MAMA-MIA: A Large-Scale Multi-Center Breast Cancer DCE-MRI Benchmark Dataset with Expert Segmentations
Welcome to the MAMA-MIA dataset repository! This dataset contains 1,506 cases of breast cancer dynamic contrast-enhanced magnetic resonance images (DCE-MRI) with expert tumor segmentations. Below, you will find all the necessary information to download and use the dataset, as well as instructions on how to run inference using our pre-trained nnUNet model.
If you use the MAMA-MIA dataset or the pretrained model in your research, please cite our publication and the dataset publications of the images included in the dataset.
@article{Garrucho2024MAMAMIA,
title={MAMA-MIA: A Large-Scale Multi-Center Breast Cancer DCE-MRI Benchmark Dataset with Expert Segmentations},
author={Lidia Garrucho and Claire-Anne Reidel and Kaisar Kushibar and Smriti Joshi and Richard Osuala and Apostolia Tsirikoglou and Maciej Bobowicz and Javier del Riego and Alessandro Catanese and Katarzyna Gwoździewicz and Maria-Laura Cosaka and Pasant M. Abo-Elhoda and Sara W. Tantawy and Shorouq S. Sakrana and Norhan O. Shawky-Abdelfatah and Amr Muhammad Abdo-Salem and Androniki Kozana and Eugen Divjak and Gordana Ivanac and Katerina Nikiforaki and Michail E. Klontzas and Rosa García-Dosdá and Meltem Gulsun-Akpinar and Oğuz Lafcı and Ritse Mann and Carlos Martín-Isla and Fred Prior and Kostas Marias and Martijn P.A. Starmans and Fredrik Strand and Oliver Díaz and Laura Igual and Karim Lekadir},
journal={arXiv preprint},
year={2024},
doi={arXiv:2406.13844},
url={https://arxiv.org/abs/2406.13844}
}
Some of the potential applications of the dataset are:
- Classification, Detection and Segmentation of Breast Cancer in MRI using Deep Learning.
- Radiomic extraction and Machine Learning.
- Segmentation Model Finetuning using the Pretrained nnUNet Weigths.
- Fine-tuning of Foundational Models.
- Treatment Response to Neoadjuvant Chemotherapy.
- Patient Survival Prediction.
- Segmentation Quality Control.
- Image Synthesis.
The MAMA-MIA dataset is hosted on Synapse. You can download the dataset using the CLI or Python with the following code:
synapse get syn60868042
import synapseclient
syn = synapseclient.Synapse()
syn.login()
entity = syn.get("syn60868042")
Check Synapse documentation for more info.
The pre-trained vanilla nnUNet model has been trained using the 1506 full-image DCE-MRIs and the expert segmentations from the MAMA-MIA dataset.
Full-image Tumor Segmentation | Dice | IoU | HD95 | MSD |
---|---|---|---|---|
nnUNet Fold 0 | 0.7600 | 0.6512 | 39.90 | 11.89 |
nnUNet Fold 1 | 0.7617 | 0.6500 | 41.25 | 12.10 |
nnUNet Fold 2 | 0.7749 | 0.6714 | 31.12 | 8.86 |
nnUNet Fold 3 | 0.7616 | 0.6577 | 36.25 | 11.74 |
nnUNet Fold 4 | 0.7520 | 0.6396 | 38.57 | 10,81 |
nnUNet Mean | 0.7620 | 0.6539 | 37.41 | 11.08 |
The metrics correspond to Dice Coefficient, Intersection over Union (IoU), the 95 Percentile of Hausdorff Distance (HD95) and to the Mean Surface Distance (MSD). The distances are in mm. The library used to compute these metrics was seg-metrics v1.2.7.
Clone the forked repository of nnUNet.
git clone https://github.com/LidiaGarrucho/MAMA-MIA
cd MAMA-MIA/nnUNet
To run the pre-trained nnUNet model, follow the installation instructions.
The nnUNet pretrained weights can be dowloaded from Synapse.
Unzip the folder inside nnUNet GitHub repository under nnUNet/nnunetv2/nnUNet_results
.
The recommended preprocessing steps to get optimum performance are:
- z-score normalization. For DCE-MRI, use the mean and standard deviation of all the phases (from pre to last post-contrast) to z-score the DCE-MRI sequence.
- isotropic pixel spacing. The MRIs were resampled using a uniform pixel spacing of [1,1,1].
nnUNetv2_predict -i /path/to/your/images -o /path/to/output -d 101 -c 3d_fullres
- Replace
/path/to/your/images
with the directory containing your input images. - Replace
/path/to/output
with the directory where you want to save the output segmentations.
