According to a validation study published in the journal Gut, Mayo Clinic has created an AI model that can identify pancreatic cancer up to three years prior to clinical diagnosis. The AI system, named REDMOD, detects subtle changes in routine CT scans and identified pre-clinical pancreatic ductal adenocarcinoma by spotting what researchers called an ‘invisible’ signature. This was achieved with 88% specificity, correctly identifying patients without the disease.
Ajit Goenka, M.D., a Mayo Clinic Radiologist and Nuclear Medicine Specialist who led the study, emphasized the challenge of detecting pancreatic cancer early enough to make it treatable. “Our inability to see the disease when it is still curable has been our greatest barrier,” he said. REDMOD reliably identifies cancer from a normal-appearing pancreas across diverse clinical settings.
In cases detected over two years before clinical diagnosis, REDMOD was nearly three times more accurate than human radiologists, with 68% accuracy compared to 23%. This development comes as AI advances continue in the realm of pancreatic cancer detection. For instance, PanDx, another AI framework analyzing contrast-enhanced CT scans, recently topped the PANORAMA challenge with an AUROC score of 0.9263.
Mayo Clinic researchers are now working on a prospective clinical study named AI-PACED to evaluate how clinicians can incorporate AI-guided detection for patients at elevated risk. This study aims to transition REDMOD’s laboratory success into practical clinical use.
Early detection is vital, as pancreatic cancer is expected to be the second-leading cause of cancer-related deaths in the U.S. by 2030, primarily due to its late diagnosis—over 85% of patients are diagnosed after it has metastasized. An estimated 67,530 Americans will likely be diagnosed with pancreatic cancer in 2026, and treatment options remain limited once the disease progresses beyond the pancreas. REDMOD originated from Mayo Clinic’s Precure initiative, which focuses on predicting and preventing diseases by identifying early biological changes before symptoms manifest.