So, you’re about to get your annual mammogram when your breast imaging clinic presents you with a new option that you’ve never heard of before – would you like to add an artificial intelligence (AI) analysis to your mammogram?
Like many others, you probably have a lot of questions about this new tool. How exactly is AI used in screening mammography, and is it worth adding to your annual screening?
Across the country, some breast imaging clinics are beginning to offer women the option of adding an AI-assisted reading to their mammograms (often for an additional cost). While mammogram images are traditionally double read by two radiologists, this new option replaces one of the readings by a radiologist using AI software.
Is my regular mammogram screening not accurate enough?
According to the National Cancer Institute, mammograms correctly identify about 87% of breast cancers present at the time of screening. Researchers are investigating whether AI software can help radiologists detect breast cancer with even more accuracy.
While some European studies have shown improved breast cancer detection through AI-assisted mammograms, it is unknown whether these results will be the same in the U.S. or will benefit everyone in the diverse U.S. population. Today, there are no published clinical trial results that show AI-assisted mammograms perform better than the current process. While more research is still needed, scientists are optimistic about these early results.
Could AI software replace my radiologist?
AI alone cannot replace the perspective and experience of a radiologist or a substitute for the unique insight that comes from human interactions. However, as a second set of eyes, AI software could help radiologists read more screenings. AI software also has the potential to detect abnormalities faster, which could help radiologists prioritize their time reviewing suspicious results.
How much does an AI-assisted mammogram cost? Is it covered by insurance?
Because AI-assisted mammograms are relatively new in mammography screening, there is no code for radiologists to bill insurance companies for them. In response, some clinics are asking people to pay an additional $40 to $100 fee for an AI reading. As the body of evidence grows regarding the value of AI-assisted screening mammograms, they may become standard of care in the future, so insurance companies may be more likely to reimburse their cost in the future.
Could an AI-assisted mammogram benefit me?
Some studies show that some AI-assisted software may be able to detect some cancers earlier than radiologists. AI may also be able to “see” more details in difficult readings like highly dense breast tissue. However, AI-assisted mammograms may also have false positives where the software incorrectly “sees” breast cancer that isn’t there, which might lead to additional stress and costs. This technology is still under study, and over time, we hope to have a clearer understanding about its value and how it may be used in the mammography screening setting.
Can everyone access AI-assisted mammograms?
One potential drawback of AI-assisted mammograms is that any tool by any one company may only benefit certain populations. AI software is trained to recognize different breast cancers based on specific datasets that may not be representative of a broad and diverse community. Another possible barrier is that not everyone can afford the extra cost of AI-assisted mammograms. This technology is also only available in certain geographic areas, so many who might benefit from the use of AI-assisted mammograms do not have access to them.
Conclusion
Komen supports new advancements in breast cancer research that provide high-quality breast cancer screenings, like AI-assisted mammograms, but they must be supported by evidence and oversight to make sure everyone can benefit from them equally.
As a community, we must be careful that use of AI doesn’t deepen disparities or create additional barriers to high-quality care due to increased costs, limited availability, lack of regulation and standardization, or use of datasets that don’t represent the populations they serve. While AI holds potential to improve the screening and early detection of breast cancer, its promise is still unfolding.
If you have any questions or would like to learn more about AI-assisted mammograms, visit the Emerging Areas in Early Detection page on komen.org.
