Researchers at Immanuel Kant Baltic Federal University, in collaboration with Southern Federal University, have developed an AI system that can detect early-stage pancreatic cancer from CT scans using a neural network based on the U-Net architecture. The model is designed as a clinical decision-support tool that highlights and segments suspicious regions in medical images, helping radiologists and oncologists identify tumors that are often missed in early stages when treatment is most effective. Tested on real CT datasets, the system reportedly achieved strong performance metrics, including 88% accuracy and 98% sensitivity and specificity, suggesting high reliability in distinguishing abnormal tissue from healthy pancreatic structures. Importantly, the tool is intended to assist—not replace—medical professionals by speeding up image analysis and improving consistency in diagnosis while leaving final clinical decisions to doctors. The research reflects a broader trend across BRICS countries, where AI and biotechnology are increasingly being applied to cancer detection, surgical support, and precision medicine to improve early diagnosis and treatment outcomes.
Read more here
