Development and testing of a step-wedge based method for the volumetric measurement of breast density

Breast density is known to be a major risk factor for breast cancer, and could therefore be important in predicting an individual’s chance of developing the disease. In order to provide quantitative estimates of breast density we have developed a method for determining the volume of dense breast tissue from digitized mammograms. The method is based on imaging a calibrated step-wedge alongside the breast. Together with an accurate measure of breast thickness this enables the thickness of dense tissue to be determined at each pixel. The compressed breast thickness is calculated from the magnification of small markers fixed to the edges of the compression plate, while a semi-circular profile is used to estimate thickness in the breast margin. Corrections are made for x-ray field non-uniformity. By imaging the step-wedge on each mammogram, the method is insensitive to variations in exposure parameters and film processing. The standard deviation on the dense tissue thickness is estimated to be 2.5 mm and is mainly due to uncertainties in deriving the matching step-wedge thickness. Using phantoms, repeatability and reproducibility tests gave standard deviations of 0.4 mm and 1.6 mm respectively. The method has been used in a two-year prospective study on the effect of weight loss on glandular and fatty breast tissue amongst premenopausal women. In women with a greater than 5% weight loss, the results showed significant reductions in breast volume and fat content, but no reduction in dense tissue volume. Comparing the cranio-caudal and medio-lateral oblique mammograms, the difference between measurements of the volume of dense tissue was on average 0.4 cm3 with a standard deviation of 3.0 cm3. Improvements have been made to the step-wedge design, which will be implemented in a forthcoming study of approximately 3000 women in the UK National Breast Screening Programme.