OBJECTIFICATION OF MEDICAL FINDINGS--INDIVIDUAL IMAGING PROTOCOLS

这项提议的广泛、长期目标是建立一个具有以下特点的医疗保健提供系统：(1)注重成像，以促进对象化、循证的医学实践；(2)能够量身定做适当的成像方案，以便为个别患者量身定制适当的成像方案；以及(3)将远程放射学转变为在各种硬件平台上运行的基础设施。该提案的具体目的是通过根据专家的输入和结构化的患者数据定义单独定制的成像检查序列，将患者的主观临床表现客观化，开发自动化数据收集的方法，以促进与过程、结果和资源利用相关的医疗服务研究，开发将患者数据与医学文献相关联的算法，设计能够实现硬件独立性的模块化系统，并实施和评估具有明确定义的患者群体的原型。该项目促进了循证医学实践，应该会减少医疗不确定性，提高护理质量。研究设计和方法包括：(1)用于单独定制的成像协议的结构化数据输入、远程会诊和有监督的机器学习；(2)用于卫生服务数据收集的灵活的输入体系结构和数据表示；(3)用于内容关联的特征提取、模式发现和基于内容的索引技术；以及(4)以网络为中心的体系结构，其使用用于硬件独立性的TCP/IP、Java和CORBA标准。拟议基础设施的发展将促进循证医学实践，改善护理过程和结果，并通过提供更准确的卫生服务研究数据，促进成本效益研究。

The broad, long-term objective of this proposal a health care delivery system with the following characteristics: (1) a focus on imaging that facilitates an objectified, evidence-based practice of medicine, (2) the ability to tailor an appropriate imaging protocol to tailor an appropriate imaging protocol for an individual patient, and (3) transform teleradiology to an infrastructure that runs on a variety of hardware platforms. The proposal specifically aims to objectify a patient's subjective clinical presentation by defining individually tailored imaging work-up sequences based on input from specialists and structured patient data, develop methods for automated data collection to facilitate health services research related to the process, outcomes, and resource utilization, develop algorithms to correlate patient data with medical literature, design a modular system that enables hardware independence and implement and evaluate a prototype with a well-defined patient population. The project's facilitation of evidence-based practice of medicine should reduce medical uncertainties and improve quality of care. Research design and methods include: (1) structured data input, teleconsultation, and supervised machine learning for individually tailored imaging protocols; (2) a flexible input architecture and data representation for health services data collection; (3) feature extraction, pattern discovery, and content-based indexing techniques, for content correlations; and (4) a network-centric architecture which use TCP/IP, JAVA, and CORBA standards for hardware independence. Development of the proposed infrastructure will promote evidence-based practice of medicine, improve the process and outcome of care, and facilitate cost-effectiveness studies by providing more accurate health services research data.