Abstract
Purpose: :
A number of different treatments for Stargardt disease (STGD) have been proposed, but how to measure a treatment effect remains a difficult problem. Autofluorescence images were analyzed to determine whether there were clearly measurable features, and the degree of concordance between eyes of each patient.
Methods: :
The images for each patient were deidentified and coded so that each image had a code that was distinct for each eye of the patient. The images were analyzed for quantifiable features. These features were outlined using Adobe Photoshop, and measured using ImageJ. The features deemed measurable were then compared for the two eyes of each patient.
Results: :
Twenty seven patients had adequate autofluorescence images for both eyes. The features used to define lesions were homogeneous dark 64%, homogeneous plus mottled dark 14%, mottled 7%, rings of autofluorescence 14%. The lesions ranged from 0.15 to 28.37 sq mm, with a median of 3 sq mm. The mean difference between eyes was 0.13 sq mm, stdev 1.4. The average absolute difference was 0.84 sq mm. The mean percent difference between eyes was 20%.
Conclusions: :
There are features in the imaging of Stargardt disease that are quantifiable in a reproducible manner. These data will help to assess concordance between the two eyes of a patient. Although these features may not correlate directly with visual function, they are measures of progression of disease and may be used as evidence of treatment efficacy.
Keywords: degenerations/dystrophies • macula/fovea