Breast density assessed from film-screen mammography is one of the strongest risk factors for breast cancer and provides important information not only for risk assessment, but also for tailoring breast screening approaches and determining response to therapies. To date, the clinical potential of breast density has not been fully realized, in part due to lack of a standard, reproducible, objective and automated clinical density measure. Currently, the American College of Radiology Breast Imaging Reporting and Data System (BI-RADS) 4-category qualitative assessment is the most widely used clinical density measure, but it is subjective, requires a trained radiologist, is not always reported, and s limited in its ability to assess small, but potentially meaningful changes in density in response t therapies. We and others have developed promising automated mammographic measures on full field digital mammography (FFDM), including area, volumetric and variation density measures, and shown their associations with breast cancer. To date, studies have not compared the performance of automated density measures or their combination on FFDM images from the same women, to inform the best measure(s) for use in the clinical setting. Since over 85% of mammography in the US is FFDM, density measures need to be developed and examined in the FFDM environment to translate for use in today's clinical care. We propose to examine and compare the association of mammographic measures from FFDM with breast cancer and evaluate their ability to detect changes in response to menopause and hormone therapies among women ages 35-75 receiving FFDM within the Mayo Clinic and San Francisco Mammography Registry breast screening practices between 2006-2015. Specifically, we propose 1) To retrieve serial, digital mammograms prior to diagnosis (or corresponding date for controls), BI-RADS density and covariates on a large nested case-control study of 2800 incident breast cancers and 5600 matched controls, and to estimate automated area, volumetric and variation density measures on all images;2) To examine the association of breast density measures from both the earliest and multiple subsequent mammograms with breast cancer and to assess whether these associations differ by age, invasive vs. in situ breast cancer and ER, PR and HER-2 defined subtypes;and 3) To estimate menopause- and therapy-related changes in density measures on FFDM from 500 healthy perimenopausal women, 1000 peri or postmenopausal women initiating hormone therapy, and 1000 breast cancer cases initiating endocrine therapy. This protocol will answer: 1) Which automated density measure or combination of measures are most clinically useful in the FFDM environment as a risk factor for breast cancer and for detecting changes in density in response to therapy? 2) Do density measures perform similarly in younger vs. older women and for specific types of breast cancer? And 3) Do multiple measures of breast density improve risk associations compared to a single measure? This proposal will impact breast density assessment across the US, allowing for standardization in clinical practice and opportunities to integrate density measures into individualized risk assessment and screening.
The clinical potential of breast density, one of the strongest risk factors for breast cancer, has not been realized, in part due to lack of a standard, reproducible, objective and automated density measure. Our proposal will be the first to examine and compare clinical BI-RADS, automated breast density measures and combinations of these measures from full field digital mammography (FFDM) images with associated breast cancer outcomes to inform the best available clinical breast density measures to assist in risk prediction, tailored screen approaches and response to therapy.
|Fowler, Erin E E; Vachon, Celine M; Scott, Christopher G et al. (2014) Automated Percentage of Breast Density Measurements for Full-field Digital Mammography Applications. Acad Radiol 21:958-70|