CHALLENGES IN GENERALIZING BREAST MRI TUMOR SEGMENTATION ACROSS MULTIPLE DATASETS
DOI:
https://doi.org/10.37943/23AAOF8219Keywords:
Breast Cancer, Magnetic Resonance Imaging, Tumor Segmentation, Deep Learning, Model Generalizability, Medical Image Analysis, model robustnessAbstract
Accurate segmentation of breast tumors in dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is essential for precise diagnosis, treatment planning, and quantitative analysis. While deep learning methods have achieved strong performance in controlled research settings, their ability to generalize across diverse clinical datasets remains underexplored and poses a major barrier to clinical adoption. In this study, we evaluate the cross-dataset generalizability of a 3D Residual U-Net model using the multicenter MAMA-MIA benchmark, which consolidates four publicly available breast MRI collections annotated by expert radiologists. A leave-one-out experimental design is employed, with three datasets used for training and validation, and the remaining dataset held-out for independent testing to simulate real-world deployment scenarios. Model performance is assessed using Dice coefficient, Precision, and Recall, alongside quantitative analysis of tumor volume estimation accuracy. The best Dice score achieved by our model was 0.683 when tested on the NACT subset. Results show a consistent degradation in segmentation accuracy when models are applied to unseen datasets, indicating that performance declines significantly outside the distribution of the training data. The most pronounced drop occurs when the DUKE dataset serves as the held-out test set, where the model struggles to adapt to differences in pre-release preprocessing strategies. A targeted qualitative review of 160 representative scans further reveals key factors contributing to both successful and failed segmentations, including variations in image field of view, temporal enhancement patterns, acquisition era, and artifact prevalence. Overall, these findings underscore the importance of accounting for dataset heterogeneity, domain shift, and standardized preprocessing in the development of robust, clinically deployable breast MRI segmentation models capable of generalizing across institutions and imaging protocols.
References
Bahdanau, D., Cho, K., & Bengio, Y. (2016). Neural machine translation by jointly learning to align and translate. https://arxiv.org/abs/1409.0473
Cardoso, M. J., Li, W., Brown, R., Ma, N., Kerfoot, E., Wang, Y., Murrey, B., Myronenko, A., Zhao, C., Yang, D., Nath, V., He, Y., Xu, Z., Hatamizadeh, A., Myronenko, A., Zhu, W., Liu, Y., Zheng, M., Tang, Y., . . . Feng, A. (2022). Monai: An open-source framework for deep
learning in healthcare. https://arxiv.org/abs/2211.02701
Chitalia, R., Pati, S., Bhalerao, M., Thakur, S., Jahani, N., Belenky, J. V., McDonald, E. S., Gibbs, J., Newitt, D., Hylton, N., Kontos, D., & Bakas, S. (2021). Expert tumor annotations and radiomic features for the ISPY1/Acrin 6657 trial data collection. https://doi.org/10.7937/TCIA.XC7A-QT20
Garrucho, L., Kushibar, K., Reidel, C.-A., Joshi, S., Osuala, R., Tsirikoglou, A., Bobowicz, M., Riego, J. d., Catanese, A., Gwo´zdziewicz, K., Cosaka, M.-L., Abo-Elhoda, P. M., Tantawy, S. W., Sakrana, S. S., Shawky- Abdelfatah, N. O., Salem, A. M. A., Kozana, A., Divjak, E., Ivanac, G., . . . Lekadir, K. (2025). A large-scale multicenter breast cancer DCE-MRI benchmark dataset with expert segmentations. Scientific Data, 12 (1), 453. https://doi.org/10.1038/s41597-025-04707-4
Guo, Y. Y., Huang, Y. H., Wang, Y., Huang, J., & ... (2022). Breast MRI tumor automatic segmentation and triple-negative breast cancer discrimination algorithm based on deep learning [Publisher: Wiley Online Library]. . . . Methods in Medicine. https://doi.org/10.1155/2022/2541358
He, J., Zhao, X., Luo, Z., Su, S., Li, S., & Zhang, G. (2024). TSESNet: Temporal-spatial enhanced breast tumor segmentation in DCE-MRI using feature perception and separability. Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, 803–811. https://doi.org/10.24963/ijcai.2024/89
Huang, R., Xu, Z., Xie, Y., Wu, H., Li, Z., Cui, Y., Huo, Y., Han, C., Yang, X., Liu, Z., & Wang, Y. (2023). Joint-phase attention network for breast cancer segmentation in DCE-MRI [Publisher: Elsevier BV]. Expert Systems with Applications, 224, 119962. https://doi.org/10.1016/j.eswa.2023.119962
