MICHIGAN — McLaren Health care has partnered with technology company Bunker Hill to launch a new artificial intelligence program designed to screen more than 40,000 patients annually for heart disease. The initiative uses FDA-cleared algorithms to identify early signs of cardiovascular risk by reviewing CT scans performed for unrelated medical reasons.
The program aims to address heart disease, which remains the leading cause of death in the United States. By analyzing existing imaging for incidental findings like calcium buildup in arteries and valves, the technology helps clinicians identify “orphaned” patients who are at risk but have not yet seen a cardiologist.
Dr. Justin Klamaris serves as the executive vice president and chief clinical officer for McLaren Health care. He noted that while heart disease is the leading cause of death in the United States, many cases are preventable through early detection. “But we see so many patients across the state of Michigan who are impacted by cardiovascular disease,” Klamaris said. “And what we know is much of cardiovascular disease is preventable if we could recognize patients who have symptoms or have evidence of disease earlier.”
The program utilizes FDA-cleared algorithms to review imaging studies. Klamaris described the initiative as an extension of existing medical teams. “This project and what we are launching with Bunker Hill... is to use artificial intelligence and use the abilities that we have to extend our clinical care teams beyond the nurses and doctors who are taking care of you when you seek care in our facilities,” Klamaris said.
Nish Kandawala is the CEO and co-founder of Bunker Hill. He explained that the AI model, which was developed at Stanford University, identifies blockages in patients who visit the hospital for unrelated issues like pneumonia or car accidents. “And so what he and his team built was this AI model that can look at scans that are being done on patients for non-cardiovascular reasons and then actually drive them to get preventive cardiology care so that they have better outcomes,” Kandawala said.
Kandawala shared an example from a partnering hospital involving a patient who sought emergency care for a turkey bone stuck in their throat. A CT scan for the bone led the AI to identify a high risk for heart disease. “This program was able to identify that this patient actually had a very high risk for heart disease, for heart attack,” Kandawala said. The patient eventually underwent bypass surgery after the AI finding prompted a visit to a cardiologist.
The health system expects the technology to process a high volume of data. “We are expecting to screen more than 40,000 patients every year,” Kandawala said. “And we project that we will be hopefully finding thousands of patients and driving them to better care.” Officials estimate that 8% to 10% of patients undergoing chest CT scans will be identified as having previously unknown risk factors.
Dr. Sam Kezia, the chief medical director for the McLaren Heart and Vascular Institute, said the AI specifically looks for coronary calcium and aortic valve calcification. “But the ability of AI now to be able to calculate your calcium score automatically and provide this number to your primary care physician would be very, very valuable,” Kezia said. He noted the system creates a “red flag” for patients who may not even know they have a problem.
Once the AI identifies a high-risk patient, McLaren will initiate a notification workflow. This includes sending a traditional letter to the patient and electronic notifications to their primary care physician. If a patient does not have a designated cardiologist, the system will help coordinate care with available providers. Dr. Kezia emphasized that the goal is to be proactive. “So AI is going to be able to identify our citizens, our community patients who are having tests in the system ahead of time and alert them that something needs to be checked out and become proactive to take care of that for them short and long term,” Kezia said.
The initiative also includes a retrospective review of the past year of medical records to find patients with existing risks in the system’s archives. Klamaris noted that McLaren already uses AI in other areas, such as its stroke program to identify blood vessel abnormalities and “ambient AI” for medical documentation during doctor visits.
McLaren Health care plans to launch the program within the next two weeks. The system will initially process a retrospective lookback of imaging from the past year before moving into real-time prospective screening for all patients at McLaren facilities.