Adaptive Optics Photoreceptor Imaging

In the present study, we investigated the cone packing density distribution along the horizontal meridian passing through the fovea in a population of young healthy subjects using a compact Adaptive Optics (AO) retinal camera (rtx1, Imagine Eyes, Orsay, France).

Nineteen healthy volunteer subjects (5 men and 14 women; age range, 24–38 years) participated in this study and gave a written informed consent. All subjects had 20/20 or better monocular best-corrected visual acuity and the spherical equivalent refractive errors ranged from −0.25 to −5.75 diopters (D) with astigmatism less than −1.50 D when referenced to the spectacle plane. The axial length (AxL) ranged between 22.61 and 26.29 mm. The protocol had approval of the local Ethical Committee and adhered to the tenets of Declaration of Helsinki. Exclusion criteria for this study included any ocular or systemic diseases.

Adaptive Optics imaging sessions were conducted after the pupils were dilated with 1 drop each of 0.5% tropicamide and 10% phenylephrine hydrochloride. A program provided by manufacturer correlated and averaged the captured image frames to reduce noise artefacts and produce a final image. Image analysis of the photoreceptor mosaic was performed using Image J (version 1.45a; NIH, Bethesda, MD). Cone density (cells/mm²) was estimated within two 50×50 μm windows at specified eccentricities (250-, 420-, 760- and 1300-μm) from the foveal center (Fig 1; available https://www.aaojournal.org). The spectacle-corrected magnification factor (RMF_corr) was determined in all the eyes.

With the exception of the central fovea (<160 μm), the photoreceptor structure was well resolved in most of the eyes. Cones were in close proximity to each other at 200 μm from the foveal center; at increasing retinal eccentricities, cones tended to become progressively larger and the intercellular space was wider between cells, with rods intruding between cones (Fig 2; available at https://www.aaojournal.org), in accordance with the histologic studies of the human retina.¹ A variation in brightness between adjacent areas of cones was seen in all the eyes. The mean cone density was 50 574±6031 cells/mm² at 250 μm eccentricity, falling to 14 198±2114 cells/mm² at 1300 μm eccentricity (analysis of variance [ANOVA]; P<0.05). In general, subjects with higher cone density close to the foveal center had higher cone density at increasing eccentricities. The intersubject variability in parafoveal cone density distribution, estimated by the coefficient of variation, was within 15%.

Adaptive Optics technology opens a new frontier for the research in clinical Ophthalmology. The accurate measurements of retinal microscopic sized features in normal populations, according to age, refractive defects, etc., represents the basis for detecting early pathological changes of the photoreceptor layer. The cone density found in the present study could be considered representative of a healthy population of myopic adults. In previous works using AO- scanning laser ophthalmoscope (SLO), Li et al² found an average decline in cone density from ∼120 000 to ∼45 000 cell/mm² from 0.10– to 0.30–mm eccentricity from the foveal center in a population of 18 adult young subjects (23–43 years; AxL 22.86–28.31 mm). Chui et al³ found an average cone density of ∼35 000 cell/mm² at 0.5 mm, ∼20 000 cell/mm² at 1.0 mm, and ∼12 000 cell/mm² at 1.5 mm eccentricity from the fovea respectively in 11 subjects (21–31 years; AxL: 22.00–28.00 mm). Song et al⁴ found a mean cone density of ∼70 000 cell/mm² at 0.18 mm from the fovea falling to 37 000 cell/mm² and 19 000 cell/mm² at 0.5- and 1.1-mm eccentricity respectively in a population of 10 young adults (22–35 years; AxL, 22.10–26.30 mm).

We based our method of cone counting on the results of a previous work by Hirsch and Miller.⁵ The authors demonstrated that a 56×56 μm was less subject to error than smaller window sizes when estimating cone density across increasing eccentricity from the fovea. Previous authors recently used a 50×50 μm sampling window to locate cone photoreceptor positions,⁴ further showing a high repeatability in cone density estimates taken 6 months apart at the same retinal location.

Data on populations of healthy eyes are fundamental in characterizing the density, distribution, and appearance of normal photoreceptor cells in vivo. This will permit measurement of the normal ranges, which allows comparison with pathological photoreceptors, even in early stages of retinal diseases.