Isensee, F., Jaeger, P. F., Kohl, S. A. A., Petersen, J., & Maier-Hein, K. H. (2020). nnU-Net: A self-configuring method for deep learning-based biomedical image segmentation. Nature Methods, 18 (2), 203–211. https://doi.org/10.1038/s41592-020-01008-z
Khaled, R., Vidal, J., Vilanova, J. C., & Mart´ı, R. (2022). A u-net ensemble for breast lesion segmentation in DCE MRI. [Place: United States]. Computers in biology and medicine, 140, 105093. https://doi.org/10.1016/j.compbiomed.2021.105093
Newitt, D., & Hylton, N. (2016). Single site breast DCE-MRI data and segmentations from patients undergoing neoadjuvant chemotherapy. https://doi.org/10.7937/K9/TCIA.2016.QHSYHJKY
Newitt, D. C., Partridge, S. C., Zhang, Z., Gibbs, J., Chenevert, T., Rosen, M., Bolan, P., Marques, H., Romanoff, J., Cimino, L., Joe, B. N., Umphrey, H., Ojeda-Fournier, H., Dogan, B., Oh, K. Y., Abe, H., Drukteinis, J., Esserman, L. J., & Hylton, N. M. (2021). Acrin 6698/I-SPY2 breast DWI. https://doi.org/10.7937/TCIA.KK02-6D95
Qin, C., Wu, Y., Zeng, J., Tian, L., Zhai, Y., Li, F., & Zhang, X. (2022). Joint transformer and multi-scale CNN for DCE-MRI breast cancer segmentation [Publisher: Springer]. Soft Computing, 26 (17), 8317–8334.
Rahimpour, M., Saint Martin, M.-J., Frouin, F., Akl, P., Orlhac, F., Koole, M., & Malhaire, C. (2022). Visual ensemble selection of deep convolutional neural networks for 3D segmentation of breast tumors on dynamic contrast enhanced MRI [Publisher: Springer Science and Business Media LLC]. European Radiology, 33 (2), 959–969. https://doi.org/10.1007/s00330-022-09113-7
Ronneberger, O., Fischer, P., & Brox, T. (2015). U-net: Convolutional networks for biomedical image segmentation. https://arxiv.org/abs/1505.04597
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. https://doi.org/10.7937/TCIA.E3SV-RE93
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2023). Attention is all you need. https://arxiv.org/abs/1706.03762
World Health Organization. (2021). Breast cancer. Geneva, Switzerland: World Health Organization. Retrieved from https://www.who.int/news-room/fact-sheets/detail/breast-cancer
Xu, Z., Rauch, D. E., Mohamed, R. M., Pashapoor, S., Zhou, Z., & ... (2023). Deep learning for fully automatic tumor segmentation on serially acquired dynamic contrast-enhanced MRI images of triple-negative breast cancer [Publisher: mdpi.com Type: HTML]. Cancers. https://www.mdpi.com/2072-6694/15/19/4829
Yue, W., Zhang, H., Zhou, J., Li, G., Tang, Z., Sun, Z., & ... (2022). Deep learning-based automatic segmentation for size and volumetric measurement of breast cancer on magnetic resonance imaging [Publisher: frontiersin.org Type: HTML]. Frontiers in . . . https://doi.org/10.3389/fonc.2022.984626
Zhang, J., Cui, Z., Shi, Z., Jiang, Y., Zhang, Z., Dai, X., Yang, Z., Gu, Y., Zhou, L., Han, C., Huang, X., Ke, C., Li, S., Xu, Z., Gao, F., Zhou, L., Wang, R., Liu, J., Zhang, J., . . . Shen, D. (2023). A robust and efficient AI assistant for breast tumor segmentation from DCE-MRI via
a spatial-temporal framework. [Place: United States]. Patterns (New York, N.Y.), 4 (9), 100826. https://doi.org/10.1016/j.patter.2023.100826
Zhang, Q., Xiao, J., & Zheng, B. (2023). Image segmentation of triple-negative breast cancer by incorporating multiscale and parallel attention mechanisms (S. Hussain, Ed.) [Publisher: Wiley]. Scientific Programming, 2023, 1–13. https://doi.org/10.1155/2023/6629189
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Articles are open access under the Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Authors who publish a manuscript in this journal agree to the following terms:
- The authors reserve the right to authorship of their work and transfer to the journal the right of first publication under the terms of the Creative Commons Attribution License, which allows others to freely distribute the published work with a mandatory link to the the original work and the first publication of the work in this journal.
- Authors have the right to conclude independent additional agreements that relate to the non-exclusive distribution of the work in the form in which it was published by this journal (for example, to post the work in the electronic repository of the institution or publish as part of a monograph), providing the link to the first publication of the work in this journal.
- Other terms stated in the Copyright Agreement.