학술논문
Encrypted federated learning for secure decentralized collaboration in cancer image analysis.
Document Type
Academic Journal
Author
Truhn D; Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany. Electronic address: dtruhn@ukaachen.de.; Tayebi Arasteh S; Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany.; Saldanha OL; Department of Medicine III, University Hospital RWTH Aachen, Aachen, Germany; Else Kroener Fresenius Center for Digital Health, Medical Faculty Carl Gustav Carus, Technical University Dresden, Dresden, Germany.; Müller-Franzes G; Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany.; Khader F; Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany.; Quirke P; Division of Pathology and Data Analytics, Leeds Institute of Medical Research at St James's, University of Leeds, Leeds, United Kingdom.; West NP; Division of Pathology and Data Analytics, Leeds Institute of Medical Research at St James's, University of Leeds, Leeds, United Kingdom.; Gray R; Clinical Trial Service Unit, University of Oxford, Oxford, United Kingdom.; Hutchins GGA; Division of Pathology and Data Analytics, Leeds Institute of Medical Research at St James's, University of Leeds, Leeds, United Kingdom.; James JA; Precision Medicine Centre of Excellence, Health Sciences Building, The Patrick G Johnston Centre for Cancer Research, Queen's University Belfast, Belfast, United Kingdom; Regional Molecular Diagnostic Service, Belfast Health and Social Care Trust, Belfast, United Kingdom; The Patrick G Johnston Centre for Cancer Research, Queen's University Belfast, United Kingdom.; Loughrey MB; The Patrick G Johnston Centre for Cancer Research, Queen's University Belfast, United Kingdom; Department of Cellular Pathology, Belfast Health and Social Care Trust, Belfast, United Kingdom; Centre for Public Health, Queen's University Belfast, Belfast, United Kingdom.; Salto-Tellez M; Precision Medicine Centre of Excellence, Health Sciences Building, The Patrick G Johnston Centre for Cancer Research, Queen's University Belfast, Belfast, United Kingdom; Regional Molecular Diagnostic Service, Belfast Health and Social Care Trust, Belfast, United Kingdom; The Patrick G Johnston Centre for Cancer Research, Queen's University Belfast, United Kingdom.; Brenner H; Division of Clinical Epidemiology and Aging Research, German Cancer Research Center (DKFZ), Heidelberg, Germany; Division of Preventive Oncology, German Cancer Research Center (DKFZ) and National Center for Tumor Diseases (NCT), Heidelberg, Germany; German Cancer Consortium (DKTK), German Cancer Research Center (DKFZ), Heidelberg, Germany.; Brobeil A; Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany; Tissue Bank, National Center for Tumor Diseases (NCT), University Hospital Heidelberg, Heidelberg, Germany.; Yuan T; Division of Clinical Epidemiology and Aging Research, German Cancer Research Center (DKFZ), Heidelberg, Germany; Medical Faculty Heidelberg, Heidelberg University, Heidelberg, Germany.; Chang-Claude J; Cancer Epidemiology Group, University Cancer Center Hamburg, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; Division of Cancer Epidemiology, German Cancer Research Center (DKFZ), Heidelberg, Germany.; Hoffmeister M; Division of Clinical Epidemiology and Aging Research, German Cancer Research Center (DKFZ), Heidelberg, Germany.; Foersch S; Institute of Pathology, University Medical Center Mainz, Mainz, Germany.; Han T; Physics of Molecular Imaging Systems, Experimental Molecular Imaging, RWTH Aachen University, Aachen, Germany.; Keil S; Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany.; Schulze-Hagen M; Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany.; Isfort P; Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany.; Bruners P; Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany.; Kaissis G; Institute of Diagnostic and Interventional Radiology, Technical University of Munich, Munich, Germany; Artificial Intelligence in Medicine and Healthcare, Technical University of Munich, Munich, Germany; Department of Computing, Imperial College London, London, United Kingdom.; Kuhl C; Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany.; Nebelung S; Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany.; Kather JN; Department of Medicine III, University Hospital RWTH Aachen, Aachen, Germany; Else Kroener Fresenius Center for Digital Health, Medical Faculty Carl Gustav Carus, Technical University Dresden, Dresden, Germany; Division of Pathology and Data Analytics, Leeds Institute of Medical Research at St James's, University of Leeds, Leeds, United Kingdom; Medical Oncology, National Center for Tumor Diseases (NCT), University Hospital Heidelberg, Heidelberg, Germany.
