5 Takeaways from BCRF’s Webinar on AI and Breast Cancer Risk

A new AI tool is improving breast cancer prevention with smarter, more inclusive risk assessment

Just ahead of Breast Cancer Awareness Month, BCRF hosted a groundbreaking conversation on how artificial intelligence (AI) is reshaping breast cancer risk assessment, early detection, and prevention—especially in younger women.

Joining BCRF’s Sadia Zapp was Dr. Constance Lehman, breast imaging expert, BCRF investigator, and founder of Clairity Breast—the first AI-driven breast cancer risk assessment tool to receive FDA approval. Clairity Breast analyzes patterns in mammograms to predict a woman’s risk of developing breast cancer, offering a way to identify many patients who would otherwise never be flagged as high risk. In May 2025, the platform received FDA de novo authorization. It is expected to become available in late 2025, with broader rollout expected in 2026.

Below, we recap five key takeaways from their conversation. You can also watch the full webinar above and join our email list to find out about upcoming events.

1. AI can uncover individuals with high risk that traditional models miss

Zapp started the conversation by sharing her own experience of being diagnosed with breast cancer at 36 with no family history or genetic mutation known to increase risk. Her situation is all too common, but overlooked by traditional risk assessment models, like Gail and Tyrer-Cuzick. These tools focus on flagging people who have a known genetic mutation or family history, yet 85 percent of breast cancer cases occur in people without those obvious risk factors, Dr. Lehman explained.

Until recently, there hasn’t been a reliable way to identify women with a high risk of breast cancer who don’t have those major risk factors. Clairity Breast will change that, though. It uses AI to analyze subtle patterns in screening mammograms to detect who is more likely to develop breast cancer over the next five years and offer opportunities to potentially prevent it altogether.

“We’re now identifying a really important group of patients who can enter into care pathways in the absence of a genetic mutation and in the absence of a strong family history, but in the presence of a validated, strong, effective AI image-based risk score,” Dr. Lehman explained. “That’s going to open the door to improve care for more women.”

2. A baseline mammogram at 30 could personalize screening and prevention protocols

One of the most important points from the conversation was Dr. Lehman’s call for a new approach to screening younger women for breast cancer. In the U.S., women with average risk of breast cancer (meaning no known family history or genetic mutations known to affect risk) can start receiving annual mammograms at age 40.

Dr. Lehman suggested that all women instead receive a baseline mammogram at 30. That early image, when paired with Clairity Breast’s AI-informed risk score, would identify a subset of women who could benefit from earlier, more frequent screening and preventative care strategies. Those found to be at average risk could receive subsequent mammograms at 35, then 40, before following the standard screening recommendations, while those flagged as higher risk could move into more intensive care pathways right away.

While the idea of recommending mammograms to women in their 30s might sound revolutionary, it’s not a new concept, noted Dr. Lehman.

“It wasn’t that long ago that the American Cancer Society actually told all women they need a mammogram at 35. They wanted that baseline mammogram because they knew women were being diagnosed in their 30s, and this was back in 1970,” she explained.

That recommendation eventually fell away, largely because of concerns about false positives and unnecessary biopsies in younger women who, statistically, tend to be at lower risk. It took decades of debate just to lower the recommended screening age from 50 to 40 in the U.S.

Now, AI offers a way forward. Instead of recommending annual mammograms for all women in their 30s—which could put even those at average risk through unnecessary tests and the stress of receiving false positives—image-based risk scores can pinpoint only those who truly need earlier intervention. This added precision could help catch cancers sooner in the women most at risk while minimizing the stress, procedures, and potential harms for everyone else.

3. AI risk scores are dynamic, empowering women to make decisions about their lives

Another key point of the discussion was that, unlike genetic tests, AI image-based risk scores are not static. An individual’s score can change over time, from mammogram to mammogram, giving her the most current assessment of her risk of breast cancer. In BCRF-supported research, Dr. Lehman and her colleagues found that AI-powered breast cancer risk scores often begin to creep up four to six years before a breast cancer diagnosis—long before symptoms or lumps appear.

“It’s not just about your score, but how this year’s score compares to your score last year or two years ago,” said Dr. Lehman.

This evolving understanding of breast cancer risk can help women make critical decisions about their health and lives. A rising score could prompt a woman to begin preventive medications, make lifestyle changes to reduce risk (such as quitting alcohol consumption), increase the frequency of screening, or consider the timing of pregnancy or parenthood before a risk-reducing surgery.

Just as importantly, changes to a score could impact a woman’s approach to preventative therapy. If a score decreases after starting a preventive therapy, it offers reassurance that it’s working and should be continued. Or, if the score doesn’t drop, it might indicate the need to switch to a different medication or dosage, explained Dr. Lehman. This feedback loop makes breast cancer risk assessment more actionable than ever before.

“I see a lot of different areas where we can hopefully leverage this data and information to guide better choices, more precise choices, and more targeted choices for women because of the dynamic nature of this risk sore,” added Dr. Lehman.

4. AI can help close inequities in breast cancer care

During the discussion, Zapp raised the question of how risk models can better serve all women—not just those who fit the narrow criteria of age, genetics, or race. Dr. Lehman explained that traditional tools were developed using a disproportionately high number of white women’s data, which has limited their accuracy for women of other racial and ethnic backgrounds, many of whom face disparities in breast cancer risk and outcomes. AI, she said, offers a chance to correct that inequity.

“The more I was studying the AI image-based risk models, the more I studied the traditional risk models, which have struggled from the beginning [because] they were built largely on European, Caucasian women, and it’s been hard to have them generalize to women who identify as Black, Asian, and Hispanic,” said Dr. Lehman. “We didn’t want to repeat the sins of our fathers, so to speak.”

To train Clairity Breast’s AI model, the researchers used data from more than 2 million mammograms from women all around the world. As part of the process to get De Novo authorization from the FDA, the model was later externally validated on another diverse data set, the results of which will be presented at the next annual meeting of the Radiological Society of North America. It included data from hundreds of thousands of exams and images collected from 10 geographically diverse areas and people of all races.

The AI-powered risk assessment scores offer a more inclusive way to predict who is most likely to develop breast cancer, regardless of race, age, or background. For Dr. Lehman, that equity piece is just as important as the science itself.

5. AI-powered risk assessment could be available by the end of 2025, but widespread rollout will take longer

Earlier this year, Clairity Breast became the first AI-driven breast cancer risk assessment tool to receive FDA De Novo authorization—a milestone that set the stage for clinical use. Unlike earlier AI tools designed to assist radiologists in spotting cancers on images that they may have missed, Clairity was authorized to do something entirely new: predict a woman’s future risk of breast cancer, something no human can calculate from an image alone. Dr. Lehman said the company is on track to have the commercial version of the product available to the first patients at three early health system partners by late 2025.

Broader rollout will begin in 2026, starting with a self-pay model while coverage decisions from Medicare, Medicaid, and private insurers are still in progress. To help bridge the gap, Clairity is working with partners to keep the price as low as possible (around $199) and developing programs to subsidize costs for patients with Medicaid. The goal is to make sure cost and geography don’t become barriers to access for this life-saving tool.

“It has always been built on access and inclusion and we’re going to stay with that,” said Dr. Lehman.