Iconographic display of mammographic features for computer-aided diagnosis of breast cancer

Abstract

We have recently developed objective methods for characterization of shapes of microcalcifications and tumors in mammograms, and measures of acutance for characterization of the radiographic definition (or edge sharpness) of tumors. The measures and factors quantify subjective descriptors used by radiologists in diagnosis and decision making. While computer extracted features permit computerized pattern recognition and decision making, they may not be readily usable by the radiologist in a computer aided diagnosis system. The objective of the work presented is to improve the diagnostic efficiency of mammography by the use of visualization tools. We propose an iconographic (or iconic) display system based upon graphical icons whose characteristics are related to the computer extracted shape factors and other features of mammographic entities. The icons are presented to the radiologist in the form of an image; the locations and visual characteristics of the icons are associated with corresponding diagnostic entities such as microcalcifications and tumors. The procedure is expected to assist radiologists in interpreting the features in an objective manner, and hence improve diagnostic efficiency.