AI tool addresses women’s heart health during mammogram

Cardiovascular disease is the number one killer of women. It is underdiagnosed and undertreated. While men have a standardized diagnosis protocol that helps catch things earlier, women don’t. Women frequently don’t recognize the symptoms of heart disease–until they’re admitted to the emergency room or to cardiac ICU.

Genexia, a company that offers AI-powered solutions, has turned its attention to this problem and created a tool for catching heart disease risk during a routine mammogram. We spoke to Dino Martis, Genexia’s CEO, and Dr. Kelly Cohen, co-founder and Brian H. Rowe Endowed Chair in Aerospace Engineering at University of Cincinnati, about the work they’re doing in the space of women’s heart health.

This is an interesting application of your company’s expertise. How are you making it happen?

Martis: We studied the problem and we realized that we had to provide a solution through an easily accessible test. Kelly and his brilliant team of data scientists and the radiologists–we have full radiologists on staff–have come together to identify cardiovascular risk during routine screening mammography, with no additional equipment or images necessary. The technology is complex, obviously, but the outcome to the radiologist is very simple. It is given to the radiologist in a report format and with DICOM images.

As with breast cancer, your test seeks to catch a condition early enough to cure?

Martis: If you identify early, it is easily controlled. Cardiovascular disease is controlled sometimes with just lifestyle changes or taking statins. You don’t have to have surgeries, very complex surgeries where it degrades your quality of life. A person is not the same after a heart attack as before.

How does heart disease manifest in a mammogram?

Dr. Cohen: We look at skin thickening and breast arterial calcification. Breast arterial calcification is different from breast calcification. Breast arterial calcification can be mapped to coronary artery calcification. And thereby establish that link to cardiovascular risk. This is a tool for the radiologist to add to his own interpretation of diagnosis, and further tests can be ordered.

It’s a tool that uses artificial intelligence.

Martis: Yes. And you have to train the AI through a database of images. We currently own a repository of 35,000 images. We want to increase that and collect a database of coronary calcium CT scans and CPR symbiotic mammography. To get FDA approval, we need a very large statistical database. In fact, we’re embarking on an Indiegogo campaign on June 21 to raise money for growing those databases.

Dr. Cohen: The power of AI is its ability to look at millions and millions of images and patterns. Then we can say with a high degree of accuracy and confidence, this person is at risk. Therefore, we need to channel them through the clinical procedures that follow identification of risk and thereby save their lives, improve their quality of life. Our technology is good enough to win awards. We won an award at a Swedish International Conference for our work. I am confident in the tech. The question is, do we have access to the amount of data that we can then work with?

Dr. Cohen, your work on this has personal meaning, correct?

Dr. Cohen: Yes, I had a heart attack in 2018 while hosting an event in Orlando. My wife called an ambulance against my judgement and it saved my life—there was a 100% blockage and I had to have a stent operation. Later on, an opportunity came up at the UC Innovation Hub where I met with Dino and looked at where our AI could be applicable. I’m invested emotionally and technically in coming up with this solution.

For me, this isn’t just another equation I have to solve. I’ve been given a second chance. After being a successful engineer for 35 years, I want to devote the next 10 years or so of my professional career to making an impact and help solve this problem.