最近在自动放射治疗计划中的工作已经使用基于历史治疗计划的机器学习来直接从计划图像推断新患者的空间剂量分布。我们提出了一种基于图谱的方法，该方法在训练数据库中学习每个患者的剂量预测模型（图谱），然后学习将新患者与最相关的图谱相匹配。该方法创建空间剂量目标，其指定期望的每体素剂量，并且因此取代指定剂量体积目标的任何要求，以传达治疗计划的目标。概率剂量分布从最相关的图谱中推断，并使用条件随机场进行定标，以确定最可能的剂量空间分布，从而为相关的感兴趣区域产生特定的剂量先验（直方图）。然后，基于体素的剂量模拟将预测的剂量分布转换为可递送的治疗计划剂量分布。在这项研究中，我们研究了接受IMRT和VMAT治疗的右侧orbiliaynx头颈部患者的自动计划。我们使用54个训练和12个独立测试患者的数据库比较了我们的剂量预测管道的四个版本。我们的初步结果是有希望的;与临床相比，自动计划在7例患者中实现了更多的剂量评价标准，在4例患者中实现了相同数量的剂量评价标准。总体而言，自动化计划达到的标准相对数量高于临床（17 vs 8），自动化计划显示出对风险器官的保留增加（52 vs 44）和更好的靶覆盖/均匀性（41 vs 31）。
英文摘要：
Recent works in automated radiotherapy treatment planning have used machine learning based on historical treatment plans to infer the spatial dose distribution for a novel patient directly from the planning image. We present an atlas-based approach which learns a dose prediction model for each patient (atlas) in a training database, and then learns to match novel patients to the most relevant atlases. The method creates a spatial dose objective, which specifies the desired dose-per-voxel, and therefore replaces any requirement for specifying dose-volume objectives for conveying the goals of treatment planning. A probabilistic dose distribution is inferred from the most relevant atlases, and is scalarized using a conditional random field to determine the most likely spatial distribution of dose to yield a specific dose prior (histogram) for relevant regions of interest. Voxel-based dose mimicking then converts the predicted dose distribution to a deliverable treatment plan dose distribution. In this study, we investigated automated planning for right-sided oropharaynx head and neck patients treated with IMRT and VMAT. We compare four versions of our dose prediction pipeline using a database of 54 training and 12 independent testing patients. Our preliminary results are promising; automated planning achieved a higher number of dose evaluation criteria in 7 patients and an equal number in 4 patients compared with clinical. Overall, the relative number of criteria achieved was higher for automated planning versus clinical (17 vs 8) and automated planning demonstrated increased sparing for organs at risk (52 vs 44) and better target coverage/uniformity (41 vs 31).
Divergent Clinical Equivalence Findings from DVH and NTCP Metrics for Alternative OAR Delineations with Increasing Setup Variability in Head-and-Neck Radiotherapy
Causal Machine Learning Analysis of Empirical Relative Biological Effectiveness (RBE) for Mandible Osteoradionecrosis in Head and Neck Cancer Radiotherapy
Quantification of head and neck cancer patients' anatomical changes during radiotherapy: prediction of replanning need
McGill University（麦吉尔大学）；Research Institute of the McGill University Health Centre, McGill University（麦吉尔大学健康中心研究所，麦吉尔大学）；Department of Radiation Oncology, Hôpital Maisonneuve-Rosemont（Maisonneuve-Rosemont医院放射肿瘤科）；Department of Radiation Oncology, McGill University（麦吉尔大学放射肿瘤科）
TransAnaNet: Transformer-based Anatomy Change Prediction Network for Head and Neck Cancer Patient Radiotherapy
UT Southwestern Medical Center（德克萨斯大学西南医学中心）；University of Maryland Medical Center（马里兰大学医学中心）；Central Arkansas Radiation Therapy Institute（中央阿肯色放射治疗研究所）
Deep learning to achieve clinically applicable segmentation of head and neck anatomy for radiotherapy
University College London Hospitals NHS Foundation Trust（伦敦大学学院医院国民保健服务基金会信托）；Google Health（谷歌健康）；DeepMind（深度思维）；Clatterbridge Cancer Centre NHS Foundation Trust（克拉特布里奇癌症中心国民保健服务基金会信托）；University College London（伦敦大学学院）
Three-Dimensional Radiotherapy Dose Prediction on Head and Neck Cancer Patients with a Hierarchically Densely Connected U-net Deep Learning Architecture
University of Texas Southwestern Medical Center（德克萨斯大学西南医学中心）
