Wong Stephen T C, Hoo Kent Soo, Knowlton Robert C, Laxer Kenneth D, Cao Xinhau, Hawkins Randall A, Dillon William P, Arenson Ronald L
Department of Radiology, University of California-San Francisco, 94143, USA.
J Am Med Inform Assoc. 2002 May-Jun;9(3):239-54. doi: 10.1197/jamia.m0988.
A comprehensive data warehouse framework is needed, which encompasses imaging and non-imaging information in supporting disease management and research. The authors propose such a framework, describe general design principles and system architecture, and illustrate a multimodality neuroimaging data warehouse system implemented for clinical epilepsy research. The data warehouse system is built on top of a picture archiving and communication system (PACS) environment and applies an iterative object-oriented analysis and design (OOAD) approach and recognized data interface and design standards. The implementation is based on a Java CORBA (Common Object Request Broker Architecture) and Web-based architecture that separates the graphical user interface presentation, data warehouse business services, data staging area, and backend source systems into distinct software layers. To illustrate the practicality of the data warehouse system, the authors describe two distinct biomedical applications--namely, clinical diagnostic workup of multimodality neuroimaging cases and research data analysis and decision threshold on seizure foci lateralization. The image data warehouse framework can be modified and generalized for new application domains.
需要一个全面的数据仓库框架，该框架在支持疾病管理和研究方面涵盖成像和非成像信息。作者提出了这样一个框架，描述了一般设计原则和系统架构，并举例说明了一个为临床癫痫研究而实施的多模态神经影像数据仓库系统。该数据仓库系统构建在图像存档与通信系统（PACS）环境之上，并应用迭代式面向对象分析与设计（OOAD）方法以及公认的数据接口和设计标准。该实现基于Java CORBA（公共对象请求代理体系结构）和基于Web的架构，该架构将图形用户界面表示、数据仓库业务服务、数据暂存区和后端源系统分隔为不同的软件层。为了说明数据仓库系统的实用性，作者描述了两个不同的生物医学应用——即多模态神经影像病例的临床诊断检查以及癫痫病灶侧化的研究数据分析和决策阈值。图像数据仓库框架可以针对新的应用领域进行修改和推广。
