gOogle’s AI algorithm to help with breast cancer screening will now be part of commercial mammograms.
On Nov. 28, the company announced that it had licensed its artificial intelligence technology to iCAD, a medical technology company that provides breast cancer detection services to healthcare facilities around the world.
While iCAD already includes AI-based strategies in its cancer screening services, it will now also integrate Google’s algorithm, which Google tested with researchers from Northwestern University. “This is an inflection point for us,” says Greg Corrado, co-founder of the Google Brain team and lead scientist for Google’s AI healthcare team. “We are moving from academic research to being able to deploy our algorithm in the real world.”
In an earlier study published in 2020 in Nature, Google’s algorithm for mammograms performed better than radiologists by recording fewer false positives and false negatives when reading images. The study involved mammograms of more than 91,000 women in the US and UK. In the US, where most women aged 50-74 are recommended to be screened every two years, Google’s system reduced the false positive rate by 6%, and in the UK. , where women between the ages of 50 and 70 are asked to be screened every three years, by 1.2%. The machine learning algorithm also reduced false positives by 9% in the US and almost 3% in the UK.
This benefit will now be commercially available for the first time to the 7,500 mammography sites worldwide, including academic health systems, that use iCAD’s services. While Corrado declined to detail how Google’s algorithm differs from those tested by other researchers and companies in the field, he said the system incorporates data from a wide range of images, even at beyond those of breast tissue, to refine the machine learning process. iCAD and Google will continue to develop and refine the technology under the partnership agreement.
Read more: How AI is changing medical imaging to improve patient care
The algorithm is not designed to replace radiologists, at least not in the short term. But in Europe, according to Stacey Stevens, president and CEO of iCAD, it could help ease the burden on radiologists, as many countries (including the UK) require two readings of a mammogram image. iCAD is working with health regulators to get the proper clearance so the company’s AI-based interpretation can eventually be a part of it, she says. In the US, Stevens expects the first product including Google’s algorithm to roll out in early 2024.
Stevens also anticipates that the AI-based system will bring mammography to more people around the world, especially in low-resource areas that could not support the infrastructure required to house mammography-related hardware. storage of mammography images. With Google’s cloud-based storage capabilities, she says, “we have the ability to expand into new geographies and new regions of the world and scale our tools on more patients in regions of the world constrained by infrastructure challenges”.
As with any machine learning system, the more mammogram data the algorithm feeds, the better it detects the smallest differences that distinguish normal tissue from potentially cancerous tissue. Women receiving mammograms using the AI-based system will have their information fed back into the algorithm, minus any identifying data. Right now, most people getting mammograms probably don’t know that an AI system could be in the background to complement the radiologist, because at this time, no regulatory agency has approved a fully AI-based interpretation of mammograms. But as more AI algorithms like those from Google enter the market, that may change, and radiologists may end up talking to patients about how their images are interpreted.
Read more: How researchers are working to make IVF more effective
Ultimately, such machine-based readings could begin to show patterns that human eyes cannot see. Stevens says iCAD’s current AI-based algorithm already detects the presence of tiny calcifications in breast tissue that scientists are beginning to associate with an increased risk of heart disease. If this association is confirmed, mammograms could also become a tool for assessing the risk of heart disease in women.
For now, adding an AI perspective to mammograms could begin to improve how women’s breast cancer risk is determined. AI systems can better distinguish, for example, differences between specific racial and ethnic groups; in the United States, African American women are at a higher risk of developing more aggressive types of breast cancer and are more likely to die from the disease than other women, so training an AI system to detecting early signs of these cancers could lead to better outcomes. “We find that there are many cases of women with what appears to be a normal mammogram, but there are things in these images that cannot be seen with the human eye,” says Stevens. If these differences can be detected by an AI algorithm, these women could be sent for additional screening to determine if they are at higher risk of developing cancer. This could put them on the path to receiving treatment sooner, which ultimately leads to a better chance of survival. It could also mean cheaper medical services, which translates into cost savings for the healthcare system. “We’re in the early rounds of breast cancer risk assessment with AI,” Stevens says, “but we’re excited about its potential.”
More must-reads from TIME
#Googles #mammograms