Source
Publisher: Elsevier Country of Publication: Netherlands NLM ID: 9713490 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1361-8423 (Electronic) Linking ISSN: 13618415 NLM ISO Abbreviation: Med Image Anal Subsets: MEDLINE
Subject
Language
English
Abstract
Artificial intelligence (AI) has a multitude of applications in cancer research and oncology. However, the training of AI systems is impeded by the limited availability of large datasets due to data protection requirements and other regulatory obstacles. Federated and swarm learning represent possible solutions to this problem by collaboratively training AI models while avoiding data transfer. However, in these decentralized methods, weight updates are still transferred to the aggregation server for merging the models. This leaves the possibility for a breach of data privacy, for example by model inversion or membership inference attacks by untrusted servers. Somewhat-homomorphically-encrypted federated learning (SHEFL) is a solution to this problem because only encrypted weights are transferred, and model updates are performed in the encrypted space. Here, we demonstrate the first successful implementation of SHEFL in a range of clinically relevant tasks in cancer image analysis on multicentric datasets in radiology and histopathology. We show that SHEFL enables the training of AI models which outperform locally trained models and perform on par with models which are centrally trained. In the future, SHEFL can enable multiple institutions to co-train AI models without forsaking data governance and without ever transmitting any decryptable data to untrusted servers.
Competing Interests: Declaration of Competing Interest The Authors declare no competing financial or non-financial interests. For transparency, we provide the following information: JNK declares consulting services for Owkin, France, DoMore Diagnostics, Norway, Panakeia, UK, Scailyte, Switzerland, Cancilico, Germany, Mindpeak, Germany, and Histofy, UK; furthermore he holds shares in StratifAI GmbH, Germany, and has received honoraria for lectures by AstraZeneca, Bayer, Eisai, MSD, BMS, Roche, Pfizer and Fresenius. DT holds shares in StraifAI GmbH, Germany and received honoraria for lectures by Bayer. PQ and NW declare research funding from Roche and PQ consulting and speaker services for Roche. MST has recently received honoraria for advisory work in relation to the following companies: Incyte, MindPeak, MSD, BMS and Sonrai; these are all unrelated to this work. No other potential conflicts of interest are reported by any of the authors. The authors received advice from NVIDIA when performing this study, but NVIDIA did not have any role in study design, conducting the experiments, interpretation of the results or decision to submit for publication.
(Copyright © 2023 The Authors. Published by Elsevier B.V. All rights reserved.)
Competing Interests: Declaration of Competing Interest The Authors declare no competing financial or non-financial interests. For transparency, we provide the following information: JNK declares consulting services for Owkin, France, DoMore Diagnostics, Norway, Panakeia, UK, Scailyte, Switzerland, Cancilico, Germany, Mindpeak, Germany, and Histofy, UK; furthermore he holds shares in StratifAI GmbH, Germany, and has received honoraria for lectures by AstraZeneca, Bayer, Eisai, MSD, BMS, Roche, Pfizer and Fresenius. DT holds shares in StraifAI GmbH, Germany and received honoraria for lectures by Bayer. PQ and NW declare research funding from Roche and PQ consulting and speaker services for Roche. MST has recently received honoraria for advisory work in relation to the following companies: Incyte, MindPeak, MSD, BMS and Sonrai; these are all unrelated to this work. No other potential conflicts of interest are reported by any of the authors. The authors received advice from NVIDIA when performing this study, but NVIDIA did not have any role in study design, conducting the experiments, interpretation of the results or decision to submit for publication.
(Copyright © 2023 The Authors. Published by Elsevier B.V. All rights reserved.)