Key Takeaways
Pancreatic cancer is relatively rare on a population level but highly fatal because it is usually diagnosed at an advanced stage.
An artificial intelligence model that drew on comprehensive health data from nearly 40,000 patients aimed at identifying subtle clues of pancreatic cancer early.
The model showed a high accuracy to distinguish between people at risk for pancreatic cancer and those with low risk up to three years before diagnosis.
These findings will be presented at the American College of Surgeons Clinical Congress 2026 in Washington, Sept. 26-29.
WASHINGTON (September 25, 2026) — Researchers at Mayo Clinic have designed an artificial intelligence model that can potentially predict an individual’s risk of developing pancreatic cancer years before diagnosis.
The research will be presented at the American College of Surgeons (ACS) Clinical Congress 2026 in Washington, Sept. 26-29, where thousands of surgeons will convene to advance surgical quality, patient safety, and access to care.
Pancreatic cancer is relatively rare but highly deadly, with about 67,000 new diagnoses and 52,000 deaths in 2026, according to the American Cancer Society. Its share of cancer deaths is outsized: pancreatic cancer accounts for about 3% of all new cancers but 8% of all cancer deaths.
“Pancreatic cancer can be curable, but only when we catch it early — and fewer than one in five patients is diagnosed in time,” said study co-author Cornelius Thiels, DO, MBA, FACS, a surgical oncologist at Mayo Clinic in Rochester, Minnesota. “As a result, survival for many patients is still measured in months, not years.”
Unfortunately, universal screening for pancreatic cancer isn’t feasible, Dr. Thiels said, so his team set out to develop an AI model that can identify patients at greatest risk of developing cancer of the pancreas.
“We know that pancreatic cancer forms over five to seven years, but the things that a clinician or patient sees don’t happen until it’s too late,” he said.
The model Dr. Thiels, lead study author Chris Varghese, MBChB, and their team developed used individual patients’ longitudinal health history — essentially the detailed, comprehensive patient information in a patient’s electronic health record to get a full picture of a patient’s health over time — from the Mayo Clinic system. The model combined that data with results of routine laboratory tests obtained over an average of a decade or more.
The study dataset included 6,066 individuals with pancreatic cancer and 33,396 controls with 7.5 to 19 years of clinical histories. The goal was to identify subtle clues that could point to a risk of pancreatic cancer early on, Dr. Thiels said.
To test the model’s effectiveness at predicting pancreatic cancer three years prior to diagnosis, the researchers calculated area under the receiver operating characteristic (AUROC) curve to distinguish between people at risk for pancreatic cancer and those with low risk. The AUROC was 0.853, on a scale where 1.0 would represent perfect discrimination and 0.5 would be no better than chance. The model also showed a strong ability to identify patients truly at risk while limiting false positives, with an area under the precision-recall curve (AUPRC) of 0.712.
The study also showed the model was well calibrated on the calibration curve, a measure of how closely a model’s predicted risk matches what actually happens, with a calibration plot slope of 1.08. “Our model showed that a greater than 50% risk of pancreas cancer predicted by our model indicated an 88% likelihood of being diagnosed with pancreatic cancer in one year,” said Dr. Varghese, a surgical data scientist at Mayo Clinic in Rochester.
“We built this to be as generalizable, scalable, and easy to put into practice as possible,” Dr. Varghese added. The data inputs the model relies on are captured almost universally in hospital systems worldwide, Dr. Varghese said. “If it’s shown to work, it could be used in almost any setting,” he added.
The researchers are deploying the model on a research basis, Dr. Thiels said. “We’re proving that we can move this from a retrospective research tool into our clinical environment and run it prospectively for validation,” he said.
They are also working to further validate the model within Mayo prospectively and, this year, at a non-Mayo system, Dr. Thiels said. “We are also working on developing more advanced machine learning architectures, which appear to improve the performance even more,” he added.
Study co-authors with Dr. Thiels and Dr. Varghese are Leo Yan Li-Han, PhD; Tanios S. Bekaii-Saab, MD; Richa Bisht, MD; Ajit H. Goenka, MD; John D. Halamka, MD, MS; Ellen L. Larson, MD; Frank G. Lee, MD; Michael L. Kendrick, MD, FACS; Shounak Majumder, MD; Hojjat Salehinejad, PhD; and Mark J. Truty, MD, MS.
Disclosures: Authors have no disclosures to report.
Citation: Varghese C, et al. Enabling Digital Screening for Pancreatic Cancer using Artificial Intelligence Analysis of Disease Trajectories . Scientific Forum, American College of Surgeons (ACS) Clinical Congress 2026.
Note: Research abstracts presented at the ACS Clinical Congress Scientific Forum are reviewed and selected by a program committee but are not yet peer reviewed.
# # #
About the American College of Surgeons
The American College of Surgeons is a scientific and educational organization of surgeons that was founded in 1913 to raise the standards of surgical practice and improve the quality of care for all surgical patients. The College is dedicated to the ethical and competent practice of surgery. Its achievements have significantly influenced the course of scientific surgery in America and have established it as an important advocate for all surgical patients. The College has approximately 95,000 members and is the largest organization of surgeons in the world. "FACS" designates that a surgeon is a Fellow of the American College of Surgeons.
Follow the ACS on social media: X | Instagram | YouTube | LinkedIn | Facebook
Enabling Digital Screening for Pancreatic Cancer using Artificial Intelligence Analysis of Disease Trajectories.
25-Sep-2026