New Image Analysis Technique Increases Oximetry Sensitivity

Abstract

Abstract: : Purpose: Optical density (OD) is defined as log₁₀(1/transmission), with transmission equal to transmitted light/illumination. In the case of reflectance oximetry utilizing an 8 bit digital image (0 = black and 255 = white), the illumination has a level of 255. The purpose of this study was to determine the effect of an improved unconventional pigment insensitive method of optical density calculation on reproducibility in retinal oximetry. Methods: All procedures were approved by an IRB, and informed consent obtained. Oximetry images were obtained in 15 normal healthy adults while breathing room air, and after 5 minutes of breathing 100% oxygen. Images were analyzed using a conventional OD calculation, and an innovative OD calculation; replacing the illumination level of 255 with the average reflectance of tissue surrounding blood vessels. OD, therefore, becomes log₁₀(tissue/blood vessel). The increase in venous blood oxygen saturation before and after hyperoxia, as calculated by each technique, was compared by Wilcoxon paired ranked comparison. Results: The conventional calculation of OD resulted in an increase in venous oxygen saturation of 9% (p=0.05). Analysis of the same veins within same images, using an improved OD calculation, found an increase of 35% (p=0.0038), as opposed to 9% with the conventional method. Conclusions: Replacing the 255 constant term with a varying tissue reflectance level increases sensitivity to changes in venous oxygen saturation in retinal oximetry analysis.

Keywords: imaging/image analysis: clinical • metabolism • retina