Cone photoreceptors are central to vision and die in many retinal degenerative diseases. High-resolution retinal imaging methods–notably adaptive optics optical coherence tomography (AO-OCT)–use these cells’ reflectance profiles to characterize their morphologic and functional properties in the living human eye to assess their health. While some cone cells reveal reflections that correspond to identifiable features such as the inner segment/outer segment junction (IS/OS) and cone outer segment tip (COST), other cells can generate additional unexplained reflections that complicate our ability to characterize their reflectance profile. Here, we present a new quantitative method to properly identify cone reflections in AO-OCT images that correspond to their features. We use this method to estimate the prevalence of any additional cone reflections in healthy eyes and eyes with retinitis pigmentosa (RP) and to identify the true COST reflection. Using our method as a ground truth, we find that the conventional method (which identifies COST as the brightest reflection between IS/OS and retinal pigment epithelium) misidentified COST in 6.1±1.5% of cones in healthy controls. In the transition zones of RP, this rate can increase to 18.8%. In these cones, our method’s estimate of cone outer segment length and optoretinogram response differed by 36.8 ± 12.8% and 20.7 ± 17.6%, respectively, in healthy controls, and by 79.5 ± 21.2% and 34.9 ± 24.8% in the transition zone of RP.
|