Human-like automated radiotherapy treatment planning via imitation learning

通过模仿学习制定类似人类的自动放射治疗计划

基本信息

批准号：
10610971
负责人：
Xun Jia
金额：
$ 60.6万
依托单位：
UT SOUTHWESTERN MEDICAL CENTER
依托单位国家：
美国
项目类别：
财政年份：
2021
资助国家：
美国
起止时间：
2021-05-18 至 2026-04-30
项目状态：
未结题

来源：
https://reporter.nih.gov/project-details/10610971
关键词：
Anatomy Area Artificial Intelligence Attention Back Beds Behavior Cancer Patient Caring Clinical Clinical Research Communities Complex Complication Data Decision Making Deterioration Development Dose Ensure Environment Evaluation Evolution Failure Feedback Goals Growth Head and Neck Cancer Human Intention Joints Learning Malignant neoplasm of prostate Manuals Mathematics Measures Medical Medical center Mind Modality Modeling Modernization Normal tissue morphology Operative Surgical Procedures Outcome Patient-Focused Outcomes Patients Performance Physicians Play Probability Process Radiation therapy Research Resource-limited setting Resources Role Scheme Site Structure System Testing Time Training Translating Translations Treatment Step Treatment outcome Validation algorithm development cancer therapy chemotherapy clinical translation design experience head and neck cancer patient improved individual patient industry partner innovation learning algorithm learning progression model development negative affect new technology optimal treatments population based product development prototype response routine practice success treatment planning treatment strategy tumor validation studies

项目摘要

PROJECT SUMMARY Radiation therapy is one of the major approaches for cancer treatment. Treatment planning, the process of designing the optimal treatment plan for each patient, is one of the most critical steps. If a treatment is poorly designed, a satisfactory outcome cannot be achieved, regardless of the quality of other treatment steps. Treatment planning in modern radiotherapy is formulated as a mathematical optimization problem defined by a set of hyperparameters. While there exists several quantifiable metrics to quantify plan quality and guide the planning process, these are simplified representations that cannot fully describe the physician’s intent. In addition, these metrics only measure plan quality from a population-based perspective, and cannot guide treatment planning to achieve the patient-specific best treatment plans. Hence, the best physician-preferred solution often sits in a gray area, only achievable by an extensive trial-and-error hyperparameter tuning process and interactions between the planner and physician. Consequently, planning time can take up to a week for complex cases and plan quality may be poor, if the planner is inexperienced and/or under heavy time constraints. These consequences substantially deteriorate treatment outcomes, as having been clearly demonstrated in clinical studies. Recently, the advancement in artificial intelligence (AI), particularly in imitation learning allows human- like decision making by observing a human expert’s actions and internally building its own decision-making system. In response to PAR-18-530, the goal of this project is to develop and translate an AI planner that mimics human experts’ behavior to generate a high quality plan. The AI planner will not replace human planners. Instead, the AI plan will be used as a starting point in the current planning process to improve plan quality and planning efficiency. The human planner’s actions on further plan improvement can feed back to the AI planner through continuous learning for its continuous evolution. We will pursue this goal using prostate cancer as the test bed through an academic-industrial partnership, jointing strong research and clinical expertise at UT Southwestern Medical Center with extensive commercial product development experience at Varian Medical Systems Inc. The following specific aims are defined. Aim 1: Model and algorithm development. We will collect experts’ behavior data in routine treatment planning and train the AI planner. Aim 2: System validation and translation. We will integrate the AI planner into Varian Eclipse treatment planning system and validate the system in a clinically realistic setting. The innovations include the use of a state-of-the-art AI imitation learning algorithm to solve a clinically important problem, the novel technological capabilities enabled by the developed system, as well as coherent translation activities to deliver new capabilities to end users. Deliverability is ensured by extensive preliminary studies and the partnership integrating complementary expertise and resources. Clinical translation of the AI planner will bring substantial impacts to radiotherapy by providing high-quality and efficient treatment planning to benefit patients, especially those in resource-limited regions.

