Automated assessment of area of dense tissue in the breast: a comparison with human estimation

Both interactive thresholding tools and human visual assessment have been related to the risk of developing breast cancer In this paper we explore the relationship between human assessment of area of dense tissue and the actual thickness of tissue in the breast by using a volumetric density technique to compute areas of dense tissue, varying the threshold below which areas of low density are discounted and observing the correlation with visual assessment of density at different thresholds Based on analysis of thresholds used in the automated method, radiologists' definition of a dense pixel is one in which the percentage of glandular tissue is between 10% and 20% of the total thickness of the compressed breast at that point.