Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Patlak Parametric Image Estimation from Dynamic PET Using Diffusion Model Prior

Ziqian Huang, Boxiao Yu, Siqi Li, Savas Ozdemir, Sangjin Bae, Jae Sung Lee, Guobao Wang, Kuang Gong

TL;DR

Dynamic PET parametric imaging suffers from low quality due to ill-posed kinetic fitting and limited counts. This work introduces a diffusion-model–based kinetic framework for Patlak imaging that leverages a score function trained on static total-body PET as a prior and enforces data fidelity through the Patlak model, implemented via RED-Diff and HQS. Across normal- and low-dose total-body datasets, the approach improves CNR, PSNR, and SSIM while preserving structural details, addressing convergence issues seen in fully coupled methods. The method offers a practical route to more reliable voxel-wise Patlak maps and can be extended to other kinetic models and tracers in clinical and research settings.

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.

Patlak Parametric Image Estimation from Dynamic PET Using Diffusion Model Prior

TL;DR

Dynamic PET parametric imaging suffers from low quality due to ill-posed kinetic fitting and limited counts. This work introduces a diffusion-model–based kinetic framework for Patlak imaging that leverages a score function trained on static total-body PET as a prior and enforces data fidelity through the Patlak model, implemented via RED-Diff and HQS. Across normal- and low-dose total-body datasets, the approach improves CNR, PSNR, and SSIM while preserving structural details, addressing convergence issues seen in fully coupled methods. The method offers a practical route to more reliable voxel-wise Patlak maps and can be extended to other kinetic models and tracers in clinical and research settings.

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.
Paper Structure (19 sections, 27 equations, 9 figures, 1 algorithm)

This paper contains 19 sections, 27 equations, 9 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. Method
  3. Diffusion Models
  4. Patlak Model for Parametric Imaging
  5. Patlak Estimation Based on Diffusion Models
  6. Patlak Estimation Based on RED-Diff and HQS
  7. Implementation Details and Reference Methods
  8. Pre-trained score function based on static total-body PET datasets
  9. Proposed Method
  10. Reference Methods
  11. Experimental Setup
  12. Datasets
  13. Data Analysis
  14. Results
  15. Analysis of Intermediate Reconstruction Results
  16. ...and 4 more sections

Figures (9)

  • Figure 1: Comparison of estimated Patlak intercept image. (a) Gaussian filtering. (b) DDPM-based method.
  • Figure 2: Time activity curves and Patlak fitting of one subject in COVID-19 datasets. (a) Extracted time activity curves of the aorta region and organs. (b) Patlak plot for the liver TAC.
  • Figure 3: Three-view of one patch showing intermediate results across data consistency iterations for a COVID-19 subject.
  • Figure 4: Three-view of the same patch showing intermediate results across denoising iterations for a COVID-19 subject.
  • Figure 5: Sagittal and coronal views of Patlak slope images reconstructed from normal-dose datasets using different methods, including the baseline iterative method, Gaussian filtering, NLM, HYPR, and the proposed method. Results are shown for (a) one subject from the COVID-19 dataset and (b) one subject from the GUC dataset.
  • ...and 4 more figures