Privacy-preserving Federated Brain Tumour Segmentation

By Wenqi Li, Fausto Milletarì, Daguang Xu, Nicola Rieke, Jonny Hancox, Wentao Zhu, Maximilian Baust, Yan Cheng, Sébastien Ourselin, M. Jorge Cardoso and 1 more. 2019

562 citations recorded by the source corpus.

Due to medical data privacy regulations, it is often infeasible to collect and share patient data in a centralised data lake. This poses challenges for training machine learning algorithms, such as deep convolutional networks, which often require large numbers of diverse training examples. Federated learning sidesteps this difficulty by bringing code to the patient data owners and only sharing intermediate model training updates among them. Although a high-accuracy model could be achieved by appropriately aggregating these model updates, the model shared could indirectly leak the local training examples. In this paper, we investigate the feasibility of applying differential-privacy techniques to protect the patient data in a federated learning setup. We implement and evaluate practical federated learning systems for brain tumour segmentation on the BraTS dataset. The experimental results show that there is a trade-off between model performance and privacy protection costs.

Subfields: Artificial Intelligence, Health Informatics, Radiology, Nuclear Medicine and Imaging

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Health Informatics

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

FedML: A Research Library and Benchmark for Federated Machine Learning (2020, background): “…computer vision [77, 78], natural language processing [79, 80, 81, 82, 83], ﬁnance [74, 35, 84], transportation [85, 86, 87, 88, 89, 90, 91, 92, 93, 94], digital health [95, 96, 97, 98, 99, 100, 101], recommendation [102, 103, 104, 105, 106, 107], robotics [108, 109], and smart cities [110, 111].”
Federated Learning for Healthcare: Systematic Review and Architecture Proposal (2022, methodology): “Li et al. [74] provide an in-depth analysis of differential privacy applied to FL for healthcare.”
End-to-end privacy preserving deep learning on multi-institutional medical imaging (2021, background): “Li et al. 44 also demonstrate an FL segmentation task.”
A review of deep learning in medical imaging: Imaging traits, technology trends, case studies with progress highlights, and future promises (2020, methodology): “[85] study several practical FL methods while protecting data privacy for brain tumor segmentation on the BraTS data set and demonstrate a tradeoff between model performance and privacy protection costs.”
Privacy-Enhanced Federated Learning Against Poisoning Adversaries (2021, background): “Beneﬁtting from its ability to train the model without storing the training data in the cloud, FL has become one of the fundamental techniques in many security-sensitive tasks such as autonomous driving [2] and medical image analyses [3].”
Deep learning based brain tumor segmentation: a survey (2020, background): “In the future, privacy-preserved learning framework are expected to be explored aiming at protecting patients privacy [78].”
Privacy-Preserving Technology to Help Millions of People: Federated Prediction Model for Stroke Prevention (2020, background): “Li [ Li et al. , 2019 ] investigated the feasibility of applying differential-privacy techniques to protect the patient data in a federated learning setup for brain tumor segmentation.”
A Survey on Federated Learning Systems: Vision, Hype and Reality for Data Privacy and Protection (2019, background): “The targeted users of Clara FL is hospitals and medical institutions [107].”
FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space (2021, background): “[27] further compare several weights sharing strategies in FL to alleviate the effect of data imbalance among different hospitals.”
The future of digital health with federated learning (2020, methodology): “Recent research has shown that this approach can achieve a performance that is comparable to a scenario where the data was co-located in a data lake and superior to the models that only see isolated singleinstitutional data [12, 13].”
Federated learning for medical image analysis: A survey (2023, result): “Similar strategies can also be found in [85].”
FedScale: Benchmarking Model and System Performance of Federated Learning (2021, methodology): “It enables model training and evaluation on end-user data, while circumventing high cost and privacy risks in gathering the raw data from clients, with applications in diverse domains: for example, NVIDIA applies FL to create medical imaging AI [26]; Google runs federated training of NLP models in G”
Federated Learning for Smart Healthcare: A Survey (2021, methodology): “Particularly, FL is able to be deployed on medical devices such as MRI scan machine [99] which has enough local dataset and reasonable computational resources to compute AI updates to join the FL process with the cloud server.”
Collaborative Unsupervised Visual Representation Learning from Decentralized Data (2021, background): “Federated learning is an emerging distributed training technique [21] that have empowered multiple parties to collaboratively learn computer vision models, such as segmentation [18, 28], object detection [20], and person re-identiﬁcation [42].”
Multi-site fMRI Analysis Using Privacy-preserving Federated Learning and Domain Adaptation: ABIDE Results (2020, background): “Later, another work [16] empirically studied privacy-preserving issues using a sparse vector technique and investigated model weights sharing schemes for imbalanced data.”
Secure, privacy-preserving and federated machine learning in medical imaging (2020, background): “…the current paradigm of data sharing and centralized storage, have the greatest potential to enable privacy-preserving cross-institutional research in a breadth of biomedical disciplines in the near future 84,85 , with results in medical imaging 50,86 and genomics 87 recently demonstrated.”