Magnetic resonance imaging assessment of the suitability and consistency of radiotherapy treatment positioning achieved using intra-oral stents
Royal Brisbane and Women's Hospital（皇家布里斯班和妇女医院）；University of Queensland（昆士兰大学）；Queensland University of Technology（昆士兰科技大学）；Herston Biofabrication Institute, Metro North Hospital and Health Service（赫斯顿生物制造研究所，北都会区医院与健康服务）
调强放射治疗（IMRT）计划的最大挑战之一是在合理的计算时间内优化射束角度。为了证明先验知识对提高调强放射治疗中射束角度优化（BAO）效率的价值，本研究提出了一种改进的遗传算法（GA），将先验知识引入进化过程。在所谓的nowGABAO（知识引导GA for BAO）方法中，两种类型的先验知识被纳入GA的一般流程图中：（1）光束角度约束，定义不允许光束通过的角度范围，以及（2）计划模板，定义适合所研究案例的最可能的光束配置，并用于指导GA的进化进程。通过对头颈部和肺部两个复杂临床病例的优化结果表明，适当定义角度约束和规划模板可以明显提高优化速度。此外，一个相当坏的模板产生的影响可以忽略不计的优化效率的下降。总之，先验知识有助于提高波束角优化效率，并且可以通过GA适当地并入。它可以抵抗包含坏模板。
Comprehensive and Clinically Accurate Head and Neck Organs at Risk Delineation via Stratified Deep Learning: A Large-scale Multi-Institutional Study
PAII Inc.（PAII公司）；The First Affiliated Hospital Zhejiang University（浙江大学附属第一医院）；Ping An Technology（平安科技）；Chang Gung Memorial Hospital（长庚纪念医院）；Johns Hopkins University（约翰斯·霍普金斯大学）；Huadong Hospital Affiliated to Fudan University（复旦大学附属华东医院）；First Affiliated Hospital of Xi'an Jiaotong University（西安交通大学第一附属医院）；People's Hospital of Shanxi Province（山西省人民医院）；Nanfang Hospital, Southern Medical University（南方医科大学南方医院）；The First Hospital of Lanzhou University（兰州大学第一医院）；Chang Gung University（长庚大学）
Generating synthetic computed tomography for radiotherapy: SynthRAD2025 challenge report
Radiation Physics, Department of Hematology, Oncology；Radiation Physics, Skåne University Hospital, Lund, Sweden；Medical Radiation Physics, Department of Clinical Sciences Lund, Lund University, Lund, Sweden；Radiotherapy Department, University Medical Center Utrecht, Utrecht, The Netherlands；Computational Imaging Group for MR Diagnostics \& Therapy, University Medical Center Utrecht, Utrecht, The Netherlands；Division of Medical Physics in Radiation Oncology, Deutsches Krebsforschungszentrum (DKFZ), Heidelberg, Germany；Heidelberg Institute for Radiation Oncology (HIRO)；National Center for Radiation Research in Oncology (NCRO), Heidelberg, Germany；Department of Radiation Oncology；Cyberknife Center, University Hospital of Cologne, Cologne, Germany；Department of Radiation Oncology, Radboud University Medical Center, Nijmegen, The Netherlands；Department of Radiation Oncology, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands；Department of Radiation Oncology, LMU University Hospital, LMU Munich, Munich, Germany；Bavarian Cancer Research Center (BZKF), Munich, Germany；Australian eHealth Research Center, CSIRO, Brisbane, Australia；School of Computer Science, University of Sydney, Sydney, Australia；Department of Orthodontics；Pattern Recognition Lab, FAU Erlangen-Nuremberg, Germany；RIKEN Center for Integrative Medical Sciences, Tokyo, Japan；Erasmus Mundus Joint Master's Degree IPCVai, University of Bordeaux, France；Computer Science, Huazhong University of Science；Huazhong University of Science；Canon Medical Systems (China) CO., LTD., Beijing, China；Department of Radiation Oncology, Inselspital, Bern University Hospital；University of Bern, Bern, Switzerland
OpenKBP-Opt: An international and reproducible evaluation of 76 knowledge-based planning pipelines
我们建立了一个开放的框架，开发基于知识的计划（KBP）在放射治疗计划优化模型。我们的框架包括100名头颈癌患者的参考计划和OpenKBP Grand Challenge期间由不同研究小组开发的19个KBP模型的高质量剂量预测。将剂量预测输入到四个优化模型中，以形成76个独特的KBP管道，生成7600个计划。通过以下方式将预测和计划与参考计划进行比较：剂量评分，这是模型达到的剂量的平均绝对体素差异;剂量体积直方图（DVH）标准的偏差;以及临床计划标准满意度的频率。我们还进行了理论研究，以证明我们的剂量模拟模型。预测与其KBP管道之间的剂量分数的等级顺序相关性范围为0.50至0.62，这表明预测的质量通常与计划的质量呈正相关。此外，与输入预测相比，KBP生成的计划在23个DVH标准中的18个标准上的表现明显更好（P<0.05;单侧Wilcoxon检验）。类似地，每个优化模型生成的计划满足比参考计划更高百分比的标准。最后，我们的理论研究表明，剂量模拟模型生成的计划，也是最佳的常规规划模型。这是迄今为止国际上对KBP预测和优化模型组合进行评估的最大努力。为了保证可重复性，我们的数据和代码可在https://github.com/ababier/open-kbp-opt上免费获得。
