June 2022
Volume 63, Issue 7
Open Access
ARVO Annual Meeting Abstract  |   June 2022
Diagnostic Accuracy of Trained Image Graders in Screening for Retinopathy Among Youth with Diabetes
Author Affiliations & Notes
  • Philip Zhou
    Baylor College of Medicine, Houston, Texas, United States
  • Loaah Eltemsah
    Department of Pediatrics, Johns Hopkins Medicine, Baltimore, Maryland, United States
  • Mozhdeh Bahrainian
    UW Health, University of Wisconsin-Madison, Madison, Wisconsin, United States
  • Alvin Liu
    Johns Hopkins Medicine Wilmer Eye Institute, Baltimore, Maryland, United States
  • Laura Prichett
    Department of Pediatrics, Johns Hopkins Medicine, Baltimore, Maryland, United States
  • Risa Wolf
    Department of Pediatrics, Johns Hopkins Medicine, Baltimore, Maryland, United States
  • Roomasa Channa
    UW Health, University of Wisconsin-Madison, Madison, Wisconsin, United States
  • Footnotes
    Commercial Relationships   Philip Zhou None; Loaah Eltemsah None; Mozhdeh Bahrainian None; Alvin Liu None; Laura Prichett None; Risa Wolf None; Roomasa Channa None
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2022, Vol.63, 2166 – F0229. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Philip Zhou, Loaah Eltemsah, Mozhdeh Bahrainian, Alvin Liu, Laura Prichett, Risa Wolf, Roomasa Channa; Diagnostic Accuracy of Trained Image Graders in Screening for Retinopathy Among Youth with Diabetes. Invest. Ophthalmol. Vis. Sci. 2022;63(7):2166 – F0229.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : There is a growing prevalence of diabetes mellitus in the pediatric population, which increases the importance of regular screening for complications such as diabetic retinopathy (DR). The use of digital fundus photography and trained image graders can help increase access to DR screening. Our aim was to assess the diagnostic accuracy of trained image graders (with no prior experience in ophthalmology) in detecting retinopathy from digital fundus photos of pediatric patients with diabetes.

Methods : This cross-sectional analysis compared DR evaluations from two fellowship trained retina specialists and two non-ophthalmologist trained graders. The graders evaluated 90 nonmydriatic fundus photos obtained from pediatric patients with type 1 and type 2 diabetes seen at a multidisciplinary pediatric diabetes center. Trained graders took an online course on DR grading and underwent two hours of training with the retina specialists. A reference standard was determined by agreement between the retina specialists, with discrepancies adjudicated by a third fellowship-trained retina specialist and/or in-person eye exam if there continued to be a question regarding presence of DR. The main outcomes measured were percent agreement with the reference standard, and sensitivity and specificity of detecting DR for trained graders and retina specialists.

Results : Among the 90 patients from whom the retinal images were taken, the average age was 13.8 years old, 52% were non-Hispanic white, 52% were female, and 76% had type 1 diabetes. The prevalence of DR was 13% in this sample. Sensitivity for DR detection varied between 83-91% among trained graders compared to 91-92% for retina specialists. Specificity varied between 67-92% for trained graders compared to 71-83% for retina specialists. Agreement with reference standard varied between 70-91% for trained graders and 74-84% for retina specialists.

Conclusions : After undergoing a standardized training program, trained graders (with no prior experience in ophthalmology) can detect DR with high sensitivity from fundus photos of pediatric patients with diabetes, comparable to that of retina specialists. The use of trained graders in teleretinal screening networks for pediatric diabetes may be a feasible option to improve access to DR screening.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.

×
×

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.

×