AI detects signs of pancreatic cancer up to three years before diagnosis, points to a study by Mayo Clinic

An artificial intelligence model developed by Mayo Clinic detected subtle signs of pancreatic cancer in computed tomography up to three years before the formal diagnosis, when tumors were not yet visible. The discovery, published in a recent study and now under clinical trial, may represent a decisive advance in the fight against a disease that kills more than 90% of patients in up to five years.

How does AI identify signs of pancreatic cancer before the tumor is formed?

The model was trained with thousands of tomography scans of patients who later developed pancreatic cancer and healthy individuals. AI has learned to recognize microscopic patterns — such as small changes in tissue density or in the architecture of pancreatic ducts — that go unnoticed to the human eye. These invisible biomarkers appear years before any nodule is detectable.

According to the researchers, the system achieved a rate of correctness greater than 80% in the identification of future cases, even in tests performed for other clinical reasons. This means that an abdominal CT scan to investigate a biliary problem, for example, could alert the doctor to an increased risk of pancreatic cancer.

Why is early detection of pancreatic cancer crucial for survival?

Pancreatic cancer is one of the most aggressive because it is usually diagnosed in advanced stages, when surgery is no longer possible. When detected early, the five-year survival rate jumps from less than 10% to more than 80%. However, there are no routine tests recommended for the general population, and the initial symptoms — such as abdominal pain, weight loss, and jaundice — are vague and easily confused with other conditions.

Mayo Clinic technology can fill that gap. When applied to tests already performed for other reasons, it would function as an alert filter, without additional cost to the patient or overload in the health system.

What does the clinical trial seek to confirm about AI in pancreatic cancer?

The model is being tested in a multicenter clinical trial in the United States, which evaluates its effectiveness in real hospital environments. The objective is to verify that AI reduces the number of late diagnoses without exceedingly increasing false positives — which could generate unnecessary biopsies and anxiety in patients. Preliminary results should be in up to two years.

If approved, the system can be integrated into medical imaging analysis software already used in hospitals. Mayo Clinic has already started talks with regulatory agencies to define the next steps.

Frequently Asked Questions about AI and detection of pancreatic cancer

When will technology be available to the public?

There's no date for commercial release yet. The model is in the clinical trial phase, and its approval will depend on the results and regulation of agencies such as the FDA in the United States.

Will the AI replace the radiologist's work?

Nope. The tool was designed to act as a second pair of eyes, highlighting exams that deserve special attention. The final report remains the doctor's responsibility.

Does this technique work for other types of cancer?

Yes, the same approach is being studied for lung, breast and ovarian cancer, with promising results in early stages. The research was originally published on the [NBC News] portal (https://www.nbcnews.com/health/cancer/ai-early-signs-pancreatic-cancer-before-tumors-develop-rcna343099) and reflects the growing impact of AI on medicine. Meanwhile, other areas also see technology being adopted in critical contexts, such as the implementation of AI in military networks closed by the Pentagon, showing that the tool's potential goes far beyond laboratories.