The Radiological Society of North America (RSNA) has launched its latest artificial intelligence challenge, focusing on enhancing MR imaging of the knee. Researchers are invited to create machine learning models to detect and classify knee abnormalities, utilizing medical images in conjunction with radiology reports for the first time. This challenge aims to address real-world diagnostic complexities and variability in interpretations. The training dataset includes over 5,000 knee MRIs and related reports from 16 healthcare sites globally, with expert annotations to evaluate model performance. The competition will run until October 22, with $77,000 in prizes for top-performing teams, including a new award for efficiency. Winners will be announced at the RSNA 2026 event in Chicago.
Mon, 03 Aug 2026 11:06:10 GMT | Radiology Business