Note: An error might arise if your images are not in compressed NifTI format (.nii.gz).
If you use the MAMA-MIA dataset or the pretrained model in your research, please cite our publication and the dataset publications of the images included in the dataset.
Garrucho, L., Reidel, C.-A., Kushibar, K., Joshi, S., Osuala, R., Tsirikoglou, A., Bobowicz, M., del Riego, J., Catanese, A., Gwoździewicz, K., Cosaka, M.-L., Abo-Elhoda, P. M., Tantawy, S. W., Sakrana, S. S., Shawky-Abdelfatah, N. O., Abdo-Salem, A. M., Kozana, A., Divjak, E., Ivanac, G., Nikiforaki, K., Klontzas, M. E., García-Dosdá, R., Gulsun-Akpinar, M., Lafcı, O., Mann, R., Martín-Isla, C., Prior, F., Marias, K., Starmans, M. P. A., Strand, F., Díaz, O., Igual, L., and Lekadir, K. (2024). MAMA-MIA: A Large-Scale Multi-Center Breast Cancer DCE-MRI Benchmark Dataset with Expert Segmentations. arXiv preprint (arXiv:2406.13844).
@article{Garrucho2024MAMAMIA,
title={MAMA-MIA: A Large-Scale Multi-Center Breast Cancer DCE-MRI Benchmark Dataset with Expert Segmentations},
author={Lidia Garrucho and Claire-Anne Reidel and Kaisar Kushibar and Smriti Joshi and Richard Osuala and Apostolia Tsirikoglou and Maciej Bobowicz and Javier del Riego and Alessandro Catanese and Katarzyna Gwoździewicz and Maria-Laura Cosaka and Pasant M. Abo-Elhoda and Sara W. Tantawy and Shorouq S. Sakrana and Norhan O. Shawky-Abdelfatah and Amr Muhammad Abdo-Salem and Androniki Kozana and Eugen Divjak and Gordana Ivanac and Katerina Nikiforaki and Michail E. Klontzas and Rosa García-Dosdá and Meltem Gulsun-Akpinar and Oğuz Lafcı and Ritse Mann and Carlos Martín-Isla and Fred Prior and Kostas Marias and Martijn P.A. Starmans and Fredrik Strand and Oliver Díaz and Laura Igual and Karim Lekadir},
journal={arXiv preprint},
year={2024},
doi={arXiv:2406.13844},
url={https://arxiv.org/abs/2406.13844}
}
The MAMA-MIA dataset includes public DCE-MRI images from four different collection in the TCIA repository under the following licenses:
ISPY1 Trial (License CC BY 3.0)
David Newitt, Nola Hylton, on behalf of the I-SPY 1 Network and ACRIN 6657 Trial Team. (2016). Multi-center breast DCE-MRI data and segmentations from patients in the I-SPY 1/ACRIN 6657 trials. The Cancer Imaging Archive. https://doi.org/10.7937/K9/TCIA.2016.HdHpgJLK
Breast-MRI-NACT-Pilot (License CC BY 3.0)
Newitt, D., & Hylton, N. (2016). Single site breast DCE-MRI data and segmentations from patients undergoing neoadjuvant chemotherapy (Version 3) [Data set]. The Cancer Imaging Archive. https://doi.org/10.7937/K9/TCIA.2016.QHsyhJKy
ISPY2 Trial (License CC BY 4.0)
Li, W., Newitt, D. C., Gibbs, J., Wilmes, L. J., Jones, E. F., Arasu, V. A., Strand, F., Onishi, N., Nguyen, A. A.-T., Kornak, J., Joe, B. N., Price, E. R., Ojeda-Fournier, H., Eghtedari, M., Zamora, K. W., Woodard, S. A., Umphrey, H., Bernreuter, W., Nelson, M., … Hylton, N. M. (2022). I-SPY 2 Breast Dynamic Contrast Enhanced MRI Trial (ISPY2) (Version 1) [Data set]. The Cancer Imaging Archive. https://doi.org/10.7937/TCIA.D8Z0-9T85
Duke-Breast-Cancer-MRI (License CC BY-NC 4.0)
Saha, A., Harowicz, M. R., Grimm, L. J., Weng, J., Cain, E. H., Kim, C. E., Ghate, S. V., Walsh, R., & Mazurowski, M. A. (2021). Dynamic contrast-enhanced magnetic resonance images of breast cancer patients with tumor locations [Data set]. The Cancer Imaging Archive. https://doi.org/10.7937/TCIA.e3sv-re93
Thank you for using our dataset and pretrained model! If you have any questions or issues, please feel free to open an issue in this repository.