
Share
Table of Contents
Detection vs. prediction: what's the difference?
How AI predicts risk from a mammogram
Meet Clairity Breast: the first FDA-authorized tool
How accurate is AI risk prediction?
What a risk score could mean for you
Who might benefit most?
Limitations and what to keep in mind
When to talk to your healthcare provider
Frequently asked questions
Stay proactive with Everlywell
Medically reviewed by Gillian (Gigi) Singer, MPH, on October 5, 2026
To give you technically accurate, evidence-based information, content published on the Everlywell blog is reviewed by credentialed professionals with expertise in medical and bioscience fields.For decades, a mammogram has had one job: to find breast cancer that is already there. But a new generation of artificial intelligence (AI) is changing what a mammogram can do. The same image used to detect cancer today can also be analyzed by AI to estimate your risk of developing breast cancer in the future, even when no cancer is visible on the mammogram.
This article explains how AI turns a routine mammogram into a personalized risk forecast, how accurate these tools are, and what a risk score could mean for your care.
Detection vs. prediction: what's the difference?
Understanding this new technology starts with a simple distinction. Traditional mammography is about detection, while AI risk prediction is about prediction.
Traditional mammogram | AI risk prediction | |
Main question | Is there an abnormality that may indicate cancer now? | What is the likelihood of developing breast cancer in the future? |
What it looks for | Mammographic findings such as masses, calcifications, and other abnormalities | Patterns and characteristics within the mammographic image associated with future cancer risk |
Time frame | Current screening examination | A defined future period, often five years |
In other words, a screening mammogram can contain information associated with future breast cancer risk even when no cancer or abnormalities are identified on the examination.
How AI predicts risk from a mammogram
AI risk models are built using deep learning, a technique in which a computer learns to recognize patterns by studying enormous numbers of examples. These models analyze subtle patterns and characteristics in breast images that are associated with future breast cancer risk, even when cancer is not identified on the screening examination.1
Much of this work traces back to a research model called MIRAI, developed by scientists at the Massachusetts Institute of Technology and Massachusetts General Hospital. MIRAI demonstrated that a deep learning model could predict future breast cancer risk directly from a screening mammogram, and it laid the groundwork for the tools now reaching patients.
How accurate is AI risk prediction?
Early research suggests that AI can help predict a person's risk of developing breast cancer. In one study of more than 13,000 women with normal mammograms, AI was better at predicting breast cancer risk over the next five years than a standard risk calculator. The best results came when doctors used both AI and the standard risk calculator together. Another study found that an AI tool using three years of mammogram images was more than twice as accurate at predicting breast cancer risk as a standard risk questionnaire. These results are promising, but more research is needed to confirm how well AI risk tools work before they are used widely in health care.5
What a risk score could mean for you
A risk score is not a diagnosis. It is an estimate of how likely you are to develop breast cancer over a defined window, usually the next five years. A higher score does not mean you have cancer, and a lower score does not guarantee you never will.
Under the 2026 NCCN guidance, people aged 35 and older with an elevated five-year invasive breast cancer risk of at least 1.7% based on an image-based risk assessment may be considered for enhanced screening and risk-reduction strategies, including supplemental imaging such as breast MRI.6
Who might benefit most?
AI risk prediction may be especially helpful for those whose risk is underestimated, including those under the age of 50 and those without a strong family history.7 Because the model reads the mammogram itself rather than relying on questionnaires, it can flag risk that other tools miss.
Limitations and what to keep in mind
AI risk prediction is a powerful addition to breast health, but it has limits:
It predicts risk; it does not detect or diagnose existing cancer. Regular mammograms and clinical care remain essential.8
A high score is not a certainty, and a low score is not a guarantee. Scores guide decisions rather than dictate them.
Although research has shown that mammography-based AI can distinguish higher- and lower-risk groups, researchers are still studying whether using these predictions to change screening actually improves patient outcomes.4
Availability is still growing, so not every imaging center offers AI risk assessment yet.
When to talk to your healthcare provider
If you're interested in AI-based risk assessment, ask your provider or imaging center whether it is available where you receive care and how the results would be used alongside your other breast cancer risk factors. Your healthcare provider can help you interpret a score and decide whether any changes to your screening plan make sense.
Frequently asked questions
Is AI risk prediction the same as an AI reading my mammogram for cancer?
