Enhancing CBTN digital pathology processing pipeline through the use AWS cloud-based services to enable automation, parallel processing, and rapid use of AI/ML analytics

通过使用 AWS 基于云的服务增强 CBTN 数字病理处理管道，以实现自动化、并行处理和快速使用 AI/ML 分析

• 批准号: 10827712
• 负责人: Robert J Carroll
• 金额: $ 28.22万
• 依托单位: CHILDREN'S HOSP OF PHILADELPHIA
• 依托单位国家: 美国
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-09-26 至 2025-08-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10827712
• 关键词: Acceleration; Adoption; Adult; Artificial Intelligence; Automation; Big Data; Biological; Brain Neoplasms; Cancer Etiology; Child; Childhood; Childhood Brain Neoplasm; Clinical Data; Cloud Computing; Cloud Service; Communities; Data; Data Analytics; Data Set; Development; Disease; Ecosystem; Environment; Etiology; Funding; Generations; Goals; Healthcare; Human Resources; Image; Infrastructure; Ingestion; Intake; Knowledge; Life Cycle Stages; Machine Learning; Malignant Childhood Neoplasm; Malignant Neoplasms; Manuals; Medical Imaging; Methods; Modality; Modernization; Molecular; Morphology; Pathology; Pathway interactions; Patient Care; Patients; Pediatric Research; Performance; Population; Preparation; Process; Property; Research; Resolution; Resources; Sampling; Scanning; Services; Site; Slide; Software Tools; Structural Congenital Anomalies; Technology; Testing; Translations; Tumor Biology; United States National Institutes of Health; Work; advanced analytics; childhood cancer mortality; clinical phenotype; cloud based; cloud platform; cloud storage; computing resources; cost; cost estimate; data exploration; data ingestion; data integration; data management; data portal; data repository; data resource; data sharing; digital; digital pathology; empowerment; feasibility testing; flexibility; genomic data; large scale data; machine learning method; multimodality; multiple omics; novel; open source; parallel processing; parallelization; personalized medicine; phenotypic data; precision medicine; programs; research study; software development; standard of care; tool; translational model; user-friendly; web services; years of life lost

PROJECT SUMMARY The Gabriella Miller Kids First pediatric research program is a collaborative initiative with the goal of understanding the etiology and drivers of pediatric diseases. Related efforts include (1) developing data-driven platforms, workflows and tools; (2) accelerating discovery of generic causes and shared biologic pathways within and across conditions; and (3) enabling rapid translation of scientific discoveries to personalized treatments. Through this program, some of the largest, multi-modal pediatric datasets have been generated, harmonized, and released for use by the research community. In parallel, developed software platforms have enabled streamlined, user-friendly workflows in cloud environments to empower data exploration to analysis. While this work has significantly advanced the use of integrated datasets, particularly for genomics and clinical data, related advancements for medical imaging data have been slower. In the present project, we focus on digital pathology data in research contexts and aim to enable the use of modern cloud infrastructure and processes for Kids First digital pathology slides. Specifically, we will explore, implement, optimize, and assess cloud platforms and services to reduce current challenges in data ingest and release, and preparation of digital slide images for downstream analytics. The selected cloud solutions are expected to enable workflow automation, scaling of computing resources, and full support of the entire data lifecycle. This should drastically reduce the operational and computational costs of utilizing Kids First digital pathology data across research contexts. Together, this project will bridge digital pathology data and tools with high-performance cloud workspaces to accelerate their use with advanced analytics and multi-modal integration for all Kids First pediatric data.

项目摘要 加布里埃拉米勒儿童第一儿科研究计划是一个合作倡议，目标是 了解儿科疾病的病因和驱动因素。相关工作包括（1）开发数据驱动的 平台、工作流程和工具;（2）加速发现一般原因和共同的生物学途径 （3）使科学发现能够快速转化为个性化的 治疗。通过这个项目，一些最大的、多模态的儿科数据集已经生成， 统一，并发布供研究界使用。同时，开发的软件平台 在云环境中启用了简化的、用户友好的工作流，以支持数据探索和分析。 虽然这项工作大大促进了综合数据集的使用，特别是在基因组学和临床方面， 数据，医学成像数据的相关进展一直较慢。在本项目中，我们将重点放在 研究背景下的数字病理学数据，旨在使用现代云基础设施， Kids First数字病理切片的流程。具体来说，我们将探索、实施、优化、评估 云平台和服务，以减少当前在数据摄取和发布方面的挑战， 用于下游分析的幻灯片图像。选定的云解决方案预计将实现工作流程 自动化、计算资源的扩展以及对整个数据生命周期的全面支持。这将大大 降低在整个研究中利用Kids First数字病理学数据的运营和计算成本 contexts.总之，该项目将数字病理学数据和工具与高性能云 通过高级分析和多模式集成，为所有Kids First 儿科数据。