An AI Tool That Predicts Breast Cancer Risk Is Now Available Nationwide

Women can purchase the assessment through Everlywell’s platform if they’ve had a qualifying screening mammogram in the past 12 months, says Rachel Conklin, PA-C, the vice president of clinical solutions at the company. A licensed clinician in Everlywell’s network reviews and orders the assessment, then reviews the results and shares them, along with next steps based on national guidelines, she says.

“By predicting breast cancer risk over the next five years from a mammogram, it gives women the opportunity to tailor screening intensity to their risk,” says Jean Bao, MD, a clinical associate professor at Stanford Medical School and a general surgeon at Stanford Health Care in California who specializes in breast surgical oncology. Dr. Bao is not affiliated with Everlywell.

The AI Is Trained to Look for Patterns in Breast Cancer Risk

For decades, radiologists have recognized that patterns in breast tissue on a mammogram are linked to breast cancer risk, says Connie Lehman, MD, PhD, the founder and chief executive officer of Clairity, the company that developed the tool. The challenge was turning those subtle patterns into a reliable risk estimate for an individual woman, says Dr. Lehman, who is also a professor of radiology at Harvard Medical School in Boston.

Her team turned to AI for its pattern-recognition abilities. “A familiar example is a computer learning to distinguish cats from dogs by studying many images labeled ‘cat’ or ‘dog,’” says Lehman.

In this case, the AI software was fed hundreds of thousands of screening mammograms from women around the world. Images were labeled based on whether or not each woman was diagnosed with breast cancer in the next five years.

Significantly, those labels described what happened after the mammogram, not whether cancer was visible in the image, adds Lehman. “We did not tell it which tissue features to look for,” she says. When the software analyzes a new mammogram, it uses the patterns it identified to estimate individual risk, says Lehman.

How Clairity Breast Rates Risk

• Average: below 1.7 percent
• Intermediate: between 1.7 and 3 percent
• High: 3 percent or higher

The software currently works only with a Hologic screening mammogram, either 2D-only or a combination of 2D and 3D images, says Lehman. The software doesn’t detect or diagnose cancer, and it doesn’t provide care recommendations, says Lehman.

Who Should Get the Test?

The assessment is meant for women who get routine screening mammograms, says Lehman. Eligibility doesn’t depend on family history or breast density, but the test shouldn’t be used by women with known high-risk inherited gene mutations, such as BRCA1 or BRCA2, or by women who already have breast cancer, she says.

“Women with known genetic mutations that put them in the highest risk category, or who have other reasons to start screening at 30, wouldn’t fall into this group,” says Lustberg. “And similar to previous recommendations, patients should have at least a 10-year life expectancy before doing any type of breast screening,” she adds.

The Test Comes With Benefits and Risks

“One of the most common things I have heard from women after a breast cancer diagnosis is, ‘How can this be? No one in my family ever had breast cancer,’” says Lehman. This tool may help identify increased risk in women without a family history, who may benefit from additional screening with breast MRI or from medications that reduce breast cancer risk, she says.

Bao says the new AI tool is interesting but has several limitations, starting with the fact that its only input appears to be a single mammogram. That leaves out family history and a personal history of atypia (abnormal cells found on a breast biopsy), which each play a significant role in future risk, she says. “Omitting these clinical factors from a clinical risk tool is a real limitation,” says Bao.

Lehman says results from Clairity Breast are meant to be considered alongside a woman’s medical history and clinical guidelines.

The tool may also cause unnecessary anxiety if a woman doesn't have a clear plan for what to do next, says Bao. “A high-risk result means a 3 percent or greater chance over five years, which is still a low absolute risk, and most women in that group won’t develop breast cancer in that time,” she says.

“Patients may pursue more frequent mammograms or supplemental imaging such as breast MRI, which can produce false alarms leading to unnecessary biopsies and added expense,” says Bao.

She points out that earlier detection does not necessarily mean improved survival, as most breast cancers are slow-growing. The FDA authorization means the tool is safe, not that it improves outcomes. It still needs to be studied alongside other clinical risk factors, and more evidence is needed that it helps patients in practice, says Bao.

How Much Does Clairity Breast Cost? Does Insurance Cover It?

The assessment is self-pay at $249 and isn’t covered by insurance, says Conklin. It can be paid for with a health savings account (HSA) or funds from a flexible spending account (FSA), she adds. New FDA-authorized technologies often take time to gain coverage, she says.

Any Risk Calculation Should Be Shared With Your Doctor

A low-risk result isn’t a green light to skip screening or ignore symptoms. “An average risk score should never be used to dismiss a symptom or avoid or delay a recommended evaluation,” says Lehman.

“I see this as one additional tool to better understand the risk of breast cancer,” says Lustberg. It’s used alongside other clinical and genetic factors, she says.

A few free online calculators, including the National Cancer Institute’s Breast Cancer Risk Assessment Tool (also called the Gail model) and the Tyrer-Cuzick model, can also estimate your risk based on your personal and family health history.