Computer-Aided Analysis of Histopathology Images of Prostate Cancer

描述(由申请人提供)：该申请的广泛、长期的目标是通过早期发现和诊断前列腺癌来根除这种疾病作为死亡的主要原因。该项目的目标是测试在前列腺癌组织病理学的临床实践中开发一种挑战现有范式的新技术的可行性，使用计算机图像分析来辅助前列腺癌诊断。要检验的假设是，可以开发出能够准确分析前列腺癌组织病理学图像的计算机技术。其具体目标是：(1)建立前列腺癌数字组织病理学图像数据库；(2)开发在AMACR染色图像中自动识别前列腺癌的方法；(3)开发定量分析结构和形态模式的方法；(4)开发Gleason分级的自动评分方法。这项研究的设计将是与一名专业的泌尿科病理学家合作建立一个数字图像数据库，开发前列腺组织病理学图像的计算机图像分析技术，并进行前列腺癌自动检测和分级的可行性测试，以帮助病理学家。将使用的方法包括数字图像采集、图像处理、彩色图像分析、图像分割、观察者研究和计算机性能评估。追求这些目标的基本原理包括放射科医生解释的医学图像的计算机辅助诊断的进步，研究团队开发计算机辅助诊断方法的经验，以及临床组织病理学从模拟成像向数字成像的转变。这项研究与公共卫生有关，因为尽管组织病理学是其他医学分支的黄金标准，但对前列腺癌的诊断并不是100%准确。组织病理学图像的计算机辅助分析可能会使前列腺癌的组织病理学诊断更加准确，从而可能导致更有效的癌症治疗和更好的无癌生存结果。如果这项应用的目标得以实现，如果可行性被成功证明，那么一个新的、以前没有探索过的研究领域将打开，对临床实践具有潜在的未来影响。

DESCRIPTION (provided by applicant): The application's broad, long-term objective is to eradicate prostate cancer as a major cause of mortality through the early detection and diagnosis of this disease. The goal of this project is to test the feasibility of developing a new technology that challenges the existing paradigm in the clinical practice of prostate cancer histopathology, to use computer image analysis for assisting pathologists in prostate cancer diagnosis. The hypothesis to be test is that computer techniques can be developed to analyze prostate cancer histopathology images accurately. The specific aims are: (1) to establish a database of digital histopathology images of prostate cancer; (2) to develop methods for automated identification of prostate cancer in AMACR-stained images; (3) to develop methods for quantitative analysis of architectural and morphological patterns; and (4) to develop methods for automated scoring of Gleason grades. The research design will be to build a digital image database in collaboration with an expert urologic pathologist, develop computer image-analysis techniques for prostate histopathology images, and to conduct feasibility tests for automated prostate cancer detection and grading as an aid to pathologists. The methods to be used include digital image acquisition, image processing, color image analysis, image segmentation, observer study, and computer performance evaluations. The rationales for pursuing these goals include advance in computer-aided diagnosis for medical images interpreted by radiologists, the experience of the research team in developing computer-aided diagnosis methods, and a migration from analog imaging to digital imaging in clinical histopathology. This research is relevant to public health in that, while serving as a gold standard for other branches of medicine, histopathology is not one hundred percent accurate for prostate cancer diagnosis. Computer-aided analysis of histopathology images can potentially make histopathology diagnosis of prostate cancer more accurate, which could lead to more effective cancer treatment and better outcomes of cancer-free survival. If the aims of this application are achieved, and if feasibilities are demonstrated successfully, then a new, previously not explored area of research will open up, with potential future impact to clinical practice.