Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

RelA-Diffusion: Relativistic Adversarial Diffusion for Multi-Tracer PET Synthesis from Multi-Sequence MRI

Minhui Yu, Yongheng Sun, David S. Lalush, Jason P Mihalik, Pew-Thian Yap, Mingxia Liu

TL;DR

RelA-Diffusion is proposed, a Relativistic Adversarial Diffusion framework for multi-tracer PET synthesis from multi-sequence MRI that captures richer structural information to guide image generation and introduces a gradient-penalized relativistic adversarial loss to the intermediate clean predictions of the diffusion model.

Abstract

Multi-tracer positron emission tomography (PET) provides critical insights into diverse neuropathological processes such as tau accumulation, neuroinflammation, and $β$-amyloid deposition in the brain, making it indispensable for comprehensive neurological assessment. However, routine acquisition of multi-tracer PET is limited by high costs, radiation exposure, and restricted tracer availability. Recent efforts have explored deep learning approaches for synthesizing PET images from structural MRI. While some methods rely solely on T1-weighted MRI, others incorporate additional sequences such as T2-FLAIR to improve pathological sensitivity. However, existing methods often struggle to capture fine-grained anatomical and pathological details, resulting in artifacts and unrealistic outputs. To this end, we propose RelA-Diffusion, a Relativistic Adversarial Diffusion framework for multi-tracer PET synthesis from multi-sequence MRI. By leveraging both T1-weighted and T2-FLAIR scans as complementary inputs, RelA-Diffusion captures richer structural information to guide image generation. To improve synthesis fidelity, we introduce a gradient-penalized relativistic adversarial loss to the intermediate clean predictions of the diffusion model. This loss compares real and generated images in a relative manner, encouraging the synthesis of more realistic local structures. Both the relativistic formulation and the gradient penalty contribute to stabilizing the training, while adversarial feedback at each diffusion timestep enables consistent refinement throughout the generation process. Extensive experiments on two datasets demonstrate that RelA-Diffusion outperforms existing methods in both visual fidelity and quantitative metrics, highlighting its potential for accurate synthesis of multi-tracer PET.

RelA-Diffusion: Relativistic Adversarial Diffusion for Multi-Tracer PET Synthesis from Multi-Sequence MRI

TL;DR

RelA-Diffusion is proposed, a Relativistic Adversarial Diffusion framework for multi-tracer PET synthesis from multi-sequence MRI that captures richer structural information to guide image generation and introduces a gradient-penalized relativistic adversarial loss to the intermediate clean predictions of the diffusion model.

Abstract

Multi-tracer positron emission tomography (PET) provides critical insights into diverse neuropathological processes such as tau accumulation, neuroinflammation, and -amyloid deposition in the brain, making it indispensable for comprehensive neurological assessment. However, routine acquisition of multi-tracer PET is limited by high costs, radiation exposure, and restricted tracer availability. Recent efforts have explored deep learning approaches for synthesizing PET images from structural MRI. While some methods rely solely on T1-weighted MRI, others incorporate additional sequences such as T2-FLAIR to improve pathological sensitivity. However, existing methods often struggle to capture fine-grained anatomical and pathological details, resulting in artifacts and unrealistic outputs. To this end, we propose RelA-Diffusion, a Relativistic Adversarial Diffusion framework for multi-tracer PET synthesis from multi-sequence MRI. By leveraging both T1-weighted and T2-FLAIR scans as complementary inputs, RelA-Diffusion captures richer structural information to guide image generation. To improve synthesis fidelity, we introduce a gradient-penalized relativistic adversarial loss to the intermediate clean predictions of the diffusion model. This loss compares real and generated images in a relative manner, encouraging the synthesis of more realistic local structures. Both the relativistic formulation and the gradient penalty contribute to stabilizing the training, while adversarial feedback at each diffusion timestep enables consistent refinement throughout the generation process. Extensive experiments on two datasets demonstrate that RelA-Diffusion outperforms existing methods in both visual fidelity and quantitative metrics, highlighting its potential for accurate synthesis of multi-tracer PET.
Paper Structure (29 sections, 10 equations, 5 figures, 4 tables, 2 algorithms)

This paper contains 29 sections, 10 equations, 5 figures, 4 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. GAN-based PET Image Synthesis
  4. Diffusion Models for Medical Imaging
  5. Proposed Method
  6. Overall Framework
  7. Forward Diffusion Process
  8. Reverse Denoising Process
  9. Supervision on Intermediate Clean Predictions
  10. Image Constraint
  11. Gradient-Penalized Relativistic Adversarial Constraint
  12. Overall Training Objective
  13. Implementation
  14. Experiments
  15. Materials and Experimental Setup
  16. ...and 14 more sections

Figures (5)

  • Figure 1: Illustration of the proposed relativistic adversarial diffusion (RelA-Diffusion) framework that synthesizes multi-tracer PET images from multi-sequence MRI input such as T1-weighted (T1w) and T2-FLAIR (T2F) MRIs.
  • Figure 2: Visualization of test PET images for (a) TAU, (b) PBR, and (c) PIB, synthesized by 8 methods, along with corresponding difference (Diff) maps. Ground-truth (GT) and input T1w/T2F MRIs with subject IDs are shown at the bottom.
  • Figure 3: Visualization of test ADNI TAU-PET images synthesized by 8 methods, and difference (Diff) maps. The ground-truth (GT) PET images along with the input T1w and T2F MRIs are displayed at the bottom with the corresponding subject IDs.
  • Figure 4: Region-level quantitative evaluation on NFL-LONG. (a) Scatter plots of predicted vs. ground-truth volume-of-interest (VOI) mean SUVr for TAU-PET across eight methods and five tau-related VOIs (Frontal, Mesial Temporal, Meta Temporal, Temporo-Parietal, and Universal) villemagne2023centaur, with MAE and RMSE reported in each panel. (b) Comparison of SSIM results across the same five VOIs villemagne2023centaur.
  • Figure 5: Downstream regression performance for age, MMSE, and MoCA prediction using synthesized PET (syn-PET) generated from ADNI TAU images together with MRI, in comparison with MRI-only inputs.