AI in the Breast Health Journey
In this 5-part Breast Health Series, we explore the challenges related to breast cancer screening and the role AI can play in early detection. The CDC highlights mammography as the best way to find cancer for most women of screening age. Mammograms are the current standard for breast cancer detection and an essential part of every woman’s healthcare journey.
Mammography, like any screening tool, has its limitations. Part 2 of the Breast Health Series explores those limits and offers potential solutions.
Breast Health Series
Part 1: The Breast Density Dilemma
Part 2: Understanding the Limits of Mammography
Mammography is a low-dose x-ray of the breast used to screen for breast cancer. Currently, 2D and 3D mammography (also called digital tomosynthesis) are used for screening. 2D mammograms take pictures of each breast from the front and the side, creating a single image of each breast with potential image overlap that can impair image clarity. 3D mammograms take many pictures of each breast from different angles, showing each layer of breast tissue. 3D mammograms provide clearer images with greater detail, improve breast cancer detection, can reduce callbacks and the need for additional imaging exams.
A recent study by UC Davis Health shows that nearly half of all women have at least one false positive mammogram after 10 years of annual screenings. 3D mammograms provide a more detailed picture of the breast created from hundreds of layered x-ray images. 3D mammograms are helpful for women with dense breasts, but the improvement is marginal and is far from an end all solution. The next steps in revolutionizing breast cancer detection will likely focus on alternative screening methods and imaging AI rather than mammography alone.
Another obstacle is the inconsistent guidelines for routine mammograms. This involves a variety of factors such as age, family and personal medical history, and other demographic info. Surprisingly, in the last decade, there has been a decline in annual screenings for breast cancer survivors. This is just one example that highlights the need for guideline standardization and further research focused on the barriers women face in getting the care they need.
Dr. Amy Patel discusses how the lack of industry agreement on breast screening recommendations can be confusing for women and impact access to care. She notes that providing consistent recommendations will go a long way in helping patients take control of their breast health journey.
With this goal in mind, a new AI model developed by a group of scientists from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and Jameel Clinic for Machine Learning and Health created risk-based screening guidelines and recommendations for patients as to when they should return for their next screening. This personalized approach factors demographic and historical data to outline a specific care schedule for each patient. The scientists believe that by tailoring the screening to the patient’s individual risk, they can improve patient outcomes, reduce overtreatment, and eliminate health disparities.
Thanks for reading Part 2 of the Breast Health Series – The Limits of Mammography. Join us next time as we explore how AI is transforming breast cancer screening.