BREAST CANCER IMMUNOHISTOLOGICAL IMAGING DATABASE

Abstract

Breast cancer is the most common pathology among women. The death rate from breast cancer among women remainshigh. Early diagnosis and individual therapy are effective ways to extend people's lives. The main diagnostic methods arecytological, histological, and immunohistochemical. The cytological method allows assessing the qualitative and quantitativechanges in cells, as well as identifying extra- and intracellular inclusions and microorganisms. The histological method allows you toexplore changes in the location of groups of cells in a particular tissue. The immunohistochemical method is based on the use ofbiomarkers. Immunohistochemical images are the result of an immunohistochemical investigation. The aim of the work is todevelop a database of immunohistological images of breast cancer. With the developed database, a database design methodologywas used, including infological, datalogical and physical design. The scientific novelty lies in the use of an object-oriented approachfor designing a database of immunohistochemical images. The practical value of the work lies in the development of all stages ofdatabase design. As a result, an infological model, a data model, and a UML database diagram have been developed. For thepractical implementation of the server part of the database, operating systems such as Windows / Linux / macOS can be used, thedatabase server is MySQL. The developed breast cancer database contains more than 500 images for four diagnoses. The imageresolution is 4096 x 3286 pixels. For each image, two features are given: relative area and brightness level. The developedHI&IHCIDB database has medium volume, high resolution, and quantitative characteristics in the description ofimmunohistochemical images