Mammograms are specialized, low-dose X-rays of the breast that can show if and where tumors are located in breast tissue. Mammography is an effective tool for early detection of breast cancer and is considered the primary tool for breast cancer screening.
“Breast cancer is a common disease that affects millions of women globally, and detecting it early can help us guide people with appropriate treatment plans,” says Elizabeth Mittendorf, MD, PhD, chief of Multi-Disciplinary Oncology at Dana-Farber and chief of Breast Surgery in the Department of Surgery at Beth Israel Deaconess Medical Center (BIDMC).
During a mammogram, two plates compress the breast to take an image. Now, with the help of artificial intelligence (AI), mammograms can help determine if a person is at risk for developing cancer in the future.
More personalized care with AI mammography
Clairity Breast is an AI platform embedded into the imaging infrastructure at BIDMC and will soon be made available to patients at Dana-Farber. This new tool supplements the standard screening mammogram to generate an individual’s five-year risk of developing breast cancer.
“Incorporating Clairity is part of our broader effort to make risk assessments accessible, consistent, accurate, and equitable,” Mittendorf says. “We are the first in the country to do this.”
The standard mammogram will remain the same. Most women get their first screening mammogram at age 40, based on current national guidelines. While it won’t replace the need for biopsies if results require more probing, the technology will help give mammograms a dual purpose: inform care based on results, and generate risk assessment for the next five years.
“Last year alone, more than 100,000 screening mammograms were performed within the Beth Israel Lahey Health (BILH) system, and so this new AI approach will help shape these more personalized screening and prevention conversations with patients that we want to have,” Mittendorf says.
While current recommendations still advise annual mammograms, the technology could lengthen the time between screenings based on AI-assessed risk scores. For instance, if an individual’s risk falls below a certain threshold, extending the time between screenings may become possible. However, comprehensive studies still need to test this approach.
Traditional risk assessment tools are based on asking questions about an individual’s family, reproductive, and social history. The Clairity tool has been tested in different populations, which, Mittendorf says, could address some of the limitations of current risk assessment tools.
This year, the National Comprehensive Cancer Network (NCCN) changed their risk assessment guidelines, recommending that AI-based mammography risk assessment be added. The guidelines suggest that if a woman’s risk is more than 1.7%, then a healthcare provider should have conversations with her about risk prevention.
“If a radiologist says a patient’s screenings look good, but their score comes back at 3%, we can use that information in our conversations with the patient. It not only helps take care of your breasts today but also informs your breast health going forward,” Mittendorf says.