A Penn Medicine AI model called Percival finds hidden disease in CT scans ordered for other reasons.

Unlike most tools, it does not need to be told which condition to look for.

In tests it found traces of atrial fibrillation and chronic kidney disease in routine scans.

It flagged pancreatic cancer, cirrhosis and bacterial pneumonia more reliably than TotalSegmentator.

The approach is called opportunistic screening: finding results beyond a scan’s original purpose.

Percival links scans with lab values, physiology and outcomes to predict future health risks.

Co-author Charles Kahn says it could anticipate major heart and liver complications early.

Read the full article for more details on the Percival AI model.