Extrafoveal Cone Packing Density and Geometry in Retinopathy of Prematurity (ROP)

Abstract

Purpose: To study cone packing density and geometry using an adaptive optics scanning laser ophthalmoscope (AOSLO) in eyes with history of ROP and age matched control subjects.

Methods: Subjects with a history of severe, treated ROP (n=3, SROP), with a history of mild ROP that resolved without treatment (n=3, MROP), and term born controls (n=6, CT) were studied. The subjects were aged 15-23 (median: 19.8) years at test. Two to four confocal and offset-pinhole AOSLO videos (64 frames) subtending 1°×1° visual angle were obtained from each subject at retinal eccentricities 4.5°, 9°, 13.5° and 18° (Chui et al, Biomed Opt Exp, 2012) by directing the subjects’ gaze using a fixation target. Offline, non-rigid registration was used to obtain a final image for analysis. Cones were counted using a custom MATLAB program and cone geometry was assessed by fit of Voronoi polynomials. Cone density (cones/mm²) was estimated after correction of the image magnification based on axial lengths (ROP = 22.5 ± 1.11 mm vs. CT = 23.6 ± 1.10 mm). Effects of group and eccentricity were detected by repeated-measures ANOVA.

Results: Cone density decreased with increasing retinal eccentricity (4.5° to 13.5°) in all groups. Cone density was markedly lower in SROP than in CT and MROP eyes (p<0.01); however, image quality was also lower in SROP subjects. The Voronoi polygons were significantly more irregular in SROP than in MROP and CT eyes (p< 0.05).

Conclusions: SROP (but not MROP) subjects displayed reduced number and organization of cone photoreceptors. This may indicate loss of cone photoreceptors, consistent with recognized visual deficits in SROP subjects. Alternatively, low cone counts maybe due to the relatively poor quality of images obtained from SROP eyes.