Risk Classification of Mammograms Using Anatomical Linear Structure and Density Information

Mammographic risk assessment is concerned with the probability of a woman developing breast cancer. Recently, it has been suggested that the density of linear structures is related to risk. For 321 images from the MIAS database, the images were segmented in to dense and non-dense tissue using a method described by Sivaramakrishna, et al. In addition, a measure of line strength was obtained for each pixel using the Line Operator method. The above-threshold linearity was calculated in dense and non-dense tissue for each image and the images were then classified by BIRADS class using linear discriminant analysis. The results show a marked improvement when both density and linear structure information is used in classification over density information alone.