A new study from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) has developed a machine learning system that can predict the risk of developing pancreatic cancer up to three years before diagnosis. The model analyzes patterns in patients’ medical records, specifically sequences of disease codes and the timing of diagnoses, to identify subtle precursors. In tests, the AI significantly outperformed current screening methods, which typically only catch the disease at later stages. The researchers emphasize the tool is intended to assist doctors in prioritizing high-risk patients for further testing, not to provide definitive diagnoses. Pancreatic cancer is often lethal due to late detection, making early identification crucial for improving survival rates. Read the full article: https://technologyreview.com/2024/05/08/1099007/ai-predicts-pancreatic-cancer-three-years-before-diagnosis/