Abstract
Purpose :
The purpose of our study is to show how widefield optical coherence tomography (OCT) in combination with axial length measurements can be used to estimate the true retinal shape and curvature of an individual human eye.
Methods :
We have imaged eyes of normal volunteers using a prototype swept-source OCT system (ZEISS, Dublin, CA) with 200kHz scan rate, 6mm scan depth, and 21mm scan length. Axial length (AL) was measured using a commercial biometer (IOLMaster® 700 (ZEISS, Jena, Germany)). Refractive error (Rx) was measured using a commercial autorefractor (HARK Auto Refractor Model 599 (ZEISS, Dublin, CA)).
We are correcting the OCT scan using an estimation model consisting of the optical design of the OCT imaging system and a nominal model of the human eye adjusted for measured axial length. The estimation model creates corrected images that show the true retinal shape and fits the automatically-segmented retinal pigment epithelium layer to estimate the equivalent radius of curvature of the retina. The estimation model has been validated using a phantom with known geometry.
Results :
A total of 10 eyes of 10 subjects were enrolled in the study. The eyes were measured to have refractive errors between -8.5 and +2.0 diopters and axial lengths between 22.78 and 27.10 mm. We were able to acquire OCT images of all eyes. We were further able to estimate the retinal shape of all eyes. Figure 1 shows OCT B-scans before and after correction; (a) a short, hyperopic eye (Rx = +1.25D, AL = 22.78mm) with a retinal radius of curvature of 13.19mm and (b) a long, myopic eye (Rx = -8.5D, AL = 26.88mm) with a retinal radius of curvature of 11.26mm. Figure 2 shows the corrected OCT image to scale compared to the eye length suggesting that the shorter eye is likely oblate and the longer eye likely prolate.
Conclusions :
We have demonstrated an estimation model that can determine the retinal shape and curvature of an individual human eye based on widefield OCT images and axial length measurements. The estimation model provides geometrically correct retinal shapes and an estimation of the retinal curvature.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.