An artificial intelligence (AI) software that could possibly revolutionise pancreatic cancer treatment by early diagnosis, even before the tumor becomes visible on a scan, is currently being evaluated in a clinical trial.

A Mayo Clinic AI tool detected abnormalities on scans up to three years before the patients were diagnosed.

According to a research paper published in the journal Gut, an AI model developed by a Minnesota-based clinic was trained with CT scans of patients who were earlier being examined for other diseases but were later diagnosed with pancreatic cancer.

The research team then asked radiologists and AI software to review the scans and found the AI system to be three times more effective at detecting early signs of the disease compared to radiologists.

A radiologist at Mayo Clinic and an author of the study said they knew pancreatic cancer does not come all of a sudden in three months, adding, “We knew that signal was there. We just needed to find a way to be able to detect it.”

Pancreatic cancer has a five-year survival rate of around 13% and is on its way to become the second leading cause of cancer deaths by 2030; however, if successful, the deployment of Mayo Clinic’s AI software could improve early detection and treatment outcomes.

Early detection for the disease has been rare so far, but the Stage I patients had a survival rate of up to 40%.