Conversations with Dr. Lakshman Tamil, director of the Quality of Life Technology Laboratory at UTD inspired Dr. Tim Cogan BS’15, MS’17, Ph.D.’21 to focus his PhD work in Electrical Engineering on the potential uses of artificial intelligence in medical imaging.
Cogan is now Chief Technology Officer and Tamil is co-founder of MedCognetics, a company that spun out of Cogan’s PhD research at UTD. “We use AI for early cancer screening; right now, specifically for breast cancer, although we’re looking at other areas,” Cogan explained.
Cogan and his colleagues are particularly excited about the potential applications outside of the U.S., where radiologists may be prohibitively expensive or in short supply. He points to a partner organization in India where patients travel a full day or two from their village to get imaged. “And then they’re stuck with this situation where it could take a week or a week and a half to have those images looked at,” he said. “Do you just wait in the city for a whole week before going back to your village? Or if you do decide to go back to your village, do you want to come back?”
MedCognetics’ software means that patients’ images can be reviewed more immediately. “Within a couple of minutes, you could process the case and tell the patient, ‘Yes, you should hang around or no, you’re free to go,’” Cogan said.
For the U.S. market, Cogan emphasized that MedCognetics will not replace radiologists but will instead provide a second read. “I think the best thing is that we have people and AI working together, checking each other’s work,” he added. “Then you get the best possible outcome for the patient.”
Receiving FDA clearance for its breast cancer screening software in December 2022 was a key milestone, according to Cogan. “There are dozens of companies or groups doing work similar to ours, but my understanding is there’s a single digit number of mammography AI groups that have actually have FDA clearance,” he said. “It’s a pretty big hurdle to pass.”
Earlier this year, MedCognetics received a $750,000 grant from the National Institute of Health’s (NIH) Artificial Intelligence/Machine Learning Consortium to Advance Health Equity and Researcher Diversity. “From the start, we’ve really focused on getting data and partnering with doctors everywhere,” Cogan said.
By gathering data from patients all over the world, MedCognetics aims to train its AI and machine learning technology on diverse populations so it works on every patient and every demographic. “If you take your system and you just train it on a particular type of patient population, it’s going to work really well for those particular patients, but it won’t necessarily work well on another group of patients,” Cogan explained. “People in Southeast Asia have a totally different diet and totally different lifestyle than people in say Central Texas, so cancer will manifest differently. You need your system tuned for people all around the world.”
As this technology improves, it might be applied in other medical settings beyond breast cancer screenings, providing greater access to medical imaging around the world.
