Image Mining for Comparative Analysis of Expression Patterns in Tissue Microarray

在癌症组织微阵列中对表达模式进行一致性评价的大部分困难 这是观察者的主观印象。文献表明，当表征基于 通过计算机辅助分析，客观性、再现性和灵敏度大大提高。先进 成像和计算工具可能使研究人员能够检测和跟踪 可测量的参数，导致发现新的诊断和预后线索， 仅通过人的视觉检查就可以看出。本提案的中心目标是设计、开发、 部署和评估基于内容的图像检索系统，以执行快速、可靠的比较 分析癌组织微阵列中的表达模式。拟议项目是一个自然的延伸， 主要研究人员的先前工作，并利用几个关键资源和计算工具，包括 基于网络的图像引导决策支持系统，分布式显微镜系统，虚拟显微镜 系统，DataCutter允许在计算机和存储集群上分布式执行算法， 智能图像存档系统和癌症生物医学信息学的网格中间件组件 Grid（caBIGTM）体内成像工作区。拟议项目的特点是 支持查询和执行跨大型数据集的比较的能力， 标准的机器人和虚拟显微镜以及自动定位和检索那些成像的能力 来自分布式“金标准”档案中的组织盘，其表现出表达模式， 最类似于给定查询盘的那些。基于排序检索查询盘的多数逻辑， 将被客观和可重复地评估和分类。为了测试这些技术， 将在位于纽约癌症研究所的战略地点建立网格实验室 泽西大学（CINJ）、哥伦比亚大学（CU）、俄亥俄州州立大学（OSU）、罗格斯大学（RU）、 新泽西医学与牙科大学（UMDNJ）和宾夕法尼亚大学牙科学院 医学（宾夕法尼亚大学）。该实验室将使用caBIG caGrid基础设施建设。完成后 项目，软件和基础技术将提供给科学界， 符合caBIG标准的资源，用于协作研究、教育和临床决策支持。

Much of the difficulty in rendering consistent evaluation of expression patterns in cancer tissue microarrays arises from subjective impressions of observers. The literature shows that when characterizations are based upon computer-aided analysis, objectivity, reproducibility and sensitivity improve considerably. Advanced imaging and computational tools could potentially enable investigators to detect and track subtle changes in measurable parameters leading to the discovery of novel diagnostic and prognostic clues which are not apparent by human visual inspection alone. The central objective of this proposal is to design, develop, deploy and evaluate a content-based image retrieval system for performing quick, reliable comparative analysis of expression patterns in cancer tissue microarrays. The proposed project is a natural extension of the lead investigators' prior work and leverages several key resources and computational tools including a web-based image guided decision support system, a distributed telemicroscopy system, a virtual microscopy system, DataCutter which allows distributed execution of algorithms on computer and storage clusters, an intelligent image archival system, and Grid middleware components from the cancer Biomedical Informatics Grid (caBIGTM) In-vivo Imaging Workspace. The distinguishing characteristics of the proposed project are the capacity to support queries and perform comparisons across large datasets originating from both standard robotic and virtual microscopes and the capacity to automatically locate and retrieve those imaged tissue discs from within distributed, "gold-standard" archives which exhibit expression patterns which are most similar to those of a given query disc. Based upon the majority logic of the ranked retrievals query discs will be objectively and reproducibly assessed and classified. To test these technologies a multi-institutional, Grid-enabled laboratory will be established among strategic sites located at The Cancer Institute of New Jersey (CINJ), Columbia University (CU), the Ohio State University (OSU), Rutgers University (RU), the University of Medicine & Dentistry of New Jersey (UMDNJ), and the University of Pennsylvania School of Medicine (UPenn). This laboratory will be built using the caBIG caGrid infrastructure. Upon completion of the project, the software and underlying technologies will be made available to the scientific community as caBIG compliant resources for collaborative research, education and clinical decision support.