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.

R21EB006466

2007-06-15

2010-05-31