Federated learning and differential privacy for medical image analysis

Adnan, Mohammed; Kalra, Shivam; Cresswell, Jesse C.; Taylor, Graham W.; Tizhoosh, Hamid R.

doi:10.1038/s41598-022-05539-7

Cited by 163 publications

(84 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Homomorphic encryption 58 , secure multiparty compute 59 , and trusted execution environments (TEEs) 60 , 61 allow for collaborative computations to be performed with untrusted parties while maintaining confidentiality of the inputs to the computation. Differentially private training algorithms 62 – 64 allow for mitigation of information leakage from both the collaborator model updates and the global consensus aggregated models. Finally, assurance that remote computations are executed with integrity can be designed for with the use of hardware-based trust provided by TEEs, as well as with some software-based integrity checking 65 .…”

Section: Discussionmentioning

confidence: 99%

“…This study is meant to be used as an example for future FL studies between collaborators with an inherent amount of trust that can result in clinically deployable ML models. Further research is required to assess privacy concerns in a detailed manner 63 , 64 and to apply FL to different tasks and data types 66 – 69 . Building on this study, a continuous FL consortium would enable downstream quantitative analyses with implications for both routine practice and clinical trials, and most importantly, increase access to high-quality precision care worldwide.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Federated learning enables big data for rare cancer boundary detection

et al. 2022

View full text Add to dashboard Cite

Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Federated learning enables big data for rare cancer boundary detection

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Perhaps the most known and commonly used FL algorithm is FedAvg which was initially proposed by McMahan et al 13 and then extended for medical image analysis recently by Lu et al 20 and Adnan et al 24 . It learns a global model by aggregating local models trained on independent identically distributed data, as shown in Fig.…”

Section: Methodsmentioning

confidence: 99%

“…Also, we train a centralized model following the traditional way by collecting all training data at the server. For the federated learning paradigm, we use Federated Learning Average (FedAvg) 20 , 24 as the baseline model to compare with. At the evaluation stage, all developed modes are evaluated with the global test set for a fair comparison.…”

Section: Methodsmentioning

confidence: 99%

Federated learning with hyper-network—a case study on whole slide image analysis

Lin

Wang

Dong

et al. 2023

Sci Rep

View full text Add to dashboard Cite

Federated learning(FL) is a new kind of Artificial Intelligence(AI) aimed at data privacy preservation that builds on decentralizing the training data for the deep learning model. This new technique of data security and privacy sheds light on many critical domains with highly sensitive data, including medical image analysis. Developing a strong, scalable, and precise deep learning model has proven to count on a variety of high-quality data from different centers. However, data holders may not willing to share their data considering the restriction of privacy. In this paper, we approach this challenge with a federated learning paradigm. Specifically, we present a case study on the whole slide image classification problem. At each local client center, a multiple-instance learning classifier is developed to conduct whole slide image classification. We introduce a privacy-preserving federated learning framework based on hyper-network to update the global model. Hyper-network is deployed at the global center that produces the weights of the local network conditioned on its input. In this way, hyper-networks can simultaneously learn a family of the local client networks. Instead of communicating raw data with the local client, only model parameters injected with noise are transferred between the local client and the global model. By using a large scale of whole slide images with only slide-level labels, we mensurated our way on two different whole slide image classification problems. The results demonstrate that our proposed federated learning model based on hyper-network can effectively leverage multi-center data to develop a more accurate model which can be used to classify a whole slide image. Its improvements in terms of over the isolated local centers and the commonly used federated averaging baseline are significant. Code will be available.

show abstract

“…Differential privacy in federated learning is often achieved using differentially-private stochastic gradient descent (DP-SGD) [ 7 , 41 , 42 ], an algorithm that determines the appropriate noise scale and how to clip the model parameter. The combination of federated learning and differential privacy has been explored in multiple medical use cases, including prediction of mortality and adverse drug reactions from electronic health records [ 43 ], brain tumor segmentation [ 9 ], classification of pathology whole slide images [ 20 ], detection of diabetic retinopathy in images of the retina [ 44 ], and identification of lung cancer in histopathologic images [ 45 ].…”

Section: Related Workmentioning

confidence: 99%

Defending against Reconstruction Attacks through Differentially Private Federated Learning for Classification of Heterogeneous Chest X-ray Data

Ziegler

Pfitzner

Schulz

et al. 2022

Sensors

View full text Add to dashboard Cite

Privacy regulations and the physical distribution of heterogeneous data are often primary concerns for the development of deep learning models in a medical context. This paper evaluates the feasibility of differentially private federated learning for chest X-ray classification as a defense against data privacy attacks. To the best of our knowledge, we are the first to directly compare the impact of differentially private training on two different neural network architectures, DenseNet121 and ResNet50. Extending the federated learning environments previously analyzed in terms of privacy, we simulated a heterogeneous and imbalanced federated setting by distributing images from the public CheXpert and Mendeley chest X-ray datasets unevenly among 36 clients. Both non-private baseline models achieved an area under the receiver operating characteristic curve (AUC) of 0.94 on the binary classification task of detecting the presence of a medical finding. We demonstrate that both model architectures are vulnerable to privacy violation by applying image reconstruction attacks to local model updates from individual clients. The attack was particularly successful during later training stages. To mitigate the risk of a privacy breach, we integrated Rényi differential privacy with a Gaussian noise mechanism into local model training. We evaluate model performance and attack vulnerability for privacy budgets ε∈{1,3,6,10}. The DenseNet121 achieved the best utility-privacy trade-off with an AUC of 0.94 for ε=6. Model performance deteriorated slightly for individual clients compared to the non-private baseline. The ResNet50 only reached an AUC of 0.76 in the same privacy setting. Its performance was inferior to that of the DenseNet121 for all considered privacy constraints, suggesting that the DenseNet121 architecture is more robust to differentially private training.

show abstract

Federated learning and differential privacy for medical image analysis

Cited by 163 publications

References 27 publications

Federated learning enables big data for rare cancer boundary detection

Federated learning enables big data for rare cancer boundary detection

Federated learning with hyper-network—a case study on whole slide image analysis

Defending against Reconstruction Attacks through Differentially Private Federated Learning for Classification of Heterogeneous Chest X-ray Data

Contact Info

Product

Resources

About