Mayo Clinic AI Model Detects Pancreatic Cancer Up to Three Years Early

Mayo Clinic announced its Radiomics-based Early Detection Model (REDMOD) can identify pancreatic cancer on routine abdominal CT scans up to three years before diagnosis. The finding, published in *Gut*, comes from analysis of roughly 2,000 scans previously read as normal.

KMSP - Fox 9 - MinneapolisThe National NewsKARE11AI Spots ‘Invisible’ Pancreatic Cancer Three Years Before Doctors CanAI model detects pancreatic cancer earlier and more accurately, study finds

Pancreatic ductal adenocarcinoma has a five-year survival under 15 % because most tumors are detected after spreading. It is projected to become the U.S. second-leading cause of cancer death by 2030, driving demand for earlier detection.

AI model detects pancreatic cancer earlier and more accurately, study finds

The senior author is Dr. Ajit Goenka, a Mayo Clinic radiologist and nuclear-medicine specialist. Funding came from the NIH, the Hoveida Family Foundation, the Mayo Clinic Comprehensive Cancer Center, and the Funk-Zitiello Foundation’s Champions for Hope Pancreas Cancer Research Program.

The National NewsKMSP - Fox 9 - MinneapolisKARE11AI model detects pancreatic cancer earlier and more accurately, study findsAI Spots ‘Invisible’ Pancreatic Cancer Three Years Before Doctors Can

REDMOD detected pre-diagnostic cancer in 73 % of cases at a median of 16 months before diagnosis—about twice the 39 % rate of radiologists. For scans >2 years prior, detection rose to 68 % versus 23 %. Sensitivity was 73 % vs 39 %; specificity ranged 81-87.5 % across independent cohorts of 539 and 80 patients. Predictions were stable in 90-92 % of repeat scans. Detection windows were 40 % (3-12 months), 35 % (12-24 months) and 25 % (>24 months).

AI model detects pancreatic cancer earlier and more accurately, study finds

Modeling suggests that raising the proportion of localized tumors from 10 % to 50 % could more than double survival. By flagging high-risk patients—especially those with new-onset diabetes or unexplained weight loss—REDMOD could enable curative surgery for cases that would otherwise be metastatic.

AI model detects pancreatic cancer earlier and more accurately, study finds

Mayo Clinic’s press release states REDMOD runs automatically, needs no manual image preparation, and is built for integration into radiology workflows as part of the Precure initiative. Researchers note that prospective validation in high-risk cohorts is required, the study lacked ethnic diversity, and specificity varies (81-87.5 %) across datasets, with real-world screening outcomes still unknown.

AI model detects pancreatic cancer earlier and more accurately, study finds

• “The greatest barrier to saving lives from pancreatic cancer has been our inability to see the disease when it is still curable.” — Ajit Goenka, M.D.

KMSP - Fox 9 - Minneapolis

• “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.” — Ajit Goenka, M.D.

KMSP - Fox 9 - Minneapolis

• “This temporal window holds profound significance, as attaining such early detection would substantially augment the probability of cure and improved survival.” — Study authors

AI Spots ‘Invisible’ Pancreatic Cancer Three Years Before Doctors Can

• “While prospective validation is paramount to confirm clinical utility, the REDMOD framework represents a significant advance towards shifting the paradigm for sporadic pancreatic ductal adenocarcinoma from a late-stage symptomatic diagnosis to proactive pre-clinical interception, offering tangible hope for improving outcomes in this challenging disease.” — Study authors

AI model detects pancreatic cancer earlier and more accurately, study finds

Mayo Clinic will conduct prospective trials in patients with new-onset diabetes and other high-risk profiles, aiming to embed REDMOD into routine CT interpretation pending regulatory approval.

AI Spots ‘Invisible’ Pancreatic Cancer Three Years Before Doctors Can