AI Meets Heart Health
Cardiologist Amogh Karnik (MSAI ‘24) has a new opportunity to combine his patient care with AI innovation, including a chatbot he developed to help patients understand test results and plan their next steps.
A patient’s test results come back with reason for concern–an elevated cholesterol level with potential cardiac complications.
Amogh Karnik (MSAI '24), a cardiologist at Northwestern Medicine, is spending late nights and long weekends developing AI systems to interact with patients to anticipate and answer their questions about what comes next.
This type of non-traditional work for a practicing physician is powered by what Karnik learned throughNorthwestern Engineering's Master of Science in Artificial Intelligence (MSAI) program. Karnik is a recent graduate of the MSAI+X program, an offshoot of the traditional offering geared toward those with advanced degrees such as medical doctorates.
“I was always interested in merging the two areas of engineering and medicine,” Karnik said. “That's what led me to pursue MSAI as part of my cardiology fellowship training.”
Now that training stands to benefit cardiac patients in Northwestern Medicine’s system.
Karnik sees patients during the day and works on his AI projects on his own time. In September, however, he transferred from Northwestern Memorial Hospital to the system’s Palos Hospital in Palos Heights, Illinois, a smaller facility that tests new approaches before launching successful ones systemwide.
He spends four days seeing patients at Palos and has one protected day for his AI research. That work to integrate AI into the doctor-patient relationship comes at a time of growing trust between humans and medically empowered technology. Right now, nearly 40 percent of Americans surveyed by research consultancy Censuswide said they trust tools such as ChatGPT in navigating their healthcare decisions.
That number is actually higher—52 percent—for those aged 45 to 54. This age range matches the ages people most frequently develop heart issues.
Cardiac patients undergo a plethora of tests, the results of which provide hundreds and sometimes thousands of data points that can create a mountain of information no human can efficiently rifle through to accurately and effectively spot trends. AI excels at doing exactly that.
The large language model Karnik built interprets patients’ lab test results and offers them the opportunity to engage with a chatbot to review them and discuss how they can improve their cardiovascular health.
Karnik credited Tyler Smith (MSAI '24), an AI fellow in Northwestern's Bluhm Cardiovascular Institute, and Jonathan Hourmozdi (MSAI '24), an advanced heart failure cardiologist at Northwestern Medicine, with helping the chatbot interpret test results.
“We have so much data that we collect on patients, and so that makes it a very rich source of material for machine learning and AI,” he said. “That's what this field is about: How do you tell a story with the data that you have?”
That story, personalized to each patient, becomes one that can increasingly be told well by AI systems like the ones Karnik is creating.
“There are not a lot of places and people doing this type of work,” he said. “It’s really exciting to be at the forefront and be thinking about these things, not just from an academic standpoint but from an actual clinical patient care standpoint. These are things that can actually improve how we care for patients. I think it's really important.”
