Calibration Of Retinal Oximetry Devices By Fundus-simulating Phantoms

Retinal blood vessel oxygen supply reflects retinal metabolism through the index of blood oxygen saturation（SO2）.A large number of studies have shown that retinal vascu-lar S0₂ is abnormal in many common diseases such as diabetic retinopathy（DR）,glau-coma and retinal occlusions,with statistical differences from the measurements healthy samples.Therefore,the accurate measurement of retinal vessel S0₂ has great signif-icance for the diagnosis of related diseases.Commonly used retinal oximetry device adopts the method of dual-wavelength fundus imaging,and it calculates SO2 through the linear relationship between S0₂ and the optical density ratio（ODR）of blood vessels under an oxygen-sensitive wavelength and an oxygen non-sensitive wavelength.However,the lack of traceable calibration methods has become a key problem restricting the clinical application of such devices.To establish a traceable calibration standard for retinal SO₂ measurement,we developed a fundus-simulating phantom.The phantom obtains a precise biomimetic microchannel network structure through lithography processing,simulates the optical parameters of the human fundus tissue through the ratio adjustment of the absorption and the scattering agent,and controls the S0₂ by perfusion of red blood cell solution.Using this phantom,we calibrated a dual-wavelength retinal oximetry device,studied the linear relationship between SO₂ and ODR,and analyzed the influence of factors such as fundus pigmentation and retinal blood vessel diameter on SO₂ calculation,thus deduced an S0₂ algorithm based on fundus-simulating phantom calibration.The algorithm was used to calculate retinal vessel S0₂ in five volunteers（aged 24-27 years）in the validation group,compared with the empirical value calibration method.The results of repeated experiments showed that the mean S0₂ of the arteries and veins obtained by the fundus-simulating phantom calibration algorithm were consistent with the calculation results of the empirical value calibration algorithm,and had similar reproducibility.This study demonstrates the potential of clinical applications for the calibration of retinal oximetry devices using fundus-simulating phantoms.