PREcision Care In Cardiac ArrEst - ICECAP (PRECICECAP)

心脏骤停的精准护理 - ICECAP (PRECICECAP)

• 批准号: 10412861
• 负责人: Jonathan Elmer
• 金额: $ 32.9万
• 依托单位: STANFORD UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-12-15 至 2022-11-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10412861
• 关键词: Acute Brain Injuries; Address; Architecture; Artificial Intelligence; Attention; Automation; Biomedical Research; Brain; Brain imaging; Cardiopulmonary; Cardiopulmonary Resuscitation; Caring; Cessation of life; Clinical; Collaborations; Common Data Element; Communities; Complex; Computer software; Critical Illness; Custom; Data; Data Aggregation; Data Scientist; Data Set; EEG-based imaging; FAIR principles; Funding; Goals; Heart; Heart Arrest; Human; Individual; Industry; Intensive Care; Knowledge; Label; Libraries; Lung; Machine Learning; Manuals; Metadata; Mission; Modern Medicine; Modernization; Monitor; Morbidity - disease rate; Neurologic; Neurosciences Research; Ontology; Organ; Outcome; Parents; Patients; Physiological; Public Health; Readiness; Recovery; Recovery of Function; Reproducibility; Research; Research Personnel; Resolution; Scientist; Series; Signal Transduction; Site; Software Tools; Source; Standardization; System; Techniques; Terminology; Testing; Time; United States National Institutes of Health; Visualization; Work; application programming interface; base; biomarker signature; complex data; cost; dashboard; data archive; data ecosystem; data harmonization; data pipeline; data visualization; design; disability; feature extraction; flexibility; graphical user interface; improved; innovation; mortality; natural hypothermia; novel; novel marker; patient subsets; personalized care; precision medicine; prospective; quality assurance; software development; supervised learning; tool; unsupervised learning; user-friendly

Project Summary: The goals of the parent PREcision Care In Cardiac ArrEst - ICECAP (PRECICECAP) study are to discover novel biomarker signatures after cardiac arrest that predict treatment responsiveness and long-term recovery. Cardiac arrest is a major public health problem with high morbidity and mortality. Improving survival and functional recovery are critical public health objectives. We hypothesize that not all patients are identical and that through innovative, multi-parametric data- driven approaches we will be able to identify novel signatures to identify distinct patient subgroups. The PRECICECAP analysis plan allows us to meet our study needs but is not intrinsically generalizable or usable by others. An ever-growing number of NIH-funded projects amass and analyze related datasets from hospitalized patients and develop their own custom solutions. Such an ad hoc approach is costly, inefficient, and threatens research rigor and reproducibility. The objective of this supplement to PRECICECAP is to develop a freely-available software platform that allows artificial intelligence/machine learning (AI/ML) analysis of complex neurocritical care data and to provide a curated dataset from PRECICECAP ready for AI/ML. The knowledge learned here will be applied to help achieve NIH goals for modernizing the biomedical research data ecosystem by developing a software product that can handle AI/ML on this type of complex neurocritical care data. It will also allow the sharing of a cleaned, annotated comprehensive data set. Through a thoughtful collaboration between clinician investigators, data scientists and industry, we will take NIH-supported data from the PRECICECAP study and make it broadly available and easily usable. The project will deliver an important software tool that can be used by others conducting similar research, advancing the NIH’s mission to make complex data FAIR (Findable, Accessible, Interoperable, and Reusable). We will develop a series of modular functions (for example, to support data harmonization, annotation or visualization) that permit users to graphically construct processing pipelines maximizing automation where appropriate and allowing facile interaction with data when needed. We will develop a user-friendly dashboard interface to allow individual study sites and coordinating hubs to understand complex data, inspect key meta-data features, and identify potential errors. Modular design facilitates dynamic configuration of AI/ML architectures within the same interface, allowing individual modules to combine synergistically to maximize efficiency and reproducibility. The result will facilitate an open, wide collaboration between scientists using similar data.

项目概要： 家长 PREcision Care In Cardiac ArrEst - ICECAP (PRECICECAP) 研究的目标是 发现心脏骤停后预测治疗反应的新生物标志物特征 长期康复。心脏骤停是一个发病率高且严重的重大公共卫生问题 死亡。改善生存和功能恢复是重要的公共卫生目标。我们 假设并非所有患者都是相同的，并且通过创新的多参数数据- 通过驱动方法，我们将能够识别新的特征来识别不同的患者亚组。 PRECICECAP 分析计划可以让我们满足我们的研究需求，但本质上并不是 可推广或可供他人使用。 NIH 资助的项目数量不断增加 分析住院患者的相关数据集并开发自己的定制解决方案。这样的 临时方法成本高昂、效率低下，并且威胁研究的严谨性和可重复性。这 PRECICECAP 补充的目的是开发一个免费的软件平台， 允许对复杂的神经重症监护数据进行人工智能/机器学习 (AI/ML) 分析 并提供来自 PRECICECAP 的精选数据集，为 AI/ML 做好准备。在这里学到的知识 将用于帮助实现 NIH 生物医学研究数据生态系统现代化的目标 通过开发一种可以处理此类复杂神经重症监护中的 AI/ML 的软件产品 数据。它还将允许共享经过清理、带注释的综合数据集。通过一个 临床医生研究人员、数据科学家和行业之间的深思熟虑的合作，我们将采取 NIH 支持 PRECICECAP 研究的数据，并使其广泛可用且易于使用。 该项目将提供一个重要的软件工具，其他人可以使用该工具进行类似的工作 研究，推进 NIH 的使命，使复杂数据公平（Findable、Accessible、 可互操作和可重用）。我们将开发一系列模块化功能（例如， 支持数据协调、注释或可视化），允许用户以图形方式构建 处理管道在适当的情况下最大化自动化并允许与 需要时提供数据。我们将开发一个用户友好的仪表板界面，以允许个人学习 站点和协调中心来理解复杂的数据，检查关键的元数据特征，以及 识别潜在的错误。模块化设计有利于AI/ML架构的动态配置 在同一界面内，允许各个模块协同组合以最大化 效率和再现性。结果将促进科学家之间开放、广泛的合作 使用类似的数据。