Patlak Parametric Image Estimation from Dynamic PET Using Diffusion Model Prior

2025-12-22

AI Summary

Dynamic PET Patlak parametric images suffer from low spatial resolution, high noise, and substantial quantitative bias due to limited photon counts, discontinuous acquisition, and ill-posedness of the kinetic model. To address these challenges, we propose the first end-to-end generative estimation framework that jointly integrates diffusion-based structural priors with physiological kinetic modeling. Specifically, we leverage a score-based diffusion model—pretrained on static whole-body PET images—as a structural prior for Patlak parameter maps, enforce data consistency via the Patlak linear kinetic model, and incorporate patch-wise similarity regularization with iterative denoising optimization. Evaluated on multi-bed, multi-dose whole-body dynamic PET data, our method achieves an average 8.2 dB PSNR gain in slope parameter maps, reduces noise by 37%, and preserves high quantitative accuracy—significantly outperforming conventional voxel-wise fitting and state-of-the-art deep learning approaches.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Dynamic PET enables the quantitative estimation of physiology-related parameters and is widely utilized in research and increasingly adopted in clinical settings. Parametric imaging in dynamic PET requires kinetic modeling to estimate voxel-wise physiological parameters based on specific kinetic models. However, parametric images estimated through kinetic model fitting often suffer from low image quality due to the inherently ill-posed nature of the fitting process and the limited counts resulting from non-continuous data acquisition across multiple bed positions in whole-body PET. In this work, we proposed a diffusion model-based kinetic modeling framework for parametric image estimation, using the Patlak model as an example. The score function of the diffusion model was pre-trained on static total-body PET images and served as a prior for both Patlak slope and intercept images by leveraging their patch-wise similarity. During inference, the kinetic model was incorporated as a data-consistency constraint to guide the parametric image estimation. The proposed framework was evaluated on total-body dynamic PET datasets with different dose levels, demonstrating the feasibility and promising performance of the proposed framework in improving parametric image quality.

Problem

• Enhances parametric image quality in dynamic PET
• Addresses ill-posed kinetic model fitting challenges
• Incorporates diffusion model prior for Patlak estimation

Innovation

• Diffusion model prior for Patlak image estimation
• Pre-trained score function on static PET images
• Kinetic model as data-consistency constraint during inference