EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Quality assessment of colour fundus and fluorescein angiography images using deep learning

Michael König, Philipp Seeböck, Bianca S Gerendas iD, Georgios Mylonas, Rudolf Winklhofer, Ioanna Dimakopoulou iD, Ursula Margarethe Schmidt-Erfurth iD

British Journal of Ophthalmology · 2024-01 · 108 (1) · 98-104

DOI: 10.1136/bjo-2022-321963 ↗

Kurz gefasst

Fachartikel in der Zeitschrift British Journal of Ophthalmology, online erschienen am 23.11.2022. Verfasst von Michael König und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Ophthalmology and Optometry, Medical University of Vienna, Wien, Austria. Hauptthemen laut MEDLINE: Deep Learning. Die Arbeit wurde bisher 21-mal zitiert (Crossref) und nennt 25 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background/aims Image quality assessment (IQA) is crucial for both reading centres in clinical studies and routine practice, as only adequate quality allows clinicians to correctly identify diseases and treat patients accordingly. Here we aim to develop a neural network for automated real-time IQA in colour fundus (CF) and fluorescein angiography (FA) images.

Methods

Training and evaluation of two neural networks were conducted using 2272 CF and 2492 FA images, with binary labels in four (contrast, focus, illumination, shadow and reflection) and three (contrast, focus, noise) modality specific categories plus an overall quality ranking. Performance was compared with a second human grader, evaluated on an external public dataset and in a clinical trial use-case.

Results

The networks achieved a F1-score/area under the receiving operator characteristic/precision recall curve of 0.907/0.963/0.966 for CF and 0.822/0.918/0.889 for FA in overall quality prediction with similar results in most categories. A clear relation between model uncertainty and prediction error was observed. In the clinical trial use-case evaluation, the networks achieved an accuracy of 0.930 for CF and 0.895 for FA.

Conclusion

The presented method allows automated IQA in real time, demonstrating human-level performance for CF as well as FA. Such models can help to overcome the problem of human intergrader and intragrader variability by providing objective and reproducible IQA results. It has particular relevance for real-time feedback in multicentre clinical studies, when images are uploaded to central reading centre portals. Moreover, automated IQA as preprocessing step can support integrating automated approaches into clinical practice.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningHumansFluorescein AngiographyColorFundus OculiNeural Networks, Computer

Stichwörter der Autor:innen

Imaging

Autor:innen und Institutionen

  • Department of Ophthalmology and Optometry, Medical University of Vienna, Wien, Austria.
  • Department of Ophthalmology and Optometry, Medical University of Vienna, Wien, Austria ursula.schmidt-erfurth@meduniwien.ac.at.

ORCID-Profile: Bianca S Gerendas, Ioanna Dimakopoulou, Ursula Margarethe Schmidt-Erfurth

Zitieren

APA

König, M., Seeböck, P., Gerendas, B. S., Mylonas, G., Winklhofer, R., Dimakopoulou, I., & Schmidt-Erfurth, U. M. (2024). Quality assessment of colour fundus and fluorescein angiography images using deep learning. British Journal of Ophthalmology, 108(1), 98-104. https://doi.org/10.1136/bjo-2022-321963

Vancouver

König M, Seeböck P, Gerendas BS, Mylonas G, Winklhofer R, Dimakopoulou I, et al. Quality assessment of colour fundus and fluorescein angiography images using deep learning. British Journal of Ophthalmology. 2024;108(1):98-104. doi:10.1136/bjo-2022-321963

BibTeX
@article{Knig2024,
  author = {König, Michael and Seeböck, Philipp and Gerendas, Bianca S and Mylonas, Georgios and Winklhofer, Rudolf and Dimakopoulou, Ioanna and Schmidt-Erfurth, Ursula Margarethe},
  title = {Quality assessment of colour fundus and fluorescein angiography images using deep learning},
  journal = {British Journal of Ophthalmology},
  year = {2024},
  volume = {108},
  number = {1},
  pages = {98--104},
  doi = {10.1136/bjo-2022-321963}
}
Kostenlos starten →