Percival AI can uncover unseen health issues in CT scans
When analyzed by a new artificial intelligence model developed at Penn Medicine, CT scans reveal much about a patient's health beyond the specific condition they were originally ordered for, according to a new study published in npj Digital Medicine. The AI model, dubbed Percival, can uncover nearly hidden patterns in disease and potential future health risks that might have otherwise gone…
A new artificial intelligence model called Percival, developed at Penn Medicine, can detect hidden health issues in CT scans beyond the condition for which the scan was originally ordered, according to a recent study. Percival, named after a knight of King Arthur's Round Table and radiologist Eugene Percival Pendergrass, was trained using data from the Penn Medicine BioBank, which contains information from over 350,000 individuals collected over two decades.
The model can uncover patterns in disease and potential future health risks that may have otherwise gone unnoticed by human eyes. Researchers found that Percival can detect several chronic conditions, including atrial fibrillation and chronic kidney disease, and it outperformed a similar AI model called TotalSegmentator in identifying pancreatic cancer, cirrhosis, and bacterial pneumonia.
The developers hope to test Percival in clinics early next year, believing that routine imaging may provide opportunities for earlier recognition and intervention in patients with pre-existing heart or metabolic diseases. The AI model's code and training specifications are open source, which could help improve its performance and robustness.
Written by urgent.news from Medical Xpress's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