项目摘要放射治疗是癌症治疗的主要方法之一。治疗计划，过程为每个患者设计最佳治疗计划是最关键的步骤之一。如果治疗差设计，无论其他治疗步骤的质量如何，都无法实现令人满意的结果。现代放疗中的治疗计划被提出为数学优化问题一组超参数。虽然存在几个可量化的指标来量化计划质量并指导计划过程，这些是简化的表示，无法完全描述物理学的意图。此外，这些指标仅从基于人群的角度衡量计划质量，并且无法指导治疗计划实现特定于患者的最佳治疗计划。因此，经常是最好的物理优先解决方案坐在灰色区域，只能通过广泛的试用超级参数调谐过程才能实现计划者与物理之间的相互作用。因此，计划时间最多需要一周的复杂时间如果规划师没有经验和/或在很大的时间限制下，案例和计划质量可能很差。这些后果基本上是侦探治疗结果，正如临床中清楚地证明的那样研究。最近，人工智能（AI）的进步，特别是在模仿学习中，可以使人类就像通过观察人类专家的行动并在内部建立自己的决策来做出决策系统。为了响应Par-18-530，该项目的目标是开发和翻译一个模仿的AI计划者人类专家的行为是为了产生高质量计划。 AI计划者将不会取代人类规划者。反而， AI计划将用作当前计划过程的起点，以提高计划质量和计划效率。人类规划师对进一步计划改进的行动可以通过持续学习的持续发展。我们将使用前列腺癌作为测试床实现这一目标通过学术工业合作伙伴关系，在UT Southwestern结合强大的研究和临床专业知识在Varian Medical Systems Inc.拥有丰富商业产品开发经验的医疗中心。定义了以下特定目标。目标1：模型和算法开发。我们将收集专家的行为常规治疗计划中的数据并培训AI计划者。目标2：系统验证和翻译。我们将将AI计划者集成到Varian Eclipse治疗计划系统中，并在诊所中验证该系统现实设置。创新包括使用最先进的AI模仿学习算法来解决A 临床上重要的问题，开发系统实现了新颖的技术功能，以及连贯的翻译活动为最终用户提供新功能。通过广泛确保可交付性初步研究和合作伙伴关系整合完长的专业知识和资源。临床翻译 AI规划师将通过提供高质量和有效的治疗来对放射疗法产生重大影响计划使患者受益，尤其是在资源有限地区的患者。

项目成果

期刊论文数量（10）

专著数量（0）

科研奖励数量（0）

会议论文数量（0）

专利数量（0）

Advances in Automated Treatment Planning.

DOI：
10.1016/j.semradonc.2022.06.004
发表时间：
2022-10
期刊：
Seminars in radiation oncology
影响因子：
3.5
作者：
D. Nguyen;Mu-Han Lin;D. Sher;Wei Lu;X. Jia;Steve B Jiang
通讯作者：
D. Nguyen;Mu-Han Lin;D. Sher;Wei Lu;X. Jia;Steve B Jiang

Site-agnostic 3D dose distribution prediction with deep learning neural networks.

DOI：
10.1002/mp.15461
发表时间：
2022-03
期刊：
Medical physics
影响因子：
3.8
作者：
通讯作者：

Single patient learning for adaptive radiotherapy dose prediction.

单个患者学习自适应放疗剂量预测。

DOI：
10.1002/mp.16799
发表时间：
2023
期刊：
Medical physics
影响因子：
3.8
作者：
Maniscalco,Austen;Liang,Xiao;Lin,Mu-Han;Jiang,Steve;Nguyen,Dan
通讯作者：
Nguyen,Dan

Towards a safe and efficient clinical implementation of machine learning in radiation oncology by exploring model interpretability, explainability and data-model dependency.

DOI：
10.1088/1361-6560/ac678a
发表时间：
2022-05-27
期刊：
Physics in medicine and biology
影响因子：
3.5
作者：
Barragán-Montero A;Bibal A;Dastarac MH;Draguet C;Valdés G;Nguyen D;Willems S;Vandewinckele L;Holmström M;Löfman F;Souris K;Sterpin E;Lee JA
通讯作者：
Lee JA

Intentional deep overfit learning for patient-specific dose predictions in adaptive radiotherapy.