Technical and topographical variability in automatic retinal oximetry

Abstract

AbstractPurpose To assess if retinal vessel oxygen saturation measurements are affected by (1) which eye is measured (2) photographic flashes (3) image angle (4) location within the retina (quadrant).Methods The Oxymap retinal oximeter simultaneously acquires fundus images at 570nm and 600nm and calculates relative oxygen saturation in retinal vessels. Oximetry images were obtained from 26 healthy individuals, 18‐30 years. Images were taken of one eye at different angles and at the same angle with at least two photographic flashes in between. In addition, one image from the other eye was obtained. Retinal vessel oxygen saturation was measured with the Oxymap analysis software.Results (1) No difference was found between the left and right eyes (n=24, p>0.68). (2) Saturation increased by 0.5±1.0%, (p=0.028, mean±SD, hemoglobin saturation percentage) in arterioles and 0.9±1.9% (p=0.020) in venules after three or more photographic flashes. (3) Saturation was higher in both arterioles (1.3±1.7%, p=0.0004) and venules (1.9±2.4%, p=0.0007) when the subject gazed down than when looking straight. (4) Significant topographical variations were found within the retina, up to ‐3.1±2.4% difference from average saturation in arterioles and ‐5.4±4.0% in venules.Conclusion The left and right eye are comparable in retinal vessel oxygen saturation measurements. A few photographic flashes have a small effect on the measurement. The difference in saturation by angle is significant when the subject is looking down and the topographical difference in saturation is notable between quadrants. It is important to standardise image acquistion and analysis. Commercial interest