Hannah Huth is completing her final year of medical school at the University of Tennessee Health Science Center
I'm currently a medical student studying to become a pediatric oncologist. I came to NIH through the Medical Research Scholars Program. It's a year-long research immersion program for future clinician-scientists.
While I was here, I joined a research project led by James Anibal. He is an NIH Oxford Cambridge Scholar who is completing his PhD at Oxford University. James has such an incredible and unmatched understanding of AI and the computational side of it, so I was thrilled to be able to bring a deeper understanding of the clinical aspect to this project.
Overall, the project is a big one. It aims to use AI and large language models to analyze people's voice recordings on a mobile phone application to screen for disease. That's the big picture goal, but that is still a ways off. At the moment, we're focused on the data collection components. I've been in charge of looking at how we might design the data collection and the clinical partnerships we might form to make that successful. It's still a work in progress. But we have trials currently underway in Vietnam, Rwanda, Baltimore, Washington, D.C., and Memphis, Tenn.
In medicine, we used to play catch up with disease and put all our money and research into therapeutics. Now there is a massive emphasis on biomarkers, early ways to detect diseases. Most biomarkers are being evaluated in urine or blood samples with patients who have a higher socioeconomic status and live closer to high revenue hospital systems.
So it becomes really exciting when we start to think about the ways we can do this from an iPhone—inexpensively and at scale. We want this app to be a tool for people around the world who can't get to the hospital easily or can't afford to. People who might be sitting at home with diabetes or Parkinson's or Alzheimer's or other diseases or viruses and not know it. No one ever thought that diabetes would cause changes in your voice. That sounds crazy, but it's true. So what other diagnoses can we look at that might have these implications?
My start in medicine came from being a patient. I've had four carniopharyngioma brain tumors. I was misdiagnosed and experienced delayed treatment. My family had no medical experience or expertise at all. Back then, AI didn't exist. But now we can use AI to facilitate diagnostics and that overall patient experience, within and outside the walls of the hospital. That's been my approach in my very short career so far in medicine—to see a problem clinically and solve it with technology.
- Interview by Sean Markey