Winning Solution of the ODELIA Breast MRI Challenge 2025
Meisen Meister is the official winning entry for the ODELIA Breast MRI Challenge 2025, a competition focused on advancing deep learning methods for breast cancer MRI classification.
Our solution achieved first place on the final leaderboard, demonstrating strong generalization and robustness across diverse MRI datasets and acquisition settings.
- Top-ranked model on the final leaderboard
- Robust domain adaptation across institutions
- Optimized pipeline for efficiency and reproducibility
The ODELIA Breast MRI Challenge 2025 aimed to benchmark automated methods for breast lesion classification using MRI data.
Teams were evaluated on a hidden test set to ensure generalization and reliability.
Leaderboard: Final Evaluation Results
We are currently cleaning and organizing the code before public release.
All scripts, model weights, and documentation will be made available here soon.
Stay tuned — a detailed setup guide and training pipeline will be added in the coming weeks.