SBIR Phase I: Unified data description layer for magnetic resonance imaging scanners

这个小企业创新研究（SBIR）第一阶段项目的更广泛的影响/商业潜力是改善放射学和医疗成像设备的管理。拟议的分析平台将提供设备利用率的详细信息，以帮助监督操作，优化工作流程，更好地利用现有设备，并评估投资的成功。更有效地使用扫描仪预计将大大有利于患者群体，因为它将减少等待磁共振成像（MRI）的时间，增加患者访问，缩短成像方案，减少镇静持续时间，减少和预测延迟，并最终改善患者体验。该平台解锁的数据也将为放射科医生和研究人员开辟新的研究途径。这个小企业创新研究（SBIR）第一阶段项目旨在开发一种技术，重新利用磁共振成像（MRI）扫描仪产生的医学数字成像和通信（DICOM）数据，建立一个统一的、可查询的成像检查知识来源。这个项目将协调DICOM元数据，并以此为基础创建一个描述成像检查所有方面的本体。开发领域包括通过分析图像和检查的算法来推断扫描仪何时真正处于活动状态，从而恢复采集时间和扫描仪活动。该项目还通过训练机器学习模型来自动检测重复图像，这是时间表延误的主要来源，从而展示了数据源的影响。总的来说，这个项目的发展从DICOM数据中构建了时间和工作流程的关键方面，使放射学中的数据分析成为一种新的形式。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值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 is to improve radiology and management of medical imaging equipment. The proposed analytics platform will provide detailed insight into device utilization to help oversee operations, optimize workflows, better leverage existing equipment, and evaluate the success of investments. More efficient use of scanners is expected to substantially benefit the patient population as it will reduce the wait time for magnetic resonance imaging (MRI), increase patient access, shorten imaging protocols, reduce sedation duration, reduce and predict delays, and ultimately improve the patient experience. The data unlocked by the platform will also open new avenues of research for radiologists and researchers.This Small Business Innovation Research (SBIR) Phase I project aims to develop a technology that repurposes the Digital Imaging and Communications in Medicine (DICOM) data created by magnetic resonance imaging (MRI) scanners to build a unified, query-able source of knowledge about imaging exams. This project will harmonize DICOM metadata and build upon it to create an ontology that describes all the facets of imaging exams. Areas of development include recovering acquisition duration and scanner activity through algorithms that analyze images and exams to infer when the scanner was truly active. The project also demonstrates the impact of the data source by training a machine learning model to automatically detect repeated images, a prominent source of schedule delays. Overall, the developments from this project construct key aspects of timing and workflow from DICOM data to enable a new form of data analytics in Radiology.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.