Generation of Heterogeneous PET Images from Uniform Organ Activity Maps Using a Pretrained Domain-Adapted Diffusion Model

arXiv:2605.20267 · 2026-05-21 공개 · arXiv · PDF

diffusion-models domain-adaptation tumor-segmentation radiomic-analysis xcat-phantoms clinical-imaging image-realism pet-image-synthesis

Abstract

Synthetic PET images are valuable for quantitative imaging workflow development, scalable virtual imaging trials, and deep learning model training, but conventional physics-based simulation approaches are computationally intensive, limited in anatomical variability, and often fail to capture heterogeneous PET uptake. This study developed a pretrained domain-adapted diffusion (PAD) model for anatomy-conditioned PET synthesis from uniform organ activity maps. PAD adopts a natural-image pretrained text-to-image decoder with an upstream conditioning encoder and a downstream PET-domain adapter. A two-phase training strategy was used, with the first phase learning coarse uptake distributions and the second refining local image details. Uniform organ activity maps were generated from CT-based segmentations by assigning each organ its mean uptake from the paired PET image. Evaluation included quantitative accuracy, noise assessment, radiomic analysis, tumor segmentation performance, and a human observer study. PAD-generated images achieved high quantitative accuracy, with concordance correlation coefficients above 0.92 between organ mean SUVs and assigned activity values. The synthesized images showed noise levels and texture characteristics similar to target PET images and produced comparable tumor segmentation performance. In a two-alternative forced-choice observer study, four readers achieved approximately 50% accuracy, indicating visual indistinguishability between synthesized and target images. PAD also generated realistic PET images from XCAT-derived activity maps, demonstrating compatibility with phantom-based anatomical priors. Overall, PAD provides a diffusion-based framework for generating clinically relevant heterogeneous PET images from uniform organ activity maps derived from clinical segmentations or digital phantoms, supporting data augmentation and downstream imaging studies.

한국어 요약

📋 한 줄 요약

**[의료 영상 합성 / 확산 모델]** 자연 이미지 사전학습 text-to-image 디코더에 PET 도메인 어댑터를 부착한 PAD 모델로 균일 장기 활성 맵에서 이질적 PET 영상을 합성.

🎯 핵심 기여도

💡 핵심 아이디어

대규모 자연 이미지로 사전학습된 확산 모델의 풍부한 시각 prior를 활용하면, PET처럼 데이터가 제한된 의료 도메인에서도 도메인 어댑터만으로 사실적인 합성이 가능하다는 transfer learning 관점.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: PET 데이터 부족 환경에서 가상 임상시험·DL 학습 데이터·정량 영상 워크플로 검증에 사용 가능한 확산 기반 합성 프레임워크 제시. **한계**: 입력 활성 맵의 품질(분할 정확도)에 의존, 매우 희귀한 흡수 패턴에 대한 일반화는 추가 검증 필요.

🚀 실용적 활용