. The neural network adjusts its reconstruction based on feedback from multi-scale feature extraction and spatial alignment, progressively improving image quality while maintaining low-dose acquisition.
Solution Approach 2:
The patent adds the dimension of deep learning feature space to bridge the quality gap between low-dose and standard-dose images. By transforming images into multi-scale feature maps and performing non-linear reconstruction through neural networks, the system recovers high-quality images from low-dose inputs in a way that traditional linear interpolation cannot achieve.
4Productivity
If sequential scanning is performed, then imaging completeness is improved, but radiation exposure and scanning time increase
Solution Approach 1:
The patent performs preliminary action by acquiring low-dose PET and CT images separately before the actual diagnostic reconstruction. These low-dose preliminary scans serve as input data for the deep learning model, which then reconstructs the final diagnostic images. This preliminary low-dose acquisition strategy maintains imaging completeness while minimizing radiation exposure compared to traditional high-dose sequential scanning.
Data Source
AI summary
A dual dose reduction attenuation correction method for a PET/CT system includes: performing multi-scale feature extraction on a low-dose ACCT image to obtain a first feature map set{fCTj❘j=1,2,…,N},and performing the multi-scale feature extraction on a low-dose non-attenuation-corrected PET image to obtain a second feature map set{f PETj<semantics definitionURL="">❘<annotation encoding="Mathematica">"\[RightBracketingBar]"</annotation></semantics>j=1,2,… ,N},where N is a set value; performing adaptive spatial alignment between a first feature mapf CTjand a second feature mapf PETj,and matching and fusing two aligned feature maps to obtain a third feature map set{fCT-PETj<semantics definitionURL="">❘<annotation encoding="Mathematica">"\[RightBracketingBar]"</annotation></semantics>j=1,2,… ,N};performing M iterations of scale-invariant feature extraction on a third feature mapfCT-PETNto obtain an attenuation-corrected feature mapfAC-PETN,where M is a set value; upscaling a feature mapfAC-PETjto match a size offCT-PETj-1,and concentratingfCT-PETj-1withfAC-PETjto obtain an attenuation-corrected feature mapfAC-PETj-1,where j-N, . . . , 2; and obtaining a visual standard-dose PET image based on a feature mapfAC-PET1.