A Fully Decentralized Federated Learning Framework for Automated Image Segmentation in Cancer Radiotherapy

用于癌症放射治疗自动图像分割的完全去中心化联合学习框架

• 批准号: 10303437
• 负责人: Yading Yuan
• 金额: $ 21.06万
• 依托单位: ICAHN SCHOOL OF MEDICINE AT MOUNT SINAI
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-09-01 至 2023-03-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10303437
• 关键词: Address; Adoption; Anatomy; Appearance; Area; Artificial Intelligence; Client; Clinical; Collaborations; Communities; Data; Data Aggregation; Data Pooling; Data Set; Data Sources; Databases; Decentralization; Development; Environment; Ethics; Goals; Health system; Healthcare Systems; Hospitals; Image; Image Analysis; Institution; Learning; Legal; Liver; Location; Machine Learning; Medical; Medical Imaging; Medical center; Medicine; Methods; Modeling; National Institute of Biomedical Imaging and Bioengineering; Organ; Ownership; Participant; Patients; Performance; Privacy; Process; Radiation Oncology; Radiation therapy; Research Priority; Resources; Risk; Scheme; Shapes; Software Tools; Source; Techniques; Time; Training; Validation; X-Ray Computed Tomography; automated segmentation; base; cancer radiation therapy; clinical practice; data sharing; deep learning; deep learning algorithm; flexibility; head and neck cancer patient; imaging Segmentation; improved; interoperability; liver imaging; metropolitan; novel; open source; patient privacy; peer; precision medicine; soft tissue; software infrastructure; success; task analysis; treatment planning; tumor

PROJECT SUMMARY While the recent surge of artificial intelligence (AI) has made remarkable progress in various image analysis tasks, their performance in a broad range of clinical environment is largely restricted by the limited generalization capability when being applied to new data, primarily because most models have been generated using data from a single institution or public datasets with limited training data. Aggregating data from different institutions could improve model training, but such centralized data sharing is practically challenging due to various technical, legal, privacy and data ownership barriers. This proposal aims to address these barriers by developing a novel gossip federated learning (GFL) framework to build an effective AI model by learning from different data sources without the need of sharing patient data. As compared to the traditional client/server federated learning such as FedAvg, the proposed framework is fully decentralized in that the models trained in local datasets will directly communicate to each other in a peer-to-peer manner, making our method more robust and efficient. We will develop and evaluate the proposed scheme in the task of automated organ segmentation in CT images for liver and head and neck (H&N) cancer patients treated with radiation therapy (RT) because accurate, robust and efficient delineation of those organs at risk (OARs) is a clinically important but technically challenging problem. We hypothesize that the model trained with our framework can achieve segmentation performance not inferior to a model with data pooled from all the resources. The dynamics of our recently created healthcare system mimic a diverse multi-institutional environment, which places us in an ideal setting to systematically evaluate our framework. Our specific aims include: 1) Establish the GFL-based automated OAR segmentation framework, and develop the supporting software infrastructure; 2) Optimize the GFL-based auto- segmentation; 3) Evaluate GFL-based OAR segmentation framework with 400 liver and 400 H&N cancer patients collected from four hospitals within a metropolitan health system. This proposal addresses two key research priorities for NIBIB: machine learning based segmentation and approaches that facilitate interoperability among annotations used in image training databases. The success of this project will substantially increase the number and variety of data for model training without sacrificing the patient privacy, and thus improve the performance and generalization of the segmentation model on new data. We will open-source this framework, which may enable a larger scale of multi-institutional collaboration and could expedite the clinical adoption of AI-driven auto- segmentation in RT. More importantly, this framework provides a flexible and robust solution to the primary barrier of applying AI to the medical domain where learning on multi-institutional data sharing is impeded by patient privacy concerns, and is expected to have a catalytic impact on precision medicine by generalizing it to broader applications within medicine where a model needs to learn across multi-institutional data without sacrificing patient privacy.

项目摘要 虽然最近人工智能 (AI) 的蓬勃发展在以下方面取得了显着进展： 各种图像分析任务，它们在广泛的临床环境中的性能在很大程度上受到限制 应用于新数据时泛化能力有限，主要是因为大多数模型已经被 使用来自单个机构的数据或训练数据有限的公共数据集生成。聚合数据来自 不同的机构可以改进模型训练，但这种集中的数据共享实际上具有挑战性 由于各种技术、法律、隐私和数据所有权障碍。该提案旨在解决这些障碍 通过开发一种新颖的八卦联邦学习（GFL）框架，通过学习来构建有效的人工智能模型 不同的数据源，无需共享患者数据。与传统的客户端/服务器相比 联邦学习，例如 FedAvg，所提出的框架是完全去中心化的，因为训练的模型 本地数据集将以点对点的方式直接相互通信，使我们的方法更加稳健 和高效。我们将在自动器官分割任务中开发和评估所提出的方案 在接受放射治疗 (RT) 治疗的肝癌和头颈 (H&N) 癌症患者的 CT 图像中，因为 准确、稳健和有效地描绘那些处于危险中的器官（OAR）在临床上很重要，但在技术上 具有挑战性的问题。我们假设用我们的框架训练的模型可以实现分割 性能不亚于从所有资源汇集数据的模型。我们最近创建的动态 医疗保健系统模仿多元化的多机构环境，这使我们处于理想的环境中 系统地评估我们的框架。我们的具体目标包括：1）建立基于GFL的自动化OAR 细分框架，开发配套软件基础设施； 2）优化基于GFL的自动 分割; 3) 使用 400 名肝癌患者和 400 名 H&N 癌症患者评估基于 GFL 的 OAR 分割框架 从大都市卫生系统内的四家医院收集。该提案涉及两项关键研究 NIBIB 的优先事项：基于机器学习的细分和促进互操作性的方法 图像训练数据库中使用的注释。该项目的成功将大幅增加 和各种数据用于模型训练而不牺牲患者隐私，从而提高性能 以及新数据上分割模型的推广。我们将开源这个框架，这可能 实现更大规模的多机构协作，并可以加快人工智能驱动的自动驾驶技术的临床采用 RT 中的分割。更重要的是，该框架为主要问题提供了灵活而强大的解决方案 将人工智能应用到医疗领域的障碍是多机构数据共享的学习受到阻碍 患者隐私问题，并有望通过将其推广到对精准医疗产生催化影响 医学领域更广泛的应用，其中模型需要跨多个机构的数据进行学习，而无需 牺牲患者隐私。