Penn Medicine AI trained on 400K CT scans spots hidden disease

Advertisement

Philadelphia-based Penn Medicine has developed an AI model that detects signs of disease and future health risks in CT scans beyond the condition each scan was ordered to assess, according to a study published in NPJ Digital Medicine.

The model, named Percival, was trained on more than 400,000 CT images from nearly 52,000 patients in the Penn Medicine BioBank. The biobank holds data on more than 350,000 people collected across the University of Pennsylvania Health System over two decades.

Unlike models that must be directed to look for specific conditions, Percival runs in the background. Researchers found it detected traces of conditions not typically diagnosed by CT, including atrial fibrillation and chronic kidney disease. It also outperformed two other AI imaging models on 279 conditions in chest CT and 230 in abdomen/pelvis CT, with its widest margins on lung cancer and acute lymphoid leukemia.

“We’re hoping to elevate ‘opportunistic screening,’” said senior author Walter Witschey, PhD, associate professor of radiology and associate vice chair of research IT, in an October news release. Dr. Witschey said the goal is to catch problems outside a scan’s original purpose, such as early signs of cancer on an image ordered for pneumonia.

Co-author Charles Kahn, MD, a professor of radiology, said many patients already have significant heart or metabolic disease by the time they reach specialty care. Dr. Kahn said routine imaging could help clinicians recognize and treat those conditions earlier.

Percival remains in preclinical testing, and the authors noted its broader disease and risk findings have not yet been validated on data from outside Penn Medicine. Developers hope to test it in early 2027 with the goal of regular clinical use, and expect their experience deploying AInSights, another imaging AI tool Penn Medicine built in-house, to help. Percival’s code is open source.

The $11.9 billion academic health system has also built an AI model that reads cardiac MRI scans, adding to a growing portfolio of homegrown imaging AI.

Becker’s 3rd Annual Fall Chief Pharmacy Officer Summit is the top tier event that the people leading the future of pharmacy convene. This 3-day event is built specifically for Chief Pharmacy Officers and senior pharmacy leaders. The sessions examine the workforce, financial and clinical priorities shaping pharmacy today. Becker’s is where candid discussions flourish and new ideas thrive. Becker's 3rd Annual Fall Chief Pharmacy Officer Summit

Advertisement

Next Up in Artificial Intelligence

Advertisement