In a world where mammograms are typically just about breast cancer, a new twist has emerged: AI is now stepping up to predict heart disease risk from those very same images.
Yes, you heard that right. Those routine screenings that most women dread could actually reveal more than just lumps and bumps. Enter deep learning algorithms, which are analyzing mammographic features alongside the patient’s age. The result? A concordance index of 0.72, not too shabby when stacked against traditional cardiovascular risk equations, which score between 0.76 and 0.79.
Routine mammograms might soon do more than check for breast cancer; AI is using them to assess heart disease risk, too.
The beauty of this new approach is that it doesn’t need a pile of clinical data. Nope. The first algorithm developed relies exclusively on mammographic features. It’s like a two-for-one deal: screening for breast cancer and evaluating heart risk, all while you’re in that uncomfortable gown.
In a study of nearly 50,000 women, 3,392 experienced significant cardiovascular events during follow-up, proving that this new method isn’t just a flashy gimmick.
Let’s talk about breast arterial calcification (BAC). It shows up as bright pixels on mammograms. AI models are now smart enough to segment these calcified vessels automatically. No need to focus on BAC alone, either. The new algorithms consider a range of mammographic features, making them more effective—especially for women under 80. For those over 80? Well, good luck. Additionally, the AI model developed at Emory Healthcare is trained on data from over 56,000 patients, enhancing its accuracy in assessing cardiovascular risks.
Integrating these AI models into routine screenings is straightforward. No extra lab tests or medical history required. This innovative approach could help address the underdiagnosis and undertreatment of cardiovascular disease in women.
Imagine this: about 40 million mammograms are conducted annually in the U.S. This makes it a goldmine for identifying cardiovascular risks. Women flagged as moderate or high risk can be nudged toward their primary care docs for further evaluation. Talk about a public health win. It’s about time we used technology to tackle more than just breast cancer.
