Medical image synthesis for data augmentation and anonymization using generative adversarial networks

Hoo Chang Shin; Neil A. Tenenholtz; Jameson K. Rogers; Christopher G. Schwarz; Matthew L. Senjem; Jeffrey L. Gunter; Katherine P. Andriole; Mark Michalski

doi:10.1007/978-3-030-00536-8_1

Medical image synthesis for data augmentation and anonymization using generative adversarial networks

Hoo Chang Shin, Neil A. Tenenholtz, Jameson K. Rogers, Christopher G. Schwarz, Matthew L. Senjem, Jeffrey L. Gunter, Katherine P. Andriole, Mark Michalski

Radiology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

140 Scopus citations

Abstract

Data diversity is critical to success when training deep learning models. Medical imaging data sets are often imbalanced as pathologic findings are generally rare, which introduces significant challenges when training deep learning models. In this work, we propose a method to generate synthetic abnormal MRI images with brain tumors by training a generative adversarial network using two publicly available data sets of brain MRI. We demonstrate two unique benefits that the synthetic images provide. First, we illustrate improved performance on tumor segmentation by leveraging the synthetic images as a form of data augmentation. Second, we demonstrate the value of generative models as an anonymization tool, achieving comparable tumor segmentation results when trained on the synthetic data versus when trained on real subject data. Together, these results offer a potential solution to two of the largest challenges facing machine learning in medical imaging, namely the small incidence of pathological findings, and the restrictions around sharing of patient data.

Original language	English (US)
Title of host publication	Simulation and Synthesis in Medical Imaging - Third International Workshop, SASHIMI 2018, Held in Conjunction with MICCAI 2018, Proceedings
Editors	Orcun Goksel, Ipek Oguz, Ali Gooya, Ninon Burgos
Publisher	Springer Verlag
Pages	1-11
Number of pages	11
ISBN (Print)	9783030005351
DOIs	https://doi.org/10.1007/978-3-030-00536-8_1
State	Published - 2018
Event	3rd International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2018 Held in Conjunction with 21st International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2018 - Granada, Spain Duration: Sep 16 2018 → Sep 16 2018

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11037 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	3rd International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2018 Held in Conjunction with 21st International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2018
Country/Territory	Spain
City	Granada
Period	9/16/18 → 9/16/18

Keywords

Brain tumor
Deep learning
GAN
Generative models
Image synthesis
MRI
Magnetic resonance imaging
Segmentation

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-030-00536-8_1

Cite this

Shin, H. C., Tenenholtz, N. A., Rogers, J. K., Schwarz, C. G., Senjem, M. L., Gunter, J. L., Andriole, K. P., & Michalski, M. (2018). Medical image synthesis for data augmentation and anonymization using generative adversarial networks. In O. Goksel, I. Oguz, A. Gooya, & N. Burgos (Eds.), Simulation and Synthesis in Medical Imaging - Third International Workshop, SASHIMI 2018, Held in Conjunction with MICCAI 2018, Proceedings (pp. 1-11). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11037 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-00536-8_1

Medical image synthesis for data augmentation and anonymization using generative adversarial networks. / Shin, Hoo Chang; Tenenholtz, Neil A.; Rogers, Jameson K. et al.
Simulation and Synthesis in Medical Imaging - Third International Workshop, SASHIMI 2018, Held in Conjunction with MICCAI 2018, Proceedings. ed. / Orcun Goksel; Ipek Oguz; Ali Gooya; Ninon Burgos. Springer Verlag, 2018. p. 1-11 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11037 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Shin, HC, Tenenholtz, NA, Rogers, JK, Schwarz, CG, Senjem, ML, Gunter, JL, Andriole, KP & Michalski, M 2018, Medical image synthesis for data augmentation and anonymization using generative adversarial networks. in O Goksel, I Oguz, A Gooya & N Burgos (eds), Simulation and Synthesis in Medical Imaging - Third International Workshop, SASHIMI 2018, Held in Conjunction with MICCAI 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11037 LNCS, Springer Verlag, pp. 1-11, 3rd International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2018 Held in Conjunction with 21st International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2018, Granada, Spain, 9/16/18. https://doi.org/10.1007/978-3-030-00536-8_1

Shin HC, Tenenholtz NA, Rogers JK, Schwarz CG, Senjem ML, Gunter JL et al. Medical image synthesis for data augmentation and anonymization using generative adversarial networks. In Goksel O, Oguz I, Gooya A, Burgos N, editors, Simulation and Synthesis in Medical Imaging - Third International Workshop, SASHIMI 2018, Held in Conjunction with MICCAI 2018, Proceedings. Springer Verlag. 2018. p. 1-11. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-00536-8_1

