SBIR Phase I: COVID-19 Imaging XNAT Suite (CIXS): An informatics platform for developing, validating, and deploying AI applications for COVID-19 imaging

这个小企业创新研究第一阶段项目的更广泛的影响/商业潜力将是实施一种人工智能(AI)方法来评估医学图像中的新冠肺炎，包括疾病诊断、疾病严重性预测和治疗反应测量。鉴于新冠肺炎疫情仍在持续，这项提议的主要目的是支持一线临床医生和研究人员护理患者和开发有效的治疗方法。成像是新冠肺炎患者护理的一个关键但仍不发达的组成部分。该项目将利用最佳实践来培训、验证和部署人工智能应用程序，从而促进对新冠肺炎成像实践的理解。此外，为创建新冠肺炎算法而开发的成像人工智能平台将广泛用于其他放射学应用。小企业创新研究第一阶段项目旨在实施一个定位独特的平台，以支持成像AI算法的全周期开发、验证、部署和持续调整，以改善对新冠肺炎患者的护理。技术任务包括：开发从不同医院信息系统聚合临床和成像数据的方法，实施服务和用户界面，以跨远程数据和计算系统安全地进行联合学习实验，以及开发发布和订阅经过验证的高质量人工智能模型的机制，以便在临床环境中部署。核心平台已经在许多大型学术医学中心运行，因此翻译将是直截了当的。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project will be to implement an artificial intelligence (AI) method to assess COVID-19 in medical images, including diagnosis of the disease, prediction of disease severity, and measurement of treatment response. Given the ongoing COVID-19 pandemic, the primary aim of this proposal is to support frontline clinicians and researchers in caring for patients and developing effective therapies. Imaging is a critical but still underdeveloped component of COVID-19 patient care. This project will advance the understanding of COVID-19 imaging practices, utilizing best practices for training, validating, and deploying artificial intelligence applications. In addition, the imaging AI platform developed for creating COVID-19 algorithms will be broadly useful for a wide range of other radiology applications. This Small Business Innovation Research (SBIR) Phase I project aims to implement a platform that is uniquely positioned to enable full cycle development, validation, deployment, and ongoing tuning of imaging AI algorithms to improve the care of COVID-19 patients. Technical tasks include: developing methods to aggregate clinical and imaging data from diverse hospital information systems, implementing services and user interfaces to conduct federated learning experiments securely across remote data and computational systems, and developing mechanisms to publish and subscribe to validated high quality AI models for deployment in clinical environments. The core platform is already operational in many large academic medical centers, so translation will be straightforward.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.