[ad_1]
Emory University’s Dr. Hari Trevedi, an associate professor in the Department of Radiology and Imaging Sciences and the Department of Biomedical Informatics, has helped connect the dots between breast tissue and cardiovascular risk by using an AI model.
Dr. Trivedi served as the corresponding author and lead investigator on a retrospective study that used data from 2013 to 2024. The study began in 2025 and was published earlier this year. It included about 120,000 initial mammogram screenings from patients at Emory and the Mayo Clinic in Phoenix, Arizona. “We applied our AI model to calculate their breast arterial calcification (BAC) score, which we used to make future predictions to study cardiovascular risk,” Dr. Trivedi says.
Dr. Trivedi says that the concept of BAC and its importance to cardiovascular health has been known for 30 years. “The concept is not particularly novel,” he says. “But what was novel about our work is that until now, there was no standardized way to measure how much there was or how severe it was.”
BAC is visible to radiologists on scans, but only about 15 percent of radiologists traditionally report BAC in mammogram results. “Over 80 percent of radiologists don’t mention it, despite there being a body of work showing it can predict cardiovascular risk. This removes the burden from radiologists of having to locate and quantify severity.”
The team has automated the entire process, and the model is now capable of quantifying the amount of BAC. “That is something a human cannot do. They would be counting pixels in the image,” he says. By implementing the model, the hope is that mammograms can be used to detect cardiovascular risk. This could be especially useful to women ages 40 to 50 who get mammograms but who are not yet getting cardiovascular screenings.
The technology is already deployed clinically at Mayo Clinic, and the team aims to deploy it at Emory later this year after FDA clearance.
Advertisement
[ad_2]
Source link
