TL;DR — Key Takeaways
– Clairity Breast uses AI to analyze routine screening mammograms and estimate a woman’s risk of developing breast cancer within five years.
– The FDA authorized the technology in 2025 as an AI-based tool specifically designed to assess future breast cancer risk from mammograms.
– Unlike traditional risk models that rely on factors such as age, family history and breast density, Clairity analyzes patterns within the mammogram itself.
For decades, estimating a woman’s likelihood of developing breast cancer has meant assembling a collection of possible indicators, including her family history, reproductive history, breast density, previous biopsies and, in some cases, genetic testing. Those factors can identify women at elevated risk, but they still cannot tell an individual woman whether cancer is waiting somewhere in her future.
Now, an AI tool is looking for another set of clues, inside the mammogram itself.
Clairity Breast analyzes a routine screening mammogram to estimate the likelihood that a woman will develop breast cancer within the next five years. In 2025, it became the first AI-based tool authorized by the U.S. Food and Drug Administration specifically to predict future breast cancer risk from a standard screening mammogram. On Oct. 1, 2026, Clairity expanded access to the technology nationwide through a partnership with digital health company Everlywell.
“The information we need to understand a woman’s breast cancer risk has been in her mammogram all along, we just haven’t been able to interpret it until Clairity Breast,” said Dr. Connie Lehman, Clairity’s founder and CEO and a professor of radiology at Harvard Medical School. “Now, it can also inform conversations with your healthcare provider about the best screening and risk reduction options for you.”
The American Cancer Society estimates that about 322,000 women in the United States will be diagnosed with invasive breast cancer in 2026, and more than 42,000 will die from the disease. Breast cancer accounts for roughly one-third of new cancers diagnosed in American women, excluding skin cancers. According to the National Cancer Institute, the median age at breast cancer diagnosis is 63.
Worldwide, breast cancer is the most commonly diagnosed cancer among women. The International Agency for Research on Cancer estimates that about 2.3 million women are diagnosed with the disease globally each year.
According to the American Cancer Society, more than 85% of women diagnosed with breast cancer have no family history of the disease.
Medicine has spent decades trying to pinpoint who is more likely to develop breast cancer. One of the best-known tools is the Gail Model, formally called the Breast Cancer Risk Assessment Tool. Developed by researchers at the National Cancer Institute and the National Surgical Adjuvant Breast and Bowel Project, it considers factors including age, reproductive history, previous breast biopsies and breast cancer among first-degree relatives to estimate a woman’s five-year and lifetime risk.
The Tyrer-Cuzick model incorporates a more extensive family history along with hormonal, reproductive and other factors. The Breast Cancer Surveillance Consortium model includes breast density, age, race and ethnicity, family history and previous breast biopsy results in estimating risk.
These methods construct a forecast from what is known about the woman.
Clairity is attempting to derive part of that forecast from what is contained in the breast image itself. The software analyzes pixel-level features and patterns within a standard bilateral screening mammogram, searching for characteristics associated with women who subsequently developed breast cancer. The result is a percentage representing the estimated probability of developing the disease within five years. According to Clairity, its model showed a 30% improvement in discrimination compared with traditional risk models.
The FDA authorized the technology, originally submitted under the name Allix5, in May 2025. According to Clairity, it was validated for FDA authorization using more than 77,000 screening mammograms from five U.S. sites and subsequently validated using more than 120,000 mammograms across 10 U.S. facilities.
This does not mean AI can peer five years into the future and see a tumor forming. Clairity Breast is a risk-assessment tool, not a prediction of what will happen to a particular woman. Someone receiving an elevated score may never develop breast cancer, while someone with a lower score still could.
It also does not replace a radiologist. According to the FDA, Clairity Breast is not intended to detect or diagnose breast cancer, interpret a mammogram or determine treatment. Instead, its five-year risk estimate provides another piece of information that a woman and her health care provider can consider when discussing screening and risk-reduction strategies.
The approach is also beginning to enter mainstream clinical guidance. The 2026 National Comprehensive Cancer Network guidelines include AI-based mammogram risk assessment as an option for identifying women at increased breast cancer risk beginning at age 35.
Until recently, access to Clairity Breast was limited largely to participating health systems, including Beth Israel Deaconess Medical Center in Massachusetts and Invision Sally Jobe in Colorado. Through Everlywell, women nationwide with a qualifying screening mammogram from the previous 12 months can now request the assessment. The $249 service includes ordering and review through licensed providers affiliated with Everlywell, and the results can be shared with a woman’s own physician.
“Women deserve clear answers about their health, and access to the most advanced tools in health care, not years from now, but today,” said Julia Cheek, Everlywell’s founder and CEO. “This partnership expands access to cutting-edge preventive care, and that means a woman can learn more about her future breast cancer risk from the mammogram she already gets, and walk into her doctor’s office with information she’s never had before.”
Clairity Breast is also part of a broader movement in medical AI, one centered on extracting information from familiar tests that clinicians were never able to see.
Researchers are doing something similar with electrocardiograms. AI systems trained on millions of ECGs and associated patient outcomes are learning to recognize subtle electrical patterns associated with heart failure, valve disease and other conditions, signals that may be too faint or complex for even an experienced cardiologist to identify consistently.
Mammograms may hold similarly unexpected information. In a recent study, researchers in Israel used AI to analyze 97,364 mammograms from 29,921 women, not for breast cancer, but for patterns associated with cardiovascular disease. The AI identified women with a history of stroke 86% of the time and distinguished women with high blood pressure and coronary heart disease from those without those conditions with reported accuracies of 79% and 78%, respectively.
The research suggests that medical images routinely collected for one purpose may contain biological information about something entirely different. Instead of requiring another scan or test, AI can potentially return to an image already sitting in a patient’s medical record and ask questions that the image was never originally intended to answer.
