Starting Thursday, women across the United States will be able to find out their risk of developing breast cancer in the near future with the help of an A.I. tool that examines a single recent mammogram.
Everlywell, a digital health company, announced it would now offer patients nationwide access to Clairity Breast, an A.I. tool approved by the Food and Drug Administration last year to predict a woman’s five-year risk of breast cancer based on imaging. The test, which was previously available only at select locations in Massachusetts and Colorado, will cost $249. It is not covered by insurance.
Doctors use mammograms to detect the presence of cancer, sometimes with the help of A.I. models that can read the images. But Clairity’s software uses the scans for a different purpose: predicting the risk of developing cancer and providing a probability score that reports the patient’s future risk in the form of a percentage, from zero to 100.
That score, which doctors should consider along with other clinical information like family and patient medical history and other risk factors, could aid in efforts to detect cancers earlier, when they are more treatable and more likely to have better outcomes, said Dr. Eleonora Teplinsky, head of breast and gynecologic medical oncology at Valley Health System in Paramus, N.J.
The broader availability of a technology that can help women understand their breast cancer risk and talk to their doctors about it “has the potential to be practice-changing for us,” said Dr. Teplinsky, who is not involved with the work.
“We’re seeing younger and younger women with breast cancer, many of them diagnosed before they would even have started screening mammography at age 40,” she said.
Most have no telltale signs — like a strong family history of breast cancer — that would indicate higher-than-average risk. In fact, a majority of breast cancers occur in women with no family history of the illness.
“For so many women, it comes out of nowhere,” said Dr. Connie Lehman, founder and chief executive of Clairity, whose research led to the development of the new technology. “We’re trying to identify women at risk for these sporadic breast cancers, and to have fewer women fall through the cracks.”
But there are other factors to consider, including that the technology could lead to unnecessary alarm or medical interventions for people who ultimately won’t develop cancer. Here’s what to know.
How does the test work?
The model detects subtle imaging patterns that are correlated with future cancer development, even in mammograms that appear normal to the human eye. It was developed and trained on hundreds of thousands of mammograms from women around the world, Dr. Lehman said, and its results have been validated in further testing since it was authorized.
Anyone who has had a routine standard screening mammogram within the past 12 months will be able to request the test on the Everlywell website. Typically, mammograms start at 40, though guidelines issued this year introduced A.I.-driven risk assessments at the age of 35 as an option for gauging the disease risk earlier. This change could lead more women to get mammograms at younger ages.
Everlywell will retrieve a patient’s image from the mammography center and then feed it into Clairity’s platform, which produces a digital report on the patient’s risk. The report classifies patients as at average, intermediate or high risk of developing breast cancer over the next five years. It does not detect cancer itself.
Someone with a score below 1.7 percent — meaning less than a 1.7 in 100 chance of developing breast cancer in the next five years — is considered to be at average risk. Patients with a score between 1.7 percent and 3 percent are considered at intermediate elevated risk; those with a score of 3 percent or more are at high risk. (Women who are undergoing cancer treatment or who have genetic mutations that put them at high risk of breast cancer are not eligible.)
Are there potential downsides or risks?
Right now, the technology only assesses risk, and cannot definitively say whether a woman will develop cancer. A low score doesn’t mean that women are absolutely in the clear, and some experts said they worried that patients informed they are at low risk might become complacent and let screening slide.
“The danger is that women will think, ‘I have a low score, I don’t have to worry about it,’” said Dr. Debra Monticciolo, chair of the American College of Radiology’s breast cancer screening leadership group. “That’s the wrong message. If the model says your risk is low, that’s good news, but doesn’t mean you won’t get cancer. You still need screening.”
On the flip side, a high score does not mean cancer is inevitable. But those who are deemed at high risk are likely to step up screenings, and may endure more biopsies and other interventions because of those extra tests.
“People can get hurt in that process,” said Dr. H. Gilbert Welch, a senior researcher at Brigham and Women’s Hospital in Boston.
He noted that more screening also almost inevitably leads to more cancers being detected. But some cancers are slow growing and never become life-threatening. There is no evidence that all the extra care will save lives, he added.
“If you had the test you really wanted from A.I., it would be: Who’s likely to die from breast cancer? Because those are the people you could really help,” Dr. Welch said.
What do patients do with the results?
Depending on the individual’s level of risk, the report delivered to the patient outlines possible next steps. Patients are directed to share the report with a primary care provider, and they can consult with Everlywell clinicians by phone or email.
National guidelines urge women at high risk to undergo more intensive screening and start screening at younger ages, and to consider other interventions that may lower their chances of developing the disease. These include taking drugs that block estrogen receptors or, for those at very high risk, prophylactic surgery like mastectomies. Lifestyle changes — including physical activity, weight management and limiting or eliminating alcohol consumption — may also help curb breast cancer risk.
Karen Hale, 65, was surprised to learn she was at high risk of breast cancer from a Clairity Breast assessment.
“I’d never had a scare, never had any red flags,” said Ms. Hale, a co-chair on the board of the Breast Cancer Research Foundation, which funded Dr. Lehman’s research and provided seed money to Clairity Inc. Though her mother had breast cancer, Ms. Hale got genetic testing that did not identify any known mutations associated with breast cancer risk.
Doctors recommended a low dose of tamoxifen — a drug shown to reduce the risk of hormone receptor-positive breast cancer — as well as alternating M.R.I.s and mammograms every six months.
“Now that I have the information, I can do something about it,” Ms. Hale said.
What other options exist?
Many other risk assessment tools are available for free online, such as the Breast Cancer Risk Assessment Tool, or Gail Model, which assesses five-year and lifetime breast cancer risk based on eight questions about personal and family medical history; and the Tyrer-Cuzick assessment, which asks for a more detailed family history and uses that and other information to predict lifetime risk.
The post Women Can Now Assess Their Breast Cancer Risk With the Help of A.I. appeared first on New York Times.
