The DRAGON benchmark for clinical NLP

Joeran Bosma and Koen Dercksen and Luc Builtjes and Romain André and Christian Roest and Stefan Fransen and Constant Noordman and Mar Navarro-Padilla and Judith Lefkes and Natália Alves (2025), NPJ Digital Medicine, 8, 289, Nature Publishing Group UK London

Introduces DRAGON, a benchmark of 28 clinical natural-language-processing tasks built from Dutch medical reports, to standardize how language models for pathology and radiology are evaluated.