Foveal Fine Structure in Retinopathy of Prematurity (ROP) Subjects: an Adaptive Optics Spectral Domain Optical Coherence Tomographic Study

Abstract

Purpose: To compare the structure of the fovea in subjects with a history of retinopathy of prematurity (ROP) to that of controls using AO-SDOCT retinal imaging.

Methods: A multimodal adaptive optics system (Hammer et al., JOSAA 2012) was used to study the fine structure of the fovea in four control and four ROP subjects. One eye from each subject was imaged (850 nm) through a dilated pupil (7.5 mm). Images were obtained over a 6 degree inferior-superior field containing the fovea. To characterize the gross structure of the pit, the vitreo-retinal boundary was objectively segmented using ImageJ and fitted with a Gaussian superimposed on a linear slope to obtain depth and standard deviation (breadth) parameters. Three regions of interest (ROI) were selected for study of retinal laminae: the foveal pit, the point at foveal half height, and the foveal rim. A custom segmentation routine using the first derivative of the intensity profile through each ROI was used to objectively measure the thickness of seven distinct retinal layers: 1) the retinal pigment epithelium and the choriocapillaries (RPE+CC), 2) the photoreceptor inner and outer segments (IS+OS), 3) the outer nuclear layer (ONL), 4) the outer plexiform layer (OPL), 5) the inner nuclear layer (INL), 6) the inner plexiform layer and ganglion cell layer (IPL+GCL), and 7) the nerve fiber layer (NFL).

Results: By Student's t-test, foveal pit depth was significantly shallower in ROP eyes. Foveal pit breadth did not differ significantly between groups but did trend narrower in ROP eyes. The postreceptor layers (NFL, GCL+IPL, INL, OPL) were unresolvable (and probably not present) in the foveal pit of controls, but were visible there in some ROP subjects. All seven layers were measurable at the foveal half-height and the foveal rim. Group×layer×eccentricity ANOVA revealed that the ROP retina was thicker than the control retina at all eccentricities and that, overall, postreceptor layers (NFL, GCL+IPL, INL, OPL) differed significantly between ROP and control but outer retinal layers (ONL, IS+OS, RPE+CC) did not.

Conclusions: There were significant differences in overall retinal thickness between ROP and control subjects that were explained mostly by changes to the inner retina.

Keywords: 706 retinopathy of prematurity • 688 retina • 552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)