Parafoveal cone function in choroideremia assessed with adaptive optics optoretinography

Abstract

Choroideremia (CHM) is an X-linked retinal degeneration leading to loss of the photoreceptors, retinal pigment epithelium (RPE), and choroid. Adaptive optics optoretinography is an emerging technique for noninvasive, objective assessment of photoreceptor function. Here, we investigate parafoveal cone function in CHM using adaptive optics optoretinography and compare with cone structure and clinical assessments of vision. Parafoveal cone mosaics of 10 CHM and four normal-sighted participants were imaged with an adaptive optics scanning light ophthalmoscope. While acquiring video sequences, a 2 s 550Δ10 nm, 450 nW/deg² stimulus was presented. Videos were registered and the intensity of each cone in each frame was extracted, normalized, standardized, and aggregated to generate the population optoretinogram (ORG) over time. A gamma-pdf was fit to the ORG and the peak was extracted as ORG amplitude. CHM ORG amplitudes were compared to normal and were correlated with bound cone density, ellipsoid zone to RPE/Bruch's membrane (EZ-to-RPE/BrM) distance, and foveal sensitivity using Pearson correlation analysis. ORG amplitude was significantly reduced in CHM compared to normal (0.22 ± 0.15 vs. 1.34 ± 0.31). In addition, CHM ORG amplitude was positively correlated with cone density, EZ-to-RPE/BrM distance, and foveal sensitivity. Our results demonstrate promise for using ORG as a biomarker of photoreceptor function.

Figure 1

AOSLO confocal image sequence processing and ORG signal extraction. (A) An exemplar confocal image of the common area averaged over all the video acquisitions from a single session for a normal-sighted participant. Cone centers were identified (blue dots). (B) Normalized and standardized intensity from every identified cone from one image sequence. (C) Standardized intensities were aggregated over all the identified cones by calculating the standard deviation of all traces shown in (B) at every time point. (D) Signals from multiple acquisitions acquired during a single session where each line comes from a unique acquisition processed in the same manner as (C). (E) The pooled standard deviation at every time point for both stimulated acquisitions (red) and control acquisitions (blue). (F) The control curve is subtracted from the stimulated curve, and the result is referred to as the ORG (black). A gamma-pdf is fit (green) to the ORG signal (black) and the peak of this fit is taken as the ORG amplitude (red asterisk).

Figure 2

ORGs from (A) normal-sighted participants and (B) CHM participants. Box and whisker plots (C) illustrating ORG amplitudes from CHM participants are significantly reduced in comparison to ORG amplitudes from normal-sighted participants, p < 0.001. The line inside the box represents the median ORG amplitude for each group. Lower and upper box boundaries represent the 25th and 75th percentiles respectively, and the lower and upper error lines show lowest and highest ORG amplitude respectively.

Figure 3

AO images of the cone mosaic and corresponding OCT images where ORGs were measured from the CHM participants with the three highest ORG amplitudes (top) and the three lowest ORG amplitudes (bottom). The number in the lower left corner of each AO image panel indicates the ORG amplitude for that particular participant. Yellow line under the OCT indicates the retinal location of the AO image.

Figure 4

ORG amplitude shows a correlation with cone density (A) and EZ-to-RPE/BrM distance (B) for CHM participants. The solid line is a linear fit to the CHM data; the dotted curves correspond to the 95% confidence interval of the fit. Pearson correlation coefficient = 0.69 for cone density and 0.72 for EZ-to-RPE/BrM distance. Circles represent data from CHM participants whereas triangles indicate data from normal-sighted participants. Gray triangles represent data from the same normal-sighted participants where only half of the cones were included in the ORG analysis.

Figure 5

Foveal sensitivity is correlated with ORG amplitude for CHM participants. The solid line is a linear fit and the dotