Diameter dependence of dual-wavelength retinal oximetry

Abstract

Purpose : Retinal vessel oxygen saturation obtained by dual-wavelength oximetry shows a dependence on vessel diameter where measured saturation is increased in smaller vessels. This dependence is considered an artifact of measurement since hemodynamic considerations cannot explain it. This study aims at understanding the artifact as a consequence of vessel light transmission and reflectance from retinal tissue.

Methods : Retinal oximetry data were compared with outcomes from a model of tissue reflectance (Kubelka Munk, KM). KM defines the retinal vessel as a slab of infinite extent over the retina, and thus considers only double-pass transmission (transmission to and from retina behind vessel) and back-scatter by lumenal blood. KM predicts that vessel reflectance, and ultimately measured saturation, follows a nonlinear (hyperbolic) relationship with vessel diameter. Light intensity returned from first and higher-order vessels (I_v) and from adjacent retina was extracted (Oxymap Analyzer) from dual-wavelength oximetry images. Vessel reflectance (R_v) was defined by the relationship R_v = I_v / Io, where illumination intensity Io was determined such that for large vessel diameter (>150 mm), vessel reflectance matched the theoretical reflectance predicted by KM.

Results : At the oximetry measurement wavelength (600 nm), there was close agreement across the sampled diameter range (20-200 mm) between KM reflectance and experimental arteriolar and venular reflectance. At reference wavelength (570 nm), for diameters below 90 mm, the experimental reflectance from both vessel types exceeded by as much as 1.4x the model prediction (Fig. 1).

Conclusions : Agreement between oximetry data and KM supports the role of double-pass transmission as a source of diameter sensitivity. As diameter decreases, reflectance behind the vessel becomes important, increasing I_v, which ultimately raises measured saturation. Smaller diameter also allows single-pass transmission through the vessel which is a departure from the KM model. At low absorption (600 nm) the contribution is weak since there is nearly equal light return inside and outside vessels. At high absorption (570 nm), single-pass transmission raises vessel reflectance above the model prediction.

This is a 2020 ARVO Annual Meeting abstract.

 

Figure 1. Venule reflectance values (points) and KM model (dashed lines): 600 nm (upper) and 570 nm (lower).