The TrackRAD2025 Challenge concludes

TrackRAD2025 is finished

The TrackRAD2025 Challenge has concluded with the prize giveaway at the MICCAI2025 Conference. The top-5 submissions were all very close and we now know that fine-tuned foundation models are the way to go to track lesions on cine-MRI, with the winning team using MedSAM2. The challenge page remains alive and the peer-reviewed method descriptions are available for each submission, ensuring reproducibility along with open source code. Thanks to all participants!

Guillaume Landry
Guillaume Landry
Professor of Image Guided Radiation Therapy

My research interests include image guidance and artificial intelligence.