Related Experiment Video
Updated: Jul 16, 2026
A Workflow to Quantitatively Determine Age-Related Macular Degeneration Lesion-Specific Variations in Fundus Autofluorescence
Published on: May 26, 2023
Computerized detection and measurement of drusen in age-related macular degeneration
Thomas R Friberg1, Linda Huang, Maria Palaiou
1UPMC Eye Center, Department of Ophthalmology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania 15213, USA.
Background And Objective:
The authors used a drusen detection algorithm to quantitate drusen on digitized images and determined its precision and accuracy.
Patients And Methods:
Fundus images from 349 participants in the Age-Related Eye Disease Study trial were digitized and analyzed with the Drusen Analyzer (IRIDEX Corporation, Mountain View, CA). The size, number, and area of the drusen in two macular regions were computed by two readers using an interactive approach. Measurements were compared to data generated by reading center methods.
Results:
For the Analyzer, inter-observer agreement was 0.79 (SE = 0.02) and 0.86 (SE = 0.01) (kappa statistic). Intra-observer precision was 0.835 and 0.880 (Spearman coefficients). Categorical agreement (weighted kappa) between the Analyzer and reading center results for the two regions was (0.76, 0.58) and (0.68, 0.68) for readers A and B. The time required to quantitate an image using the Analyzer was 105 +/- 52 and 218 +/- 102 seconds for the two regions, respectively.
Conclusions:
Quantitative detection of drusen can be performed reproducibly and efficiently. Comparisons to more labor intensive reading center techniques suggest that the results are similar but that the algorithm is more sensitive and precise.
