Mayo AI Spots Pancreatic Cancer 3 Years Earlier

A new artificial intelligence (AI) tool developed by Mayo Clinic researchers may be able to detect pancreatic cancer up to three years before it is typically diagnosed — offering hope against one of the deadliest forms of cancer.

The study, published in the journal Gut, found that the AI system can identify subtle warning signs on routine CT scans long before tumors are visible and when treatment may still be effective.

Pancreatic cancer is notoriously difficult to catch early. More than 85% of patients are diagnosed only after the disease has spread, and the five-year survival rate remains below 15%, according to the National Cancer Institute. Experts project it could become the second-leading cause of cancer-related death in the U.S. by 2030.

Researchers analyzed nearly 2,000 abdominal CT scans, including images from patients who were later diagnosed with pancreatic cancer—even though the scans had originally been read as normal.

The AI model, known as the Radiomics-based Early Detection Model (REDMOD), identified 73% of these early cancers an average of about 16 months before diagnosis. That’s nearly double the detection rate of specialists reviewing the same scans without AI assistance.

The advantage was even greater in earlier scans. In images taken more than two years before diagnosis, the AI detected nearly three times as many cancers that would have otherwise gone unnoticed.

"The greatest barrier to saving lives from pancreatic cancer has been our inability to see the disease when it is still curable," says Ajit Goenka, M.D., the study's senior author, and a Mayo Clinic radiologist and nuclear medicine specialist. "This AI can now identify the signature of cancer from a normal-appearing pancreas, and it can do so reliably over time and across diverse clinical settings."

The system works by analyzing hundreds of subtle features in CT images, including changes in tissue texture and structure that may signal early cancer development. Importantly, it can evaluate scans that were originally performed for other medical reasons — without requiring additional testing.

Researchers say the tool could be especially useful for people at higher risk, such as those with newly diagnosed diabetes, by flagging potential problems before symptoms begin.

The AI model was tested across multiple hospitals, imaging systems, and scanning protocols, and showed consistent performance. It also produced stable results over time in patients who had multiple scans, suggesting it could be used to monitor changes and catch cancer earlier.

The research team is now moving forward with clinical testing through a study called AI-PACED, which will explore how doctors can integrate the technology into routine care and evaluate its real-world impact.

Experts say the findings represent a major step toward earlier detection — something that could significantly improve survival rates for pancreatic cancer.