Revolutionizing Medical Data Sharing Using Advanced Privacy-Enhancing Technologies: Technical, Legal, and Ethical Synthesis

Scheibner, James; Raisaro, Jean Louis; Troncoso-Pastoriza, Juan Ramón; Ienca, Marcello; Fellay, Jacques; Vayena, Effy; Hubaux, J.-P.

doi:10.2196/25120

Cited by 71 publications

(57 citation statements)

References 54 publications

(70 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An example of this is the Personal Health Train initiative in the Netherlands. Such a federated setup can be combined with multiparty homomorphic encryption, which is a synthesis between two novel advanced privacy-enhancing technologies: homomorphic encryption and secure multiparty computation [63]. This will provide a mathematical guarantee of privacy, ensuring that data can be considered anonymized.…”

Section: Discussionmentioning

confidence: 99%

Literature Analysis of Artificial Intelligence in Biomedicine

Hulsen¹

2021

Preprint

View full text Add to dashboard Cite

Artificial Intelligence (AI) refers to the simulation of human intelligence in machines, using machine learning, deep learning and neural networks. AI enables machines to learn from experience and perform human-like tasks. The field of AI research has been developing fast over the past five to ten years, due to the rise of ‘big data’ and increasing computing power. In the medical area, AI can be used to improve diagnosis, prognosis, treatment, surgery, drug discovery, or for other applications. Therefore, both academia and industry are investing a lot in AI. This review investigates the biomedical literature (in the PubMed and Embase databases) by looking at bibliographical data, observing trends over time and occurrences of keywords. Some observations are made: AI has been growing exponentially over the past few years; it is used mostly for diagnosis; COVID-19 is already in the top-5 of diseases studied using AI; the United States, China, United Kingdom, South Korea and Canada are publishing the most articles in AI research; MIT is the world’s leading university in AI research; and convolutional neural networks are by far the most popular deep learning algorithms at this moment. These trends could be studied in more detail, by studying more literature databases or by including patent databases. More advanced analyses could be used to predict in which direction AI will develop over the coming years. The expectation is that AI will keep on growing, in spite of stricter privacy laws, more need for standardization, bias in the data, and the need for building trust.

show abstract

Section: Discussionmentioning

confidence: 99%

Literature Analysis of Artificial Intelligence in Biomedicine

Hulsen¹

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The properties described above have an important consequence for the applicability and adoption of FAMHE for future real-world multicentric biomedical studies. Indeed, it was established by privacy law experts that data processed using MHE can be considered anonymous data under the General Data Protection Regulation (GDPR) 11 . Therefore, our approach significantly reduces the requirements for contractual agreements and the obligations of data controllers that often hinder multicentric medical studies.…”

Section: /19mentioning

confidence: 99%

“…Several open-source software platforms have been recently developed to provide access to FA distributed algorithms in a way that is packaged for end-users 3,6,7 , e.g., DataSHIELD 6 is an open-source distributed data analysis and machine learning (ML) platform based on the open-source software R. However, none of these platforms addresses the problem of indirect data leakages that stem from the use of "vanilla" federated learning; hence their adoption poses important questions about the actual advantage in terms of eased compliance to regulations over more conventional data-analysis platforms that rely on data centralization [8][9][10] , as partial aggregates and model updates can still be considered personal identifying data 4,[11][12][13][14] .…”

Section: Introductionmentioning

confidence: 99%

“…Furthermore, our approach overcomes the performance limitations of both SMC and HE, as it can scale to a large number of data providers for the execution of complex real-world collaborative federated workflows, without revealing any intermediate information. Our approach can significantly reduce the requirements for contractual agreements and the obligations of data controllers that often hinder multicentric medical studies, because data processed using MHE can be considered anonymous data under the General Data-Protection Regulation (GDPR) 11 .…”

