Purchase this article with an account.
Allen O Eghrari, Brian Garrett, Aisha Mumtaz, Elyse Joelle McGlumphy, Benjamin W. Iliff, Armand Edalati, S. Amer Riazuddin, John D Gottsch; Automated, software-based objective assessment of Fuchs Corneal Dystrophy severity utilizing retroillumination photography analysis. Invest. Ophthalmol. Vis. Sci. 2015;56(7 ):1640.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Retroillumination photography analysis (RPA) is an objective tool for assessment of the number and distribution of guttae in eyes affected with Fuchs Corneal Dystrophy (FCD). RPA provides a distinct baseline of 0 guttae, in contrast to variable baseline measurements of pachymetry, and can provide high resolution of disease progression. Current protocols include manual processing of images; here we demonstrate a software-based system for automated analysis across various levels of FCD severity.
Retroillumination photographs of ten FCD-affected corneas across various levels of severity were acquired and total counts of guttae summated manually using methods we have previously described. For each cornea, a single image was then loaded into ImageJ software. Processing included reducing color variability and subtracting background noise. Reflection of light from each gutta was identified as a local area of maximum intensity and counted automatically. Background irregularities due to lenticular or tear film abnormalities specific to each individual were adjusted by titrating noise tolerance levels to each cornea. Quality control was conducted by examining a small region of each image with automated overlay to ensure appropriate coverage of individual guttae.
A total of ten retroillumination photographs were analyzed. The Krachmer score, a 1-to-5 grading scale assigned clinically through slit-lamp biomicroscopy, ranged from a severity level of 2 to 5 in the set of analyzed corneas, consistent with mild to severe disease. Automated counts demonstrated a strong linear correlation with manual counts in the set of images (R2=0.93) , each of which ranged from 189 to 9211 guttae.
Software-based, automated RPA allows for rapid grading of FCD severity with high resolution across mild, moderate, and advanced levels of disease severity. This technique may prove beneficial for clinical trials which require precise characterization of disease severity using efficient methods.
This PDF is available to Subscribers Only