Analyzing diameter artifacts in oximetry measures of retinal oxygen saturation

Abstract

Purpose : Metabolic assessment in ophthalmic diseases is an important, yet largely unrealized, target due to limitations of imaging and tissue target acquisition. Dual-wavelength photographic retinal oximetry measures have identified possible deficits of oxygen metabolism in glaucoma, yet often they yield non-physiological high values of retinal oxygen saturation in small vessels. This study highlights the possible contribution of single and double pass light transmission to the observed diameter artifacts.

Methods : Retinal photographic oximetry images of healthy subjects (N = 10) were obtained with the Oxymap T1 retinal oximeter (Oxymap Analyzer, Iceland). Saturation, optical density (OD), optical density ratio (ODR), and vessel diameter (D) were obtained from arteriolar and venular segments (>30 pixel length) covering a diameter range of 30-200 µm. A conceptual model of light returned from vessels by single and double pass transmission was developed based on optical properties of blood and retinal reflectance to predict their potential contribution to the diameter artifact.

Results : In Fig. 1A, venular saturation showed a greater dependence on D (slope of -0.18) than arteriolar saturation (slope of -0.03). The artifact was more pronounced in small diameter vessels (slope was -0.22 for D <100 µm and -0.07 for D>100 µm, green). The conceptual model predicted that both single and double pass light paths contribute to the steep decrease in ODR with decreasing diameter (Fig. 1B, black).

Conclusions : The widespread adoption of photographic retinal oximetry is currently limited by a lack of precision in measurements that produce non-physiological artifacts. In this analysis, the diameter artifact is manifested most strongly in venules and is predicted to depend on both double and single pass light transmission. Our data suggests improved understanding of the relationships between vessel saturation and diameter are needed to accurately characterize the error in photographic retinal oximetry.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.

 

Figure 1. A. Oximetry measurements of saturation in arterioles (red) and venules (blue) for multiple vessel diameters from ten oximetry images. Linear regression lines (black) are included. Green regression lines are given for D <100 µm and for D>100 µm. B. Optical density ratio in arterioles (red) and venules (blue); conceptual model results are shown in black.