Miller J M, Caprioli J
Yale University School of Medicine, Department of Ophthalmology and Visual Science, New Haven, CT 06510.
Invest Ophthalmol Vis Sci. 1988 Feb;29(2):320-3.
Digitized images of the optic disc, acquired videographically (Rodenstock Analyzer) under green (540 nm) and red (640 nm) illumination, were used to quantify optic disc pallor. The pallor density of each pixel was defined as twice the reflectance under green illumination divided by the sum of the reflectance under red and green illumination. Pallor densities can range from 0 (red) to 1 (white); typical median values were 0.25 for vessels, 0.40 for healthy disc rim, and 0.70 for the lamina cribrosa. The variability of pallor measurements using this technique was determined. Two uniform color fields (Farnsworth-Munsell 100-Hue Test color chips 1 and 18) were imaged five times each using a model eye. Significant (P less than 0.0001) drift of mean pallor densities occurred between images of both sets. The optic discs of seven normal eyes and of seven glaucomatous eyes were imaged nine times each. The non-normal frequency distribution of pallor densities for each image was described by trimmed means and a measure of distribution width. Variability was defined as the standard deviation of the measurements divided by the full scale pallor density. The variability of the trimmed means increased with pallor density (r = 0.99, P less than 0.0001). The variability of distribution width was smaller than that of the mean values, and averaged 3.4% in normal and glaucomatous eyes. Videographic reflectometry may provide useful, quantitative measurements of optic disc pallor.