Mayo Clinic AI Flags Pancreatic Cancer Years Before Diagnosis

The model has also demonstrated consistent performance across different hospitals, imaging systems, and patient populations, reinforcing its potential for real-world use.

An artificial intelligence system developed by Mayo Clinic has shown that it can identify early pancreatic cancer changes from standard CT scans, often months or years before diagnosis, improving detection rates and offering a meaningful opportunity for earlier intervention and better patient outcomes.

Published in Gut, the study evaluated nearly 2,000 abdominal CT scans from multiple institutions, all initially reported as normal. The AI model, known as REDMOD, detected 73% of cancers in these prediagnostic scans, on average 16 months before clinical diagnosis, nearly doubling the detection rate compared to specialists without AI support.

Its performance was even stronger in scans taken more than two years prior, identifying nearly three times as many early cases.

Unlike traditional imaging analysis, REDMOD uses radiomics to extract hundreds of quantitative features related to tissue texture and structure. This allows it to recognize subtle biological changes in the pancreas that are invisible to the human eye.

The system operates automatically without requiring additional imaging or manual input, making it suitable for integration into routine clinical workflows.

"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, the senior author of the study. "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 model has also demonstrated consistent performance across different hospitals, imaging systems, and patient populations, reinforcing its potential for real-world use.

Researchers are now advancing clinical validation through the AI-PACED study, which will assess how AI-guided detection performs in high-risk patients over time, including those with new-onset diabetes.

This effort is part of Mayo Clinic’s broader Precure initiative, which aims to identify diseases at their earliest biological stages before symptoms appear.

Backed by institutions such as the National Institutes of Health, the study highlights how AI could reshape cancer detection by leveraging existing medical data.

Stay tuned for more such updates on Digital Health News