AI could reshape breast cancer screening guidelines
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Allie Carl/Axios
For years, patients have had to navigate a maze of conflicting mammogram advice. Now, artificial intelligence could help cut through the noise — by tailoring breast cancer screening schedules to individual risk.
Why it matters: AI is beginning to bridge the gap between one-size-fits-all screening guidelines and personalized medicine. By analyzing thousands of mammograms and patient outcomes, new tools can help doctors better identify who's at highest risk — and when they should be screened.
Between the lines: Major health organizations still differ on when — and how often — women at average risk should get screened.
- American Cancer Society: Annual screening is optional for ages 40–44; recommended for ages 45–54; every 1–2 years for ages 55+.
- American College of Radiology: Annual mammograms starting at 40.
- U.S. Preventive Services Task Force: Mammograms every other year for women ages 40–74.
And in addition to mammograms, there are other available screening tests:
- Ultrasound: Often complements mammograms, uses sound waves and is helpful in distinguishing cysts from tumors.
- MRI: For high-risk patients, especially with genetic predispositions.
- CEM: Uses iodine-based dye, often useful for patients with dense breasts.
- Thermography: Has "no rigorous scientific data," despite viral online interest, says Elias Obeid, breast medical oncologist at Hackensack Meridian Health Hennessy Institute.
What they're saying: "Using AI to develop individualized screening schedules rather than one-size-fits-all recommendations" is where the field is heading, says J. Pierre Sasson, the department chair of radiology at Mount Auburn Hospital and assistant professor of radiology at Harvard Medical School.
- Yes, but: "The human connection — compassion, reassurance and judgment — is still the heart of medicine, and that must never be lost," he says.
The latest: Sasson tells Axios that AI is already integrated into his hospital's workflow.
- It flags suspicious areas on mammograms, potentially shortening time to diagnosis.
- And it helps ensure that all images are high quality, reducing the need for patients to come back for repeat images.
Zoom in: Mirai, a deep-learning model, is one emerging use of AI that could make way for more personalized screening guidelines.
- It's collected information from thousands of mammograms to develop an algorithm that predicts breast cancer risk up to five years in advance.
- Early signs indicate it could help doctors identify high-risk patients from a single mammogram — before there's anything visible to the human eye.
- Similar mammogram-based risk tools include Transpara, ProFound and Clairity Breast.
What we're watching: AI is also driving cancer screening efforts outside of the standard mammogram.
- SpotItEarly, a tech startup, is conducting a trial where users breathe into a mask. Trained beagles then sniff for disease. AI tracks the dogs' vitals and movements to spot behavior changes that could signal disease.
