Abstract
Purpose :
A device that accurately measures oxygen saturation in the retinal vasculature could be a strong tool to study and diagnose various diseases occurring in the retina. Although systems capable of measuring the retinal vessel oxygen saturation currently under clinical investigation, the need for increased accuracy of these systems clearly has to be addressed. An important factor that haemoglobin in blood is packed in discrete blood vessels and has different absorption than homogenous distribution (pigment packing effect) has to be incorporated into oximetry estimation.
Methods :
Haemoglobin is a strong absorber in visible wavelengths and confined distributions of densely packed haemoglobin significantly reduce the optical absorption path lengths for these wavelengths. We performed a detailed mathematical analysis assuming typical measurement parameters (signal-to-noise ratio) on modified oximetry equations that explicitly considers the haemoglobin packaging effect. These analyses enable choosing optimum wavelength combinations to estimate oxygen saturation in the retinal vessels with accuracies > 5% in the presence of 1% measurement noise. To validate the theory, we built a scanning laser ophthalmoscope (Fig.1A) and imaged a retina mimicking phantom (Fig.1C) in a model eye using 488, 548 and 612 nm wavelengths, which were found to be optimal for the absorbing dyes used in the phantom (Fig.1B). The scattering properties of the phantom were similar to human retina. Narrow channels with known dye concentrations were embedded in the phantoms to simulate the blood vessels.
Results :
The table shows the results from the average saturation (mean±standard deviation) estimated using our method for the various dye channels and corresponding true saturation values.
Conclusions :
The results obtained with the phantoms are consistent with our mathematical analyses and choice of wavelengths and gives a proof of principle that accurate measurement of oxygen concentration in the retinal vasculature for the whole saturation range can be achieved using our method. The SLO imaging has many advantages compared to conventional fundus camera based methods. A measurement is being planned to estimate the oxygen saturation in the retinal vessels for in vivo validation of our method.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.