Welcome to the MAMA Challenges! We are dedicated to advancing breast cancer research through open-source code, high-quality datasets, and reproducible AI benchmarks. 🚀
We are proud to showcase the MAMA-MIA Challenge, originally featured at MICCAI 2025! This challenge focuses on large-scale multicenter breast cancer DCE-MRI analysis.
-
The Dataset: 1,506 cases of breast cancer DCE-MRI with expert segmentations. Find the dataset here.
-
Two Tasks: Automated tumor segmentation and treatment response prediction.
-
Evaluation: Validated on private multicenter datasets of 572 patients to ensure real-world robustness and fairness across subgroups.
-
Explore the data and challenge repository here: LidiaGarrucho/MAMA-MIA
-
Submit your algorithms to our running long-term benchmark for comparison: Codabench
Virtual Contrast-Enhanced Breast MRI Synthesis
We are thrilled to announce MAMA-SYNTH, a challenge dedicated to the future of contrast-free breast imaging.
-
Safer Imaging: Reducing reliance on gadolinium-based agents to eliminate safety concerns and contraindications.
-
Streamlined Workflow: Lowering clinical costs and patient burden through virtual enhancement.
-
Generative Excellence: Utilizing SOTA deep generative modeling to synthesize post-contrast images from pre-contrast acquisitions.
-
Image Fidelity: Can synthetic images match the quality of real DCE-MRI?
-
Lesion Realism: Ensuring clinically accurate representation of tumors without real contrast.
-
Downstream Utility: Validating synthetic data for actual clinical decision-making and treatment monitoring.
Coming Soon! 🚀 Stay tuned to this organization challenge timeline and submission guidelines!
