The RSNA has launched its 2026 Knee Abnormality Detection AI Challenge, inviting researchers to develop machine-learning models to detect and classify knee abnormalities on MRI using both images and radiology report text.
The challenge is the first RSNA AI challenge to focus on musculoskeletal MRI and the first to combine medical images with multilingual radiology reports, the RSNA said. The training dataset includes more than 5,000 knee MRI exams and associated radiology reports in 12 languages from 16 sites across five continents. The dataset used to assess model performance was annotated by expert radiologists.
Top-performing teams will share $77,000 in prize money, including first-time awards for the most efficient models. The competition will run through October 22, with winners announced in November and recognized at RSNA 2026, held November 29 to December 3 in Chicago.
