A new study suggests artificial intelligence may be able to identify early warning signs of pancreatic cancer long before tumors can actually be seen through standard imaging, potentially opening a path to earlier and more effective treatment.

Researchers at the Mayo Clinic in Rochester, Minnesota, have developed an AI system that was able to spot subtle abnormalities on CT scans up to three years before patients received a pancreatic cancer diagnosis, according to findings published in the journal Gut. The system is now undergoing evaluation in a clinical trial.

To build the model, scientists trained it on CT scans from patients who originally underwent imaging for unrelated health concerns but were later diagnosed with pancreatic cancer. Radiologists then reviewed the same scans, and their ability to detect early indicators was compared against the AI system. The results showed the AI outperformed human experts by a factor of three in identifying early warning signs.

Dr. Ajit Goenka, a radiologist at the Mayo Clinic and one of the study’s authors, explained the reasoning behind the research direction:
“We knew, based on the biology of the disease, that this is not something which is coming all of a sudden in three months.” Goenka told NBC. “We knew that the signal was there. We just needed to find a way to be able to detect it,”

Pancreatic cancer remains one of the deadliest forms of cancer, with a five-year survival rate of just 13%. It is projected to become the second-leading cause of cancer-related deaths by 2030. Most patients roughly 80% are diagnosed only after the disease has already spread, limiting treatment options.

Unlike more commonly screened cancers such as breast or colon cancer, there is no standard screening test for pancreatic cancer in healthy individuals. The pancreas is also difficult to examine physically due to its deep abdominal location. Early symptoms like weight loss or abdominal pain typically appear only after the cancer has progressed significantly.

Doctors say that early-stage changes are extremely subtle and often invisible on imaging until the disease advances. In many cases, scans may appear completely normal just months before diagnosis.

“I analyze these images every day,” said Dr. Daniel Jeong, a diagnostic radiologist at Moffitt Cancer Center in Tampa, Florida, who was not involved in the Mayo Clinic research and told NBC News. “We’re really looking for a measurable mass that could represent the cancer. So these tumors need to grow to a certain level to become visible.”

Researchers behind the AI model believe one of the signals it detects involves abnormal cellular activity in the pancreas that helps shield cancer from immune system detection, something previously known in theory but difficult to observe on scans.

Goenka noted that the technology may be especially useful for high-risk individuals, such as those with genetic predisposition or diabetes, even before symptoms appear. In such cases, additional testing like blood work or advanced imaging could follow a positive AI signal.

“Unfortunately, if they have symptoms and if it’s truly pancreas cancer, you don’t need AI for that,” he said.

The AI system is part of a broader wave of innovation in pancreatic cancer research. Recent developments include an experimental mRNA vaccine that showed early survival benefits in a small patient group, as well as a targeted drug, daraxonrasib, which has been shown to extend life expectancy in later-stage trials. Regulators have also begun allowing expanded access to the drug in controlled settings for patients who have exhausted other options.

At the same time, scientists are working on advanced blood-based tests designed to detect early biological markers of the disease before symptoms emerge.

“We’re making, actually, major strides. It hasn’t turned this disease around,” Dr. Tamas Gonda, director of the pancreatic disease program at NYU Langone’s Perlmutter Cancer Center told NBC News.

Experts say earlier detection could significantly expand treatment options, allowing more patients to qualify for surgery, chemotherapy, or radiation before the disease spreads.

“This really could be a game changer for us for early detection,” Dr. Pam Hodul, a surgical oncologist at Moffitt Cancer Center, told NBC News.

AI is also gaining traction in other areas of medicine. In a separate study published in Science, researchers found that AI systems matched or exceeded physicians in diagnosing emergency room cases, further highlighting its diagnostic potential.

Despite the promising results, researchers caution that widespread clinical use is still years away. The ongoing trial will need to track patients over several years to determine how accurately the model predicts future cancer development.

“In a disease where we have been just wandering in darkness for decades, this is a milestone that shows us the finish line, but we still have to get to the finish line,” he said.