June 2021
Volume 62, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2021
Hierarchical curriculum learning for robust automated detection of low-prevalence retinal disease features: application to reticular pseudodrusen
Author Affiliations & Notes
  • Cristina González-Gonzalo
    A-Eye Research Group, Informatics Institute, Universiteit van Amsterdam, Amsterdam, Noord-Holland, Netherlands
    Diagnostic Image Analysis Group, Department of Radiology and Nuclear Medicine, Radboudumc, Nijmegen, Gelderland, Netherlands
  • Eric F. Thee
    Department of Ophthalmology and Epidemiology, Erasmus MC, Rotterdam, Zuid-Holland, Netherlands
  • Bart Liefers
    Department of Ophthalmology and Epidemiology, Erasmus MC, Rotterdam, Zuid-Holland, Netherlands
    Moorfields Eye Hospital NHS Foundation Trust, London, London, United Kingdom
  • Coen de Vente
    A-Eye Research Group, Informatics Institute, Universiteit van Amsterdam, Amsterdam, Noord-Holland, Netherlands
    Diagnostic Image Analysis Group, Department of Radiology and Nuclear Medicine, Radboudumc, Nijmegen, Gelderland, Netherlands
  • Caroline C. W. Klaver
    Department of Ophthalmology and Epidemiology, Erasmus MC, Rotterdam, Zuid-Holland, Netherlands
    Department of Ophthalmology, Radboudumc, Nijmegen, Gelderland, Netherlands
  • Clara I. Sánchez
    A-Eye Research Group, Informatics Institute, Universiteit van Amsterdam, Amsterdam, Noord-Holland, Netherlands
    Department of Biomedical Engineering and Physics, Amsterdam UMC Locatie AMC, Amsterdam, Noord-Holland, Netherlands
  • Footnotes
    Commercial Relationships   Cristina González-Gonzalo, None; Eric Thee, None; Bart Liefers, None; Coen de Vente, None; Caroline Klaver, None; Clara Sánchez, None
  • Footnotes
    Support  NWO STW P15-26
Investigative Ophthalmology & Visual Science June 2021, Vol.62, 86. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Cristina González-Gonzalo, Eric F. Thee, Bart Liefers, Coen de Vente, Caroline C. W. Klaver, Clara I. Sánchez; Hierarchical curriculum learning for robust automated detection of low-prevalence retinal disease features: application to reticular pseudodrusen. Invest. Ophthalmol. Vis. Sci. 2021;62(8):86.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : The low prevalence of certain retinal disease features compromises data collection for deep neural networks (DNN) development and, consequently, the benefits of automated detection. We robustify the detection of such features in scarce data settings by exploiting hierarchical information available in the data to learn from generic to specific, low-prevalence features. We focus on reticular pseudodrusen (RPD), a hallmark of intermediate age-related macular degeneration (AMD).

Methods : Color fundus images (CFI) from the AREDS dataset were used for DNN development (106,994 CFI) and testing (27,066 CFI). An external test set (RS1-6) was generated with 2,790 CFI from the Rotterdam Study. In both datasets CFI were graded from generic to specific features. This allows to establish a hierarchy of binary classification tasks with decreasing prevalence: presence of AMD findings (AREDS prevalence: 88%; RS1-6: 77%), drusen (85%; 73%), large drusen (40%; 24%), RPD (1%; 4%). We created a hierarchical curriculum and developed a DNN (HC-DNN) that learned each task sequentially. We computed its performance for RPD detection in both test sets and compared it to a baseline DNN (B-DNN) that learned to detect RPD from scratch disregarding hierarchical information. We studied their robustness across datasets, while reducing the size of data available for development (same prevalences).

Results : Area under the receiver operating characteristic curve (AUC) was used to measure RPD detection performance. When large development data were available, there was no significant difference between DNNs (100% data, HC-DNN: 0.96 (95% CI, 0.94-0.97) in AREDS, 0.82 (0.78-0.86) in RS1-6; B-DNN: 0.95 (0.94-0.96) in AREDS, 0.83 (0.79-0.87) in RS1-6). However, HC-DNN achieved better performance and robustness across datasets when development data were highly reduced (<50% data, p-values<0.05) (1% data, HC-DNN: 0.63 (0.60-0.66) in AREDS, 0.76 (0.72-0.80) in RS1-6; B-DNN: 0.53 (0.49-0.56) in AREDS, 0.48 (0.42-0.53) in RS1-6).

Conclusions : Hierarchical curriculum learning allows for knowledge transfer from general, higher-prevalence features and becomes beneficial for the detection of low-prevalence retinal features, such as RPD, in scarce data settings. Moreover, exploiting hierarchical information improves DNN robustness across datasets.

This is a 2021 ARVO Annual Meeting abstract.

 

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×