A Computer-Based Decision Support System for Breast Cancer Diagnosis

Abstract

Abstract : The goal of this project is to develop decision support system for breast cancer diagnosis, treatment option, prognosis, and risk prediction. This project focuses on the development of advanced image pattern analysis in diagnostic imaging and information integration methodology to statistically analyze the distinction between lesion-like normal site and real lesion site. The specific aims of this research project are: (1) image pattern analysis of breast tissue in mammography using both computational features and BI-RADS features provided by radiologist for the prediction of malignancy associated with masses; (2) development of visual presentation methods for radiologists' use in the consultation system; (3) performing a pre-clinical test through an ROC analysis. The clinical goal of this consultation system is to provide scientific tools for doctors to have electronic magnification views, to perform feature analysis of suspected mammographic patterns, to access a large database and investigate clinically similar cases, and to visually inspect the features of a case in various statistical distribution using graphic displays. We have accomplished (1) feature extraction, (2) feature database construction, (3) data mining visual explanation tool development, (4) feature database structure exploration, (5) feature ranking and selection, and (6) classification of mass and non-mass based on selected.