Publication:
Artificial intelligence to detect papilledema from ocular fundus photographs

Issued Date

2020-04-30

Resource Type

Article

ISSN

15334406
00284793

DOI

10.1056/NEJMoa1917130

Other identifier(s)

2-s2.0-85084305461

Rights

Mahidol University

Rights Holder(s)

SCOPUS

Bibliographic Citation

New England Journal of Medicine. Vol.382, No.18 (2020), 1687-1695

Suggested Citation

Dan Milea, Raymond P. Najjar, Jiang Zhubo, Daniel Ting, Caroline Vasseneix, Xinxing Xu, Masoud Aghsaei Fard, Pedro Fonseca, Kavin Vanikieti, Wolf A. Lagrèze, Chiara La Morgia, Carol Y. Cheung, Steffen Hamann, Christophe Chiquet, Nicolae Sanda, Hui Yang, Luis J. Mejico, Marie Bénédicte Rougier, Richard Kho, Tran Thi Ha Chau, Shweta Singhal, Philippe Gohier, Catherine Clermont-Vignal, Ching Yu Cheng, Jost B. Jonas, Patrick Yu-Wai-Man, Clare L. Fraser, John J. Chen, Selvakumar Ambika, Neil R. Miller, Yong Liu, Nancy J. Newman, Tien Y. Wong, Valérie Biousse Artificial intelligence to detect papilledema from ocular fundus photographs. New England Journal of Medicine. Vol.382, No.18 (2020), 1687-1695. doi:10.1056/NEJMoa1917130 Retrieved from: https://repository.li.mahidol.ac.th/handle/20.500.14594/56251

Research Projects

Organizational Units

Authors

Journal Issue

Thesis

Title

Artificial intelligence to detect papilledema from ocular fundus photographs

Author(s)

Dan Milea
 Raymond P. Najjar
 Jiang Zhubo
 Daniel Ting
 Caroline Vasseneix
 Xinxing Xu
 Masoud Aghsaei Fard
 Pedro Fonseca
 Kavin Vanikieti
 Wolf A. Lagrèze
 Chiara La Morgia
 Carol Y. Cheung
 Steffen Hamann
 Christophe Chiquet
 Nicolae Sanda
 Hui Yang
 Luis J. Mejico
 Marie Bénédicte Rougier
 Richard Kho
 Tran Thi Ha Chau
 Shweta Singhal
 Philippe Gohier
 Catherine Clermont-Vignal
 Ching Yu Cheng
 Jost B. Jonas
 Patrick Yu-Wai-Man
 Clare L. Fraser
 John J. Chen
 Selvakumar Ambika
 Neil R. Miller
 Yong Liu
 Nancy J. Newman
 Tien Y. Wong
 Valérie Biousse

Other Contributor(s)

Farabi Eye Hospital
 John van Geest Centre for Brain Repair
 Istituto delle Scienze Neurologiche di Bologna
 Universite Grenoble Alpes
 Duke-NUS Medical School Singapore
 Fondation Adolphe de Rothschild
 University of Cambridge
 Københavns Universitet
 Alma Mater Studiorum Università di Bologna
 Yong Loo Lin School of Medicine
 Medical Research Foundation, Chennai
 Universität Freiburg im Breisgau
 Singapore Eye Research Institute
 Universite Catholique de Lille
 Sun Yat-Sen University
 Universidade de Coimbra
 Centro Hospitalar e Universitário de Coimbra
 State University of New York Upstate Medical University
 Universität Heidelberg
 Moorfields Eye Hospital NHS Foundation Trust
 Faculty of Medicine, Ramathibodi Hospital, Mahidol University
 Save Sight Institute
 Hôpitaux universitaires de Genève
 CHU Angers
 Singapore National Eye Centre
 Centre Hospitalier Universitaire de Grenoble
 Mayo Clinic
 A-Star, Institute of High Performance Computing
 Groupe Hospitalier Pellegrin
 Chinese University of Hong Kong
 Emory University School of Medicine
 Johns Hopkins School of Medicine
 American Eye Center

Abstract

Copyright © 2020 Massachusetts Medical Society. BACKGROUND Nonophthalmologist physicians do not confidently perform direct ophthalmoscopy. The use of artificial intelligence to detect papilledema and other optic-disk abnormalities from fundus photographs has not been well studied. METHODS We trained, validated, and externally tested a deep-learning system to classify optic disks as being normal or having papilledema or other abnormalities from 15,846 retrospectively collected ocular fundus photographs that had been obtained with pharmacologic pupillary dilation and various digital cameras in persons from multiple ethnic populations. Of these photographs, 14,341 from 19 sites in 11 countries were used for training and validation, and 1505 photographs from 5 other sites were used for external testing. Performance at classifying the optic-disk appearance was evaluated by calculating the area under the receiver-operating-characteristic curve (AUC), sensitivity, and specificity, as compared with a reference standard of clinical diagnoses by neuro-ophthalmologists. RESULTS The training and validation data sets from 6779 patients included 14,341 photographs: 9156 of normal disks, 2148 of disks with papilledema, and 3037 of disks with other abnormalities. The percentage classified as being normal ranged across sites from 9.8 to 100%; the percentage classified as having papilledema ranged across sites from zero to 59.5%. In the validation set, the system discriminated disks with papilledema from normal disks and disks with nonpapilledema abnormalities with an AUC of 0.99 (95% confidence interval [CI], 0.98 to 0.99) and normal from abnormal disks with an AUC of 0.99 (95% CI, 0.99 to 0.99). In the external-testing data set of 1505 photographs, the system had an AUC for the detection of papilledema of 0.96 (95% CI, 0.95 to 0.97), a sensitivity of 96.4% (95% CI, 93.9 to 98.3), and a specificity of 84.7% (95% CI, 82.3 to 87.1). CONCLUSIONS A deep-learning system using fundus photographs with pharmacologically dilated pupils differentiated among optic disks with papilledema, normal disks, and disks with nonpapilledema abnormalities.

Keyword(s)

Medicine

Availability

URI

https://repository.li.mahidol.ac.th/handle/20.500.14594/56251

Collections

Scopus 2020