Cargando…
A deep learning approach to private data sharing of medical images using conditional generative adversarial networks (GANs)
Clinical data sharing can facilitate data-driven scientific research, allowing a broader range of questions to be addressed and thereby leading to greater understanding and innovation. However, sharing biomedical data can put sensitive personal information at risk. This is usually addressed by data...
Autores principales: | , , , , , , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Public Library of Science
2023
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10325103/ https://www.ncbi.nlm.nih.gov/pubmed/37410795 http://dx.doi.org/10.1371/journal.pone.0280316 |
_version_ | 1785069241241698304 |
---|---|
author | Sun, Hanxi Plawinski, Jason Subramaniam, Sajanth Jamaludin, Amir Kadir, Timor Readie, Aimee Ligozio, Gregory Ohlssen, David Baillie, Mark Coroller, Thibaud |
author_facet | Sun, Hanxi Plawinski, Jason Subramaniam, Sajanth Jamaludin, Amir Kadir, Timor Readie, Aimee Ligozio, Gregory Ohlssen, David Baillie, Mark Coroller, Thibaud |
author_sort | Sun, Hanxi |
collection | PubMed |
description | Clinical data sharing can facilitate data-driven scientific research, allowing a broader range of questions to be addressed and thereby leading to greater understanding and innovation. However, sharing biomedical data can put sensitive personal information at risk. This is usually addressed by data anonymization, which is a slow and expensive process. An alternative to anonymization is construction of a synthetic dataset that behaves similar to the real clinical data but preserves patient privacy. As part of a collaboration between Novartis and the Oxford Big Data Institute, a synthetic dataset was generated based on images from COSENTYX® (secukinumab) ankylosing spondylitis (AS) clinical studies. An auxiliary classifier Generative Adversarial Network (ac-GAN) was trained to generate synthetic magnetic resonance images (MRIs) of vertebral units (VUs), conditioned on the VU location (cervical, thoracic and lumbar). Here, we present a method for generating a synthetic dataset and conduct an in-depth analysis on its properties along three key metrics: image fidelity, sample diversity and dataset privacy. |
format | Online Article Text |
id | pubmed-10325103 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2023 |
publisher | Public Library of Science |
record_format | MEDLINE/PubMed |
spelling | pubmed-103251032023-07-07 A deep learning approach to private data sharing of medical images using conditional generative adversarial networks (GANs) Sun, Hanxi Plawinski, Jason Subramaniam, Sajanth Jamaludin, Amir Kadir, Timor Readie, Aimee Ligozio, Gregory Ohlssen, David Baillie, Mark Coroller, Thibaud PLoS One Research Article Clinical data sharing can facilitate data-driven scientific research, allowing a broader range of questions to be addressed and thereby leading to greater understanding and innovation. However, sharing biomedical data can put sensitive personal information at risk. This is usually addressed by data anonymization, which is a slow and expensive process. An alternative to anonymization is construction of a synthetic dataset that behaves similar to the real clinical data but preserves patient privacy. As part of a collaboration between Novartis and the Oxford Big Data Institute, a synthetic dataset was generated based on images from COSENTYX® (secukinumab) ankylosing spondylitis (AS) clinical studies. An auxiliary classifier Generative Adversarial Network (ac-GAN) was trained to generate synthetic magnetic resonance images (MRIs) of vertebral units (VUs), conditioned on the VU location (cervical, thoracic and lumbar). Here, we present a method for generating a synthetic dataset and conduct an in-depth analysis on its properties along three key metrics: image fidelity, sample diversity and dataset privacy. Public Library of Science 2023-07-06 /pmc/articles/PMC10325103/ /pubmed/37410795 http://dx.doi.org/10.1371/journal.pone.0280316 Text en © 2023 Sun et al https://creativecommons.org/licenses/by/4.0/This is an open access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/) , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. |
spellingShingle | Research Article Sun, Hanxi Plawinski, Jason Subramaniam, Sajanth Jamaludin, Amir Kadir, Timor Readie, Aimee Ligozio, Gregory Ohlssen, David Baillie, Mark Coroller, Thibaud A deep learning approach to private data sharing of medical images using conditional generative adversarial networks (GANs) |
title | A deep learning approach to private data sharing of medical images using conditional generative adversarial networks (GANs) |
title_full | A deep learning approach to private data sharing of medical images using conditional generative adversarial networks (GANs) |
title_fullStr | A deep learning approach to private data sharing of medical images using conditional generative adversarial networks (GANs) |
title_full_unstemmed | A deep learning approach to private data sharing of medical images using conditional generative adversarial networks (GANs) |
title_short | A deep learning approach to private data sharing of medical images using conditional generative adversarial networks (GANs) |
title_sort | deep learning approach to private data sharing of medical images using conditional generative adversarial networks (gans) |
topic | Research Article |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10325103/ https://www.ncbi.nlm.nih.gov/pubmed/37410795 http://dx.doi.org/10.1371/journal.pone.0280316 |
work_keys_str_mv | AT sunhanxi adeeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT plawinskijason adeeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT subramaniamsajanth adeeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT jamaludinamir adeeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT kadirtimor adeeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT readieaimee adeeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT ligoziogregory adeeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT ohlssendavid adeeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT bailliemark adeeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT corollerthibaud adeeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT sunhanxi deeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT plawinskijason deeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT subramaniamsajanth deeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT jamaludinamir deeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT kadirtimor deeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT readieaimee deeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT ligoziogregory deeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT ohlssendavid deeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT bailliemark deeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans AT corollerthibaud deeplearningapproachtoprivatedatasharingofmedicalimagesusingconditionalgenerativeadversarialnetworksgans |