Abstract
Purpose :
Using a geometrical model, foveal outer segment (OS) length can be used as a surrogate for foveal peak cone density in subjects with albinism.a However, the model did not hold for subjects with normal vision and it was suggested that challenges with resolving cones at the center of the fovea may have impacted the accuracy of peak cone density measurements. Here we used improved adaptive optics scanning light ophthalmoscopy (AOSLO) techniques to resolve the foveal cone mosaic and reexamine the relationship between peak cone density and OS length in subjects with normal vision.
Methods :
Fifty-eight subjects without vision-limiting pathology were recruited (35 female, 23 male; mean ± SD age = 31 ± 13 years). The foveal cone mosaic was imaged using confocal AOSLO using the following techniques to improve resolution: sub-airy disc pinhole, small field of view (down to 0.5 degrees), 680nm imaging wavelength, and through-focus imaging. AOSLO images were processed and montaged as previously described,b and peak cone density was measured using semi-automated analyses. For the OS length measures, 7mm horizontal transfoveal OCT B-scans were acquired using the Bioptigen OCT, with multiple B-scans being averaged to improve SNR. Maximum foveal OS length was measured and used to estimate peak cone density following previous methods.a
Results :
Images in both modalities with sufficient quality for analysis were obtained in 48 subjects. Peak OS length ranged from 37.56 to 47.21um (mean ± SD: 42.91 ± 2.09um); peak cone density ranged from 122,710 to 247,710 (mean ± SD: 181,924 ± 25,324 cones/mm2). No significant relationship was found between peak cone density and peak OS length (Spearman r = -0.011, p=0.93). The estimated cone density values from the OS-based model systematically underestimated the measured cone density values (mean bias = 49,141 cones/mm2), and the disagreement was worse in mosaics with higher cone density.
Conclusions :
The lack of agreement between OS-based peak cone density estimates and measured peak cone density values indicates that OS length cannot be used as a surrogate for peak cone density in individuals with normal vision. In our study, OS length was defined as the distance between the 2nd and 3rd hyperreflective bands on the OCT image, which may underestimate actual OS length. In addition, nonuniform OS volume and shape could be factored into the geometrical model in future studies.
aPMC5186335
bPMC3130574
This is a 2020 ARVO Annual Meeting abstract.