Comparison of Oxygen Saturation in the Optic Nerve Head Using Five Wavelengths versus Hyperspectral Image Analysis

Abstract

Purpose:: To develop an algorithm to calculate relative oxygen saturation (O₂SAT) in the optic nerve head (ONH) and the O₂SAT response during an acute increase of intraocular pressure (IOP) using three oxygen insensitive and two oxygen sensitive wavelengths. To compare results with those of a continuous wavelength algorithm obtained from hyperspectral imaging (HSI).

Methods:: Five cynomolgus monkeys were anesthetized, and IOP was set with a saline reservoir connected to a needle and inserted into the anterior chamber. Reflectance light spectroscopy of ONH structures (artery, vein, nasal, temporal, cup, nasal cup, inferior, and superior areas) was performed by HSI. HSI contained 256 images (409 nm-917 nm, with a less than 2 nm resolution). Images were obtained using constant illumination while IOP was held at 10, 30, 45, and 55 mmHg for 10 minutes each. Five sessions were conducted for each eye in each animal. Blood pressure was measured periodically throughout the procedure. The five wavelength algorithm (5W) utilized three oxygen insensitive (548, 569 and 586nm) and two oxygen sensitive wavelengths (560 and 577nm) that were selected from HSI data sets. The result of O₂SAT obtained from the developed algorithm was compared to the result obtained from a continuous wavelength algorithm.

Results:: Relative O₂SAT of artery, vein, and ONH (nasal, temporal, cup, nasal cup, inferior, and superior areas) was calculated using 5W and the developed algorithm. The ratios of O₂SAT (HSI) / O₂SAT (5W) varied between 0.97 and 1.52 at IOP 10mmHg for different regions, with the exception of the cup, where ratios were between 0.107 and 0.157. At 30 mmHg, the ratios varied between 1.01 and 1.52 and at 55 mmHg, between 1.11 and 1.78. Linear regression coefficients for the relationship between O₂SAT (HSI) and O₂SAT (5W) over the range of IOP were: R² = 0.9309 (artery), 0.9055 (vein), 0.9952 (nasal), 0.852 (temporal), 0.2984 (cup), 0.9397 (nasal cup), 0.9889 (inferior), 0.9975 (superior).

Conclusions:: HSI recordings are not yet practical in humans, as periods of 10 to 30 seconds are required to scan the fundus. Our 5W multispectral method and developed algorithm can reduce the time required to collect imagery from human subjects. Our result suggests that 5W imagery is suitable to calculate O₂SAT in artery, vein, nasal, temporal, inferior, and superior, excepting the cup.

Keywords: optic nerve • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • hypoxia