Study Confirms AI Can Identify Pancreatic Cancer Much Earlier Than Diagnosis

A silent killer may finally be losing its invisibility. Researchers at Mayo Clinic have developed an artificial intelligence tool that can detect pancreatic cancer years before it is typically diagnosed-potentially transforming outcomes for one of the deadliest cancers.

Published in Gut, the study introduces a system called the Radiomics-based Early Detection Model (REDMOD). Unlike traditional methods that rely on visible tumors, this AI scans routine abdominal CT images to identify subtle, early changes in pancreatic tissue—long before a mass can be seen.

In testing, REDMOD analyzed nearly 2,000 CT scans, including images from patients who were later diagnosed with pancreatic cancer but whose scans had originally been read as normal. The results were striking: the AI detected 73% of these hidden cancers at a median of 16 months before diagnosis—almost twice the detection rate of specialists reviewing the same scans without AI support.

Its advantage was even more pronounced at earlier stages. In scans taken more than two years before diagnosis, the system identified nearly three times as many cancers that would otherwise go unnoticed. This matters because pancreatic cancer is rarely caught early; more than 85% of patients are diagnosed after the disease has already spread, when treatment options are limited.

The technology works by measuring hundreds of imaging features—tiny variations in tissue texture and structure—that signal early biological changes. Importantly, it can run automatically on scans already performed for other reasons, making it especially useful for monitoring high-risk individuals, such as those with new-onset diabetes.

While further validation is needed, this breakthrough points to a future where pancreatic cancer is no longer detected too late—shifting the focus from late-stage treatment to early, potentially life-saving intervention.

REFERENCE: Mukherjee S, Antony A, Patnam NG, et al; Next-generation AI for visually occult pancreatic cancer detection in a low-prevalence setting with longitudinal stability and multi-institutional generalisability; Gut; doi: 10.1136/gutjnl-2025-337266