Shin, Hoo Chang ; Tenenholtz, Neil A. ; Rogers, Jameson K. et al. / Medical image synthesis for data augmentation and anonymization using generative adversarial networks. Simulation and Synthesis in Medical Imaging - Third International Workshop, SASHIMI 2018, Held in Conjunction with MICCAI 2018, Proceedings. editor / Orcun Goksel ; Ipek Oguz ; Ali Gooya ; Ninon Burgos. Springer Verlag, 2018. pp. 1-11 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{a4fd095c0cf340378b415096426a2c45,

title = "Medical image synthesis for data augmentation and anonymization using generative adversarial networks",

abstract = "Data diversity is critical to success when training deep learning models. Medical imaging data sets are often imbalanced as pathologic findings are generally rare, which introduces significant challenges when training deep learning models. In this work, we propose a method to generate synthetic abnormal MRI images with brain tumors by training a generative adversarial network using two publicly available data sets of brain MRI. We demonstrate two unique benefits that the synthetic images provide. First, we illustrate improved performance on tumor segmentation by leveraging the synthetic images as a form of data augmentation. Second, we demonstrate the value of generative models as an anonymization tool, achieving comparable tumor segmentation results when trained on the synthetic data versus when trained on real subject data. Together, these results offer a potential solution to two of the largest challenges facing machine learning in medical imaging, namely the small incidence of pathological findings, and the restrictions around sharing of patient data.",

keywords = "Brain tumor, Deep learning, GAN, Generative models, Image synthesis, MRI, Magnetic resonance imaging, Segmentation",

author = "Shin, {Hoo Chang} and Tenenholtz, {Neil A.} and Rogers, {Jameson K.} and Schwarz, {Christopher G.} and Senjem, {Matthew L.} and Gunter, {Jeffrey L.} and Andriole, {Katherine P.} and Mark Michalski",

note = "Publisher Copyright: {\textcopyright} 2018, Springer Nature Switzerland AG.; 3rd International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2018 Held in Conjunction with 21st International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2018 ; Conference date: 16-09-2018 Through 16-09-2018",

year = "2018",

doi = "10.1007/978-3-030-00536-8_1",

language = "English (US)",

isbn = "9783030005351",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "1--11",

editor = "Orcun Goksel and Ipek Oguz and Ali Gooya and Ninon Burgos",

booktitle = "Simulation and Synthesis in Medical Imaging - Third International Workshop, SASHIMI 2018, Held in Conjunction with MICCAI 2018, Proceedings",

}

TY - GEN

T1 - Medical image synthesis for data augmentation and anonymization using generative adversarial networks

AU - Shin, Hoo Chang

AU - Tenenholtz, Neil A.

AU - Rogers, Jameson K.

AU - Schwarz, Christopher G.

AU - Senjem, Matthew L.

AU - Gunter, Jeffrey L.

AU - Andriole, Katherine P.

AU - Michalski, Mark

PY - 2018

Y1 - 2018

N2 - Data diversity is critical to success when training deep learning models. Medical imaging data sets are often imbalanced as pathologic findings are generally rare, which introduces significant challenges when training deep learning models. In this work, we propose a method to generate synthetic abnormal MRI images with brain tumors by training a generative adversarial network using two publicly available data sets of brain MRI. We demonstrate two unique benefits that the synthetic images provide. First, we illustrate improved performance on tumor segmentation by leveraging the synthetic images as a form of data augmentation. Second, we demonstrate the value of generative models as an anonymization tool, achieving comparable tumor segmentation results when trained on the synthetic data versus when trained on real subject data. Together, these results offer a potential solution to two of the largest challenges facing machine learning in medical imaging, namely the small incidence of pathological findings, and the restrictions around sharing of patient data.

AB - Data diversity is critical to success when training deep learning models. Medical imaging data sets are often imbalanced as pathologic findings are generally rare, which introduces significant challenges when training deep learning models. In this work, we propose a method to generate synthetic abnormal MRI images with brain tumors by training a generative adversarial network using two publicly available data sets of brain MRI. We demonstrate two unique benefits that the synthetic images provide. First, we illustrate improved performance on tumor segmentation by leveraging the synthetic images as a form of data augmentation. Second, we demonstrate the value of generative models as an anonymization tool, achieving comparable tumor segmentation results when trained on the synthetic data versus when trained on real subject data. Together, these results offer a potential solution to two of the largest challenges facing machine learning in medical imaging, namely the small incidence of pathological findings, and the restrictions around sharing of patient data.

KW - Brain tumor

KW - Deep learning

KW - GAN

KW - Generative models

KW - Image synthesis

KW - MRI

KW - Magnetic resonance imaging

KW - Segmentation

UR - http://www.scopus.com/inward/record.url?scp=85053935181&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053935181&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-00536-8_1

DO - 10.1007/978-3-030-00536-8_1

M3 - Conference contribution

AN - SCOPUS:85053935181

SN - 9783030005351

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 1

EP - 11

BT - Simulation and Synthesis in Medical Imaging - Third International Workshop, SASHIMI 2018, Held in Conjunction with MICCAI 2018, Proceedings

A2 - Goksel, Orcun

A2 - Oguz, Ipek

A2 - Gooya, Ali

A2 - Burgos, Ninon

PB - Springer Verlag

T2 - 3rd International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2018 Held in Conjunction with 21st International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2018

Y2 - 16 September 2018 through 16 September 2018

ER -

Medical image synthesis for data augmentation and anonymization using generative adversarial networks

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this