Federated Learning for Brain Tumor Segmentation in MRI Scans: A Privacy-Preserving and Domain-Adaptable Approach
Keywords:
N\AAbstract
The application of deep learning in intricate tasks like brain tumour segmentation heavily relies on potent algorithms, which in turn depend on large and heterogeneous datasets that are often fraught with privacy issues or deep domain variations (domain shift). Contra-violent methods concerning MRI data hold little promise as sharing sensitive data captured from patients across various institutions breaches privacy laws like HIPAA; furthermore, variation in scanner equipment and its associated detailing parameters lead to a pathological model building which does not generalise well on new data.
This document posits a robust approach to tackle these issues using Federated Learning (FL). With FL, it’s possible to train a model in collaboration with various institutions without sharing raw data from patients. In contrast, model training takes place privately first at each institution, and only learned model updates (i.e., the parameters or gradients of the model) are sent for aggregation on a central server. This allows for the refinement of a global model while keeping patient data safe within their institution. In addition, FL provides and preserves anywhere the variance of medical imaging data within one defined place as compared to traditional approaches, boosting the model resilience and flexibility to cope with new challenges due to changing MRI protocol shifts.
This approach included first locally training a Convolutional Neural Network (CNN) on data partitioned at each client site and then aggregating the model weights at a central location. The model was evaluated at the client based on the loss incurred during training and accuracy during prediction, achieving the intended goals.
Despite experiencing certain issues such as data heterogeneity (as shown by one client's loss oscillation), the model was able to learn from the distributed data and predict tumours correctly on the MRI scans. This study concludes that federated learning provides an effective framework with privacy-preserving characteristics and flexible domain adaptation for constructing resilient brain tumour segmentation models with multiple decentralised MRI datasets.
Downloads
Metrics
References
Nalawade, S. et al. (2022). Federated Learning for Brain Tumor Segmentation Using MRI and Transformers. In: Crimi, A., Bakas, S. (eds) Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries. BrainLes 2021. Lecture Notes in Computer Science, vol 12963.Springer,Cham. https://doi.org/10.1007/978-3-031- 09002-8_39
Mukherkjee D, Saha P, Kaplun D, Sinitca A, Sarkar R. Brain tumor image generation using an aggregation of GAN models with style transfer. Sci Rep. 2022;12(1):9141. Published 2022 Jun 1. doi:10.1038/s41598-022-12646-y
Vaidehi Satushe, Vibha Vyas, Shilpa Metkar, Davinder Paul Singh,AI in MRI Brain Tumor Diagnosis: A Systematic Review of Machine Learning and Deep Learning Advances(2010– 2025),Chemometrics and Intelligent Laboratory Systems,2025,105414,ISSN 0169-7439,https://doi.org/10.1016/j.chemolab. 2025.105414.
Liu, C. Chen, J. Qin, Q. Dou and P. -A. Heng, "FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space," 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville, TN, USA, 2021, pp. 1013-1023, doi:10.1109/CVPR46437.2021.00107.
X. Zhang et al., "MAPSeg: Unified Unsupervised Domain Adaptation for Heterogeneous Medical Image Segmentation Based on 3D Masked Autoencoding and Pseudo-Labeling," 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, USA, 2024,pp. 5851-5862,doi:10.1109/CVPR52733.2024.00559.
Yanzhen M, Song C, Wanping L, Zufang Y, Wang A. Exploring approaches to tackle cross-domain challenges in brain medical image segmentation: a systematic review. Front Neurosci. 2024;18:1401329. Published 2024 Jun 14. doi:10.3389/fnins.2024.1401329
H. Li, C. Li, W. Huang, X. Zheng, Y. Xi and S. Wang, "Digest: Deeply Supervised Knowledge Transfer Network Learning for Brain Tumor Segmentation with Incomplete Multi- Modal MRI Scans," 2023 IEEE 20th International Symposium on Biomedical Imaging (ISBI), Cartagena, Colombia, 2023, pp.1-4,doi: 10.1109/ISBI53787.2023.10230344.
Ullah F, Nadeem M, Abrar M, et al. Brain Tumor Segmentation from MRI Images Using Handcrafted Convolutional Neural Network. Diagnostics (Basel). 2023;13(16):2650. Published 2023 Aug 11.
doi:10.3390/diagnostics13162650
Ullah F, Nadeem M, Abrar M, Amin F, Salam A, Khan S. Enhancing Brain Tumor Segmentation Accuracy through Scalable Federated Learning with Advanced Data Privacy and Security Measures. Mathematics.2023;11(19):4189.https://doi.org/10.3390/math11194189
Tran AT, Zeevi T, Payabvash S. Strategies to Improve the Robustness and Generalizability of Deep Learning Segmentation and Classification in Neuroimaging. BioMedInformatics. 2025;5(2):20. https://doi.org/10.3390/biomedinformati cs5020020
M. Feng, B. Li, X. Xu, Y. Liu, H. Fu and W. Zuo, "Learning Federated Visual Prompt in Null Space for MRI Reconstruction," 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Vancouver, BC, Canada, 2023, pp. 8064-8073, doi: 10.1109/CVPR52729.2023.00779.
Sun L, Zhang S, Chen H, Luo L. Brain Tumor Segmentation and Survival Prediction Using Multimodal MRI Scans With Deep Learning. Front Neurosci. 2019;13:810. Published 2019 Aug 16. doi:10.3389/fnins.2019.00810
Menon and R. Ramakrishnan, "Brain Tumor Segmentation in MRI images using unsupervised Artificial Bee Colony algorithm and FCM clustering," 2015 International Conference on Communications and Signal Processing (ICCSP), Melmaruvathur, India, 2015,pp.0006-0009,doi: 10.1109/ICCSP.2015.7322635.
D. ND, K. R. Naik, C. H G, A. L and D. C, "Deep Learning Framework with Explainable AI for Accurate and Interpretable Brain Tumor Segmentation," 2025 International Conference on Intelligent Systems and Computational Networks (ICISCN), Bidar, India, 2025, pp. 1-6, doi: 10.1109/ICISCN64258.2025.10934555.
I.Khan, E. Alhoniemi, E. Kontio, S. A. Khan and M. Jafaritadi, "RegAgg: A Scalable Approach for Efficient Weight Aggregation in Federated Lesion Segmentation of Brain MRIs," 2023 Eighth International Conference on Fog and Mobile Edge Computing (FMEC), Tartu, Estonia, 2023, pp. 101-106, doi: 10.1109/FMEC59375.2023.10306171.
Nakhmani. A, Kikinis . R, Tannenbaum A.MRI Brain Tumor Segmentation and Necrosis Detection Using Adaptive Sobolev Snakes. Proc SPIE Int Soc Opt Eng.2014;9034:903442. doi:10.1117/12.2042915
Zhou T, Canu S, Vera P, Ruan S. Latent Correlation Representation Learning for Brain Tumor Segmentation With Missing MRI Modalities. IEEE Trans Image Process. 2021;30:4263- 4274. doi:10.1109/TIP.2021.3070752
Verma A, Yadav AK. Brain tumor segmentation with deep learning: Current approaches and future perspectives. J Neurosci Methods. 2025;418:110424. doi:10.1016/j.jneumeth.2025.110424
Zhou Z, Ma A, Feng Q, et al. Super- resolution of brain tumor MRI images based on deep learning. J Appl Clin Med Phys.2022;23(11):e13758. doi:10.1002/acm2.13758
Munir K, Frezza F, Rizzi A. Deep Learning Hybrid Techniques for Brain Tumor Segmentation. Sensors (Basel). 2022;22(21):8201. Published 2022 Oct26. doi:10.3390/s22218201
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
You are free to:
- Share — copy and redistribute the material in any medium or format
- Adapt — remix, transform, and build upon the material for any purpose, even commercially.
Terms:
- Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
- No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.