No. Some AI tools help radiologists detect existing cancer on a mammogram. Risk prediction tools like Clairity Breast instead estimate the chance that cancer will develop in the future, even when no cancer is visible now.8
Does an AI risk score replace my regular mammogram?
No. Risk prediction works alongside routine screening, not instead of it. You still need regular mammograms to detect any cancer that may develop.8
How far into the future does the score look?
Tools like Clairity Breast estimate risk over the next five years.1 Your healthcare provider can help you understand what that means for your screening plan.
Is AI risk prediction accurate for everyone?
The FDA-authorized model was validated across differences in age, race, and breast density and is designed to provide well-calibrated estimates across these groups, but no risk-prediction tool performs identically across every individual or population.9 As with any tool, results are best interpreted with your healthcare provider.
Stay proactive with Everlywell
AI-powered risk prediction reflects a bigger shift in cancer care: using technology to understand risk and catch disease as early as possible. You no longer have to wait for it to reach your imaging center.
The Clairity Breast™ Cancer 5-Year Risk Assessment is available directly through Everlywell. Its FDA-authorized AI-driven software analyzes a routine screening mammogram from the last 12 months to estimate your personalized five-year breast cancer risk, and provider-reviewed results are delivered to your My Everly account with guideline-based information to review with your healthcare provider. If you already have a compatible screening mammogram, Everlywell can work with your imaging center to retrieve it; if your mammogram cannot be used, you receive a full refund.
Clairity Breast estimates future risk. It does not detect or diagnose cancer, and it is not a substitute for mammograms or ongoing care with your provider. Keep up with your recommended screenings, and talk with your healthcare provider about what your risk score means for your plan. Everlywell is here to help you stay proactive about your health.
- AI breast cancer detection and diagnosis. Breast Cancer Research Foundation. https://www.bcrf.org/blog/ai-breast-cancer-detection-screening/. Accessed July 17, 2026.
- De Novo classification request for Allix5 (DEN240047). U.S. Food and Drug Administration. https://www.accessdata.fda.gov/cdrh_docs/pdf24/DEN240047.pdf. Published May 30, 2025. Accessed September 23, 2026.
- AI tops density in predicting breast cancer risk. Radiological Society of North America. https://www.rsna.org/media/press/2025/2617. Published November 26, 2025. Accessed September 22, 2026.
- Lowry KP, Jeong HE, Kim KH, et al. Current state of mammography-based artificial intelligence for future breast cancer risk prediction: a systematic review. J Natl Cancer Inst. 2026;118(3):392-403. https://doi.org/10.1093/jnci/djag002
- Conner K. Using AI (artificial intelligence) to detect breast cancer. Breastcancer.org. https://www.breastcancer.org/screening-testing/artificial-intelligence. Published June 11, 2026. Accessed September 22, 2026.
- NCCN updates breast cancer guidelines to include AI-based risk assessment. Targeted Oncology. https://www.targetedonc.com/view/nccn-updates-breast-cancer-guidelines-to-include-ai-based-risk-assessment. Accessed July 17, 2026.
- Schopf CM, Ramwala OA, Lowry KP, et al. Artificial intelligence-driven mammography-based future breast cancer risk prediction: a systematic review. J Am Coll Radiol. 2024;21(2):319-328. https://doi.org/10.1016/j.jacr.2023.10.018
- FDA authorizes 1st AI tool to predict 5-year breast cancer risk from routine mammograms. Radiology Business. https://radiologybusiness.com/topics/artificial-intelligence/fda-authorizes-1st-ai-tool-predict-5-year-breast-cancer-risk-routine-mammograms. Accessed July 17, 2026.
- Clairity Breast. Clairity, Inc. https://clairity.com/. Accessed July 17, 2026.
Share
Table of Contents
Detection vs. prediction: what's the difference?
How AI predicts risk from a mammogram
Meet Clairity Breast: the first FDA-authorized tool
How accurate is AI risk prediction?
What a risk score could mean for you
Who might benefit most?
Limitations and what to keep in mind
When to talk to your healthcare provider
Frequently asked questions
Stay proactive with Everlywell
Spotlight on
Breast Cancer Awareness Month

Your mammogram looks for cancer today. What about tomorrow? A new way to look at your breast cancer risk.
Learn more

When it comes to cancer, early is everything. Screen for early signs on 55+ cancers with a single blood draw.
Know early
Explore Everlywell









