BREAST DENSITY ASSESSMENT SOFTWARE:
Radiologists have traditionally determined breast density by visually comparing the amount of light (white, fibroglandular tissue) vs. dark (gray, fatty tissue) areas on a mammogram. The radiologist typically reports breast density in one of four Breast Imaging Reporting and Data System (BI-RADS) categories: fatty, scattered fibroglandular tissue, heterogeneously dense, or extremely dense. Breasts which are heterogeneously dense or extremely dense are considered “dense breasts.” Visual determination of density is a subjective assessment, which can vary from year to year even if there is no true change in the breast. Computer software can also be used to automatically characterize breast density on a digital mammography exam. Any density assessment (radiologist/visual or radiologist + automated tool) should be tracked over time as some breasts will become fatty replaced (and no longer dense) as a woman ages. If the density assessment shows the breasts are dense, this can be used as the basis to consider supplemental screening.
What is it? Software is available which presents interpreting radiologists with an automatic assessment of the percentage of dense tissue contained within the breast. The density assessment data are obtained by evaluating the routine mammogram or digital breast tomosynthesis exam with sophisticated algorithms. Current software requires the raw digital mammogram images or tomosynthesis projection data sets, and these are not routinely saved at most facilities.
How it works: Automated assessments calculate density as either area (length x width) or volume (length x width x height) percent density. The assessed area or volume of dense tissue is divided by the area or volume of the entire breast and then multiplied by 100 to yield a percentage. This percentage is generally then correlated to one of the four BI-RADS categories: fatty, scattered fibroglandular tissue, heterogeneously dense, or extremely dense. We do not fully understand whether the absolute amount of dense tissue or the percent of dense tissue is more important, though an analysis of multiple methods* found percent density more predictive of risk. Complexity of the pattern of breast density may be more important in predicting risk than the amount of dense tissue, though further study is needed.
Benefits: Automatically provides consistent breast density calculations across all patient populations. Removes the inter- and intra-radiologist subjectivity and variability of visual assessment of breast density.
Considerations: There are differences in software technology. Some approaches calculate the amount of fibroglandular tissue, some calculate percent area or percent volume that is dense, and some also consider the texture and variability (complexity) of density within the breast. Automated software typically provides one “average” measurement across the whole breast, or the maximum density value of the left or right breast. If one quadrant of the breast is particularly dense, this may prompt supplemental screening by the radiologist even though the software average density score across the entire breast is not dense: any computerized measurement of density should be reviewed in the context of a particular woman’s mammograms.
*Pettersson A, Graff Re, Ursin G, Dos Santos Silva I, Mccormack V, Baglietto L, Et Al. Mammographic density phenotypes and risk of breast cancer: a meta-analysis. J Natl Cancer Inst. 2014;106(5).
*Eng, A, Gallant, Z, Shepherd, J, Et Al. Digital mammographic density and breast cancer risk: a case-control study of six alternative density assessment methods. Breast Cancer Research 2014;16:439