Amyloid PET Translation From FDG PET Using Conditioned Diffusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current amyloid positron emission tomography (PET) imaging methods are limited in their ability to accurately synthesize amyloid PET images from Fluorodeoxyglucose (FDG) PET images, leading to high costs and limited availability for early detection and monitoring of Alzheimer's disease progression, with existing GAN-based architectures facing issues of mode collapse and requiring multiple modalities.
Innovation Solution
A conditioned diffusion-based generative model is employed to iteratively train FDG PET images using subject-specific demographics and disease status, leveraging a forward diffusion mechanism to synthesize high-quality amyloid PET images, incorporating a four-level encoder with channel attention and a U-Net architecture for precise image translation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If GAN-based architectures are used for amyloid PET image synthesis, then image generation capability is improved, but mode collapse occurs and diverse meaningful outputs are limited
Solution Approach 1:
The patent uses diffusion models to generate multiple diverse copies of amyloid PET images from FDG PET input, avoiding mode collapse by exploring different possible outcomes through iterative noise addition and removal processes, thereby maintaining both synthesis quality and output diversity
Solution Approach 2:
The patent changes the fundamental parameters of the generation process by using diffusion models with controllable noise schedules and conditioning mechanisms, allowing systematic exploration of the output space while maintaining image quality through controlled parameter adjustments during the diffusion process
2Measurement precision
If amyloid PET imaging is used for direct assessment, then detection accuracy is improved, but cost increases and availability decreases
Solution Approach 1:
The patent introduces FDG PET imaging as an intermediary that is more widely available and less costly, using it to predict amyloid PET images through diffusion models, thereby maintaining detection accuracy while improving accessibility through a surrogate imaging modality
Solution Approach 2:
The patent creates synthetic copies of amyloid PET images from FDG PET data using diffusion-based generative models, producing realistic amyloid deposition patterns without requiring actual amyloid PET scans, thus reducing cost while preserving diagnostic information
3Manufacturing precision
If multiple modalities are used for PET image synthesis, then synthesis quality is improved, but device complexity increases
Solution Approach 1:
The patent extracts and removes the requirement for multiple PET modalities by achieving high-quality amyloid PET synthesis using only FDG PET input, simplifying the system while maintaining synthesis quality through advanced diffusion-based generative approaches
Data Source
AI summary
Amyloid deposition is considered a viable biomarker for Alzheimer's disease which is expensive and less used modality to study amyloid spatial distribution in brain. The present disclosure addresses problems of conventional approaches which exhibit trade-off between performance and mode collapse. This disclosure provides a system and method for subject-specific amyloid position emission tomography (PET) translation using conditioned diffusion-based generative model. The present disclosure discloses an architecture for synthesizing subject-specific amyloid images with a diffusion model. First effectiveness of a relationship between Fluorodeoxyglucose (FDG) and amyloid PET images is identified. Further, a framework is provided to synthesize Amyloid PET by utilizing its connection to FDG PET of same subject and cross-subject trends in amyloid deposition by incorporating age, gender and disease status information in learning process. In the other words, a diffusion model inspired image translation is provided to synthesize Amyloid PET from FDG PET and cross-subject amyloid deposition patterns.