AI may help diagnose early pancreas cancer, Elizabeth Tracey reports


AI is offering hope for making earlier diagnoses in pancreas cancer, historically diagnosed later in the disease when treatment outcomes may not be as good. William Nelson, director of the Kimmel Cancer Center at Johns Hopkins, describes the findings.

Nelson: They trained to be able to recognize seven different types of tumors. How well did it work on another data set of 6200 subjects and another data set that was substantially larger? They compared this to radiologists with the caveat that they're not trained to look for these tumors on non contrast scans. They had a sensitivity of almost 93% in a high level specificity, which outperformed the human radiologist. They were a little bit less sensitive when the lesions were smaller so it's actually quite interesting they're probably going to prove to be helpful.   :33

Nelson notes that imaging is low hanging fruit when it comes to employing AI in medicine, with computers usually providing valuable insights into diagnosis once they are trained properly. At Johns Hopkins, I’m Elizabeth Tracey.