Investigation of Practical Scoring Methods for Breast Density

Abstract

Breast density is known as a strong risk factor for breast cancer. Clinically, physicians often use the BI-RADS or the Boyd categories to describe the density of breast, measured by observing mammograms. More accurately, breast density is measured by the percentage of glandular tissue in a breast. For all these methods, there might be more easily interpretable clinical value if the breast density was reported with scoring methods which are correlated to the patient distribution. In this paper two practical scoring methods will be discussed. The first one is population-based, with each segment of the continuous scores matching the patient BI-RADS distribution found in large scale clinical study. The second one is statistics-based, with the breast density result compared with the mean and the standard deviation from a reference population. Both methods will be described in details, together with preliminary results from an evaluation study with a total of 942 patients.