Russian Scientists Develop AI to Detect Pancreatic Cancer Early
A neural network built by researchers in St. Petersburg can identify pancreatic tumors on CT scans with over 92% accuracy — a breakthrough for one of the deadliest cancers to diagnose.
Researchers at St. Petersburg Electrotechnical University (LETI) have developed an AI model designed to detect early-stage pancreatic cancer — a disease often discovered too late for effective treatment. Pancreatic cancer is known for its aggressiveness and tendency to spread rapidly, but small tumors are notoriously difficult to spot on CT images.
Accuracy Above 92%
The AI achieved an impressive 92.55% accuracy in detecting pathological changes in the pancreas. Researchers believe continued training will make the system even more reliable, minimizing the risk of missed diagnoses during CT analysis.
While the technology is not meant to replace doctors, it could significantly speed up cancer detection and help radiologists focus on the most critical findings — offering hope for earlier treatment of one of the world’s most lethal cancers.
