EUVIMEDEUVIMED entdecken
Multizentrische StudieBeobachtungsstudieOpen Access

Artificial Intelligence‐Based Detection of Central Retinal Artery Occlusion Within 4.5 Hours on Standard Fundus Photographs

Ayse Gungor iD, Ilias Sarbout iD, Aubrey L Gilbert iD, Steffen Hamann, Pierre Lebranchu iD, Cristina Hobeanu, Philippe Gohier iD, Catherine Vignal-Clermont, Oana M Dumitrascu iD, Salomon-Yves Cohen iD, Wolf A Lagrèze iD, Nicolas Feltgen iD und 6 weitere

Journal of the American Heart Association · 2025-07 · 14 (13)

DOI: 10.1161/jaha.124.041441 ↗

Kurz gefasst

Multizentrische Studie in der Zeitschrift Journal of the American Heart Association, 2025-07. Verfasst von Ayse Gungor und 17 weiteren Autor:innen. Beteiligte Einrichtung: Department of Neuro-Ophthalmology Rothschild Foundation Hospital Paris France. Hauptthemen laut MEDLINE: Artificial Intelligence, Deep Learning, Photography, Retinal Artery Occlusion. Die Arbeit wurde bisher 16-mal zitiert (Crossref) und nennt 44 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Prompt diagnosis of acute central retinal artery occlusion (CRAO) is crucial for therapeutic management and stroke prevention. However, most stroke centers lack onsite ophthalmic expertise before considering fibrinolytic treatment. This study aimed to develop, train, and test a deep learning system to detect hyperacute CRAO on retinal fundus photographs within the critical 4.5‐hour treatment window and up to 24 hours after visual loss to aid in secondary stroke prevention.

Methods

Our retrospective, cross‐sectional study included 1322 color fundus photographs from 771 patients with acute visual loss due to CRAO, central retinal vein occlusion, nonarteritic anterior ischemic optic neuropathy, and healthy controls. Photographs were collected from 9 expert neuro‐ophthalmology centers in 6 countries, including 3 randomized clinical trials. Training included 1039 photographs (517 patients), followed by testing on 2 data sets: (1) hyperacute CRAO (54 photographs, 54 patients) and (2) CRAO within 24 hours after visual loss (110 photographs, 109 patients).

Results

The deep learning system achieved an area under the receiver operating characteristic curve of 0.96 (95% confidence interval (CI), 0.95–0.98), a sensitivity of 92.6% (95% CI, 87.0–98.0), and a specificity of 85.0% (95% CI, 81.8–92.8) for detecting CRAO at hyperacute stage, with similar results within 24 hours. The deep learning system outperformed stroke neurologists on a subset of hyperacute testing data set (120 photographs, 120 patients).

Conclusions

A deep learning system can accurately detect hyperacute CRAO on retinal photographs within a time window compatible with urgent fibrinolysis. If further validated, such systems could improve patient selection for fibrinolytic trials and optimize secondary stroke prevention. Registration URL: https://www.clinicaltrials.gov ; Unique identifier: NCT06390579.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Artificial IntelligenceDeep LearningPhotographyRetinal Artery OcclusionAgedFemaleHumansMaleMiddle AgedCross-Sectional StudiesFundus OculiPredictive Value of TestsRetrospective StudiesTime Factors

Stichwörter der Autor:innen

acute strokeartificial intelligencecentral retinal artery occlusioncerebrovascular strokeearly diagnosismachine learningvisual loss

Autor:innen und Institutionen

  • Department of Neuro-Ophthalmology Rothschild Foundation Hospital Paris France.
  • Rothschild Computational and Visual Neurosciences Laboratory Rothschild Foundation Hospital Paris France.
  • Sorbonne Université, Institut du Cerveau-Paris Brain Institute-ICM, Inria, Inserm, CNRS, APHP, Hôpital de la Pitié Salpêtrière Paris France.
  • Kaiser Permanente Northern California Vallejo California USA.
  • Department of Ophthalmology Copenhagen University Hospital-Rigshospitalet Glostrup Denmark.
  • Department of Clinical Medicine University of Copenhagen Copenhagen Denmark.

ORCID-Profile: Ayse Gungor, Ilias Sarbout, Aubrey L Gilbert, Pierre Lebranchu, Philippe Gohier, Oana M Dumitrascu, Salomon-Yves Cohen, Wolf A Lagrèze, Nicolas Feltgen, Cédric Lamirel, Jost B Jonas, Michael Obadia, Daniel Racoceanu, Dan Milea

Zitieren

APA

Gungor, A., Sarbout, I., Gilbert, A. L., Hamann, S., Lebranchu, P., Hobeanu, C., Gohier, P., Vignal-Clermont, C., Dumitrascu, O. M., Cohen, S. Y., Lagrèze, W. A., Feltgen, N., van der Heide, F., Lamirel, C., Jonas, J. B., Obadia, M., Racoceanu, D., & Milea, D. (2025). Artificial Intelligence‐Based Detection of Central Retinal Artery Occlusion Within 4.5 Hours on Standard Fundus Photographs. Journal of the American Heart Association, 14(13). https://doi.org/10.1161/jaha.124.041441

Vancouver

Gungor A, Sarbout I, Gilbert AL, Hamann S, Lebranchu P, Hobeanu C, et al. Artificial Intelligence‐Based Detection of Central Retinal Artery Occlusion Within 4.5 Hours on Standard Fundus Photographs. Journal of the American Heart Association. 2025;14(13). doi:10.1161/jaha.124.041441

BibTeX
@article{Gungor2025,
  author = {Gungor, Ayse and Sarbout, Ilias and Gilbert, Aubrey L and Hamann, Steffen and Lebranchu, Pierre and Hobeanu, Cristina and Gohier, Philippe and Vignal-Clermont, Catherine and Dumitrascu, Oana M and Cohen, Salomon-Yves and Lagrèze, Wolf A and Feltgen, Nicolas and van der Heide, Frank and Lamirel, Cédric and Jonas, Jost B and Obadia, Michael and Racoceanu, Daniel and Milea, Dan},
  title = {Artificial Intelligence‐Based Detection of Central Retinal Artery Occlusion Within 4.5 Hours on Standard Fundus Photographs},
  journal = {Journal of the American Heart Association},
  year = {2025},
  volume = {14},
  number = {13},
  doi = {10.1161/jaha.124.041441}
}
Kostenlos starten →