Section: Introductionmentioning

confidence: 99%

“…For example, DataSHIELD 7 is a distributed data analysis and a machine-learning (ML) platform based on the open-source software R. However, none of these platforms address the aforementioned problem of indirect privacy leakages that stem from their use of ‘vanilla’ federated learning. Hence, it remains unclear whether these existing solutions are able to substantially simplify regulatory compliance, compared to more conventional workflows that centralize the data 9–11 , if the partial aggregates and model updates could still be considered as personal identifying data 5, 12–15 .…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Truly Privacy-Preserving Federated Analytics for Precision Medicine with Multiparty Homomorphic Encryption

Froelicher¹,

Troncoso-Pastoriza²,

Raisaro³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

In biomedical research, real-world evidence, which is emerging as an indispensable complement of clinical trials, relies on access to large quantities of patient data that typically reside at separate healthcare institutions. Conventional approaches for centralizing those data are often not feasible due to privacy and security requirements. As a result, more privacy-friendly solutions based on federated analytics are emerging. They enable to simultaneously analyse medical data distributed across a group of connected institutions. However, these techniques do not inherently protect patients' privacy as they require institutions to share intermediate results that can reveal patient-level information. To address this issue, state-of-the-art solutions use additional privacy-preserving measures based on data obfuscation, which often introduce noise in the computation of the final result that can become too inaccurate for precision medicine use cases. We propose FAMHE, a modular system based on multiparty homomorphic encryption, that enables the privacy-preserving execution of federated analytics workflows yielding exact results and without leaking any intermediate information. To demonstrate the maturity of our approach, we reproduce the results of two published state-of-the-art centralized biomedical studies, and we demonstrate that FAMHE enables the efficient, privacy-preserving and decentralized execution of analyses that range from low computational complexity, such as Kaplan-Meier overall survival curves used in oncology, to high computational complexity, such as genome-wide association studies on millions of variants.

show abstract

DeepFixCX: Explainable privacy‐preserving image compression for medical image analysis

Gaudio

Smailagic

Faloutsos

et al. 2023

WIREs Data Min & Knowl

View full text Add to dashboard Cite

Explanations of a model's biases or predictions are essential to medical image analysis. Yet, explainable machine learning approaches for medical image analysis are challenged by needs to preserve privacy of patient data, and by current trends in deep learning to use unsustainably large models and large datasets. We propose DeepFixCX for explainable and privacy‐preserving medical image compression that is nimble and performant. We contribute a review of the field and a conceptual framework for simultaneous privacy and explainability via tools of compression. DeepFixCX compresses images without learning by removing or obscuring spatial and edge information. DeepFixCX is ante‐hoc explainable and gives privatized post hoc explanations of spatial and edge bias without accessing the original image. DeepFixCX privatizes images to prevent image reconstruction and mitigate patient re‐identification. DeepFixCX is nimble. Compression can occur on a laptop CPU or GPU to compress and privatize 1700 images per second of size 320 × 320. DeepFixCX enables use of low memory MLP classifiers for vision data; permitting small performance loss gives end‐to‐end MLP performance over 70× faster and batch size over 100× larger. DeepFixCX consistently improves predictive classification performance of a Deep Neural Network (DNN) by 0.02 AUC ROC on Glaucoma and Cervix Type detection datasets, and can improve multi‐label chest x‐ray classification performance in seven of 10 tested settings. In all three datasets, compression to less than 5% of original number of pixels gives matching or improved performance. Our main novelty is to define an explainability versus privacy problem and address it with lossy compression. This article is categorized under: Fundamental Concepts of Data and Knowledge > Explainable AI Commercial, Legal, and Ethical Issues > Security and Privacy Fundamental Concepts of Data and Knowledge > Big Data Mining

show abstract

Revolutionizing Medical Data Sharing Using Advanced Privacy-Enhancing Technologies: Technical, Legal, and Ethical Synthesis

Cited by 71 publications

References 54 publications

Literature Analysis of Artificial Intelligence in Biomedicine

Literature Analysis of Artificial Intelligence in Biomedicine

Truly Privacy-Preserving Federated Analytics for Precision Medicine with Multiparty Homomorphic Encryption

DeepFixCX: Explainable privacy‐preserving image compression for medical image analysis

Contact Info

Product

Resources

About