用于放疗中危及器官分割质量保证的图像条件扩散模型
Image-Conditioned Diffusion Models for Quality Assurance of Organ-at-Risk Segmentations in Radiotherapy
浏览论文内容
中文总结 AI 辅助
该研究针对放疗中危及器官分割审阅耗时主观的问题，对比VAE框架与图像条件分割扩散模型，发现后者能更一致地定位细微边界错误，为分割质量保证提供了有前景的框架。
中文摘要 AI 辅助
准确的危及器官分割对放疗计划至关重要，但分割结果的审阅耗时且主观。我们针对头颈部CT的分割错误检测研究了规范建模，比较了VAE框架与图像条件分割扩散模型。使用RADCURE的脑干和脊髓分割，结合模拟的边界和宽度扰动评估模型，通过输入与重建分割间的戴斯相似系数（Dice similarity coefficient）和一致距离（Distance to Agreement, DTA）评估错误检测。尽管两种模型都检测到部分模拟错误，区域DTA显示扩散模型能更一致地定位细微边界错误。这些结果表明，图像条件扩散重建是一种有前景的、可定位且感知解剖结构的分割质量保证框架。
英文摘要
Accurate organ-at-risk segmentation is essential for radiotherapy planning, but reviewing segmentations is time-consuming and subjective. We investigate normative modelling for segmentation error detection in head-and-neck CT, comparing a VAE framework with an image-conditioned segmentation diffusion model. Models were evaluated on RADCURE brainstem and spinal cord segmentations using simulated boundary and width perturbations. Error detection was assessed using the Dice similarity coefficient and the Distance to Agreement (DTA) between the input and reconstructed segmentations. While both models detected some simulated errors, regional DTA showed that the diffusion model localised subtle boundary errors more consistently. These results support image-conditioned diffusion reconstruction as a promising framework for localised, anatomy-aware segmentation QA.
发表机构
- UCL Hawkes Institute（UCL霍克斯研究所）
- University College London（伦敦大学学院）
- National Physical Laboratory（英国国家物理实验室）
- University College London Hospital（伦敦大学学院医院）
- National Radiotherapy Trials Quality Assurance Group（国家放射治疗试验质量保证组）
- Mount Vernon Hospital（芒特弗农医院）
- University of Manchester（曼彻斯特大学）
机构由 AI 辅助整理，请以论文原文为准。
