arXiv cs.AI· Riqiang Gao, Simon Arberet, Martin Kraus, Han Liu, Wilko F. A. R. Verbakel, Dorin Comaniciu, Florin-Cristian Ghesu, Ali Kamen·· 4 小时前AI 评分32
生成式 AI 放疗计划演示：按用户偏好预测 3D 剂量分布
Demo: Generative AI helps Radiotherapy Planning with User Preference
AI 导读
研究者提出一种生成式模型，仅依据用户自定义偏好即可预测放疗 3D 剂量分布，无需以参考计划作为训练真值。该模型允许计划者权衡危及器官（OAR）与计划靶区（PTV）的优先级，可与临床治疗计划系统集成。对比评估显示，在部分场景下该方法在适应性和计划质量上优于 Varian RapidPlan 模型，论文获 NeurIPS 2025 GenAI4Health 最佳论文。
正文 · 原文
Abstract:Radiotherapy planning is a highly complex process that often varies significantly across institutions and individual planners. Most existing deep learning approaches for 3D dose prediction rely on reference plans as ground truth during training, which can inadvertently bias models toward specific planning styles or institutional preferences. In this study, we introduce a novel generative model that predicts 3D dose distributions based solely on user-defined preference flavors. These customizable preferences enable planners to prioritize specific trade-offs between organs-at-risk (OARs) and planning target volumes (PTVs), offering greater flexibility and personalization. Designed for seamless integration with clinical treatment planning systems, our approach assists users in generating high-quality plans efficiently. Comparative evaluations demonstrate that our method can surpasses the Varian RapidPlan model in both adaptability and plan quality in some scenarios.
| Comments: | Best paper in GenAI4Health at NeurIPS 2025 |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2512.08996 [cs.CV] |
| (or arXiv:2512.08996v3 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2512.08996 arXiv-issued DOI via DataCite |
Submission history
From: Riqiang Gao [view email]
[v1]
Mon, 8 Dec 2025 16:49:21 UTC (4,698 KB)
[v2]
Thu, 24 Sep 2026 19:29:47 UTC (4,698 KB)
[v3]
Fri, 23 Oct 2026 15:42:29 UTC (4,681 KB)
来源：arXiv cs.AI · arxiv.org