AI detects pancreatic cancer signs 475 days before diagnosis

The system also automatically identifies and outlines the pancreas, separating it from nearby organs and tissues. That removes the need for specialists to trace the organ's boundaries manually, which can introduce inconsistencies between observers.

This approach targets a central problem. Pancreatic ductal adenocarcinoma, or PDAC, is the most common type of pancreatic cancer and has a low survival rate, largely because doctors often discover it only after it advances. Early disease produces few noticeable symptoms, its tissue changes can remain subtle on standard scans, and it can progress quickly.

How did it compare with radiologists?

Across the cancer group, REDMOD achieved 73% sensitivity, versus 39% for experienced radiologists. Sensitivity measures how effectively a diagnostic test identifies people who actually have a disease, rather than overlooking genuine cases.

The gap was wider for scans taken more than two years before diagnosis. REDMOD's detection rate reached 68%, nearly three times the radiologists' rate of 23%.

Researchers also tested whether the model could recognize cancer-free scans. In an independent group of 539 patients from several hospitals, REDMOD correctly classified just over 81% as cancer-free. On the public US National Institutes of Health NIH-PCT dataset of 80 patients, it correctly identified 87.5% as free of pancreatic cancer.

When researchers compared scans taken from the same patients several months apart, REDMOD produced matching results in 90% to 92% of cases. The researchers said this consistency strengthened the evidence that the model was identifying meaningful disease-associated changes.

Could earlier diagnosis improve survival?

The researchers said the 475-day lead time could increase the chance of cure and the likelihood of improved survival. Their modelling estimate concerned localized disease rather than advanced disease.

They wrote: "In fact, modelling studies indicate that increasing the proportion of localized [pancreatic ductal carcinomas] from 10% to 50% would more than double survival rates, thereby underscoring that the timing of diagnosis is the single most critical determinant of survival outcomes."

The authors interpret the hidden signals as imaging evidence of stage 0 PDAC, the earliest stage, when treatment is more likely to succeed. Earlier detection could give patients an opportunity for treatment while the disease remains more treatable.

What must happen before clinical use?

Researchers still need prospective testing, which follows people forward in time, before REDMOD enters routine care. Testing should particularly include high-risk patients, including those with unexplained weight loss or recently diagnosed diabetes.

The study population did not include a sufficiently diverse range of ethnic backgrounds, leaving performance across populations uncertain. The researchers concluded: "This study validates REDMOD as a fully automated AI framework capable of identifying the imaging signatures of stage 0 [pancreatic ductal adenocarcinoma] in normal pancreas, achieving this with substantial lead times and performance superior to expert radiologists."

They added that prospective validation remains essential and framed the framework as a possible shift from late-stage symptomatic diagnosis toward finding PDAC before symptoms. The findings do not yet show that earlier detection improves survival in everyday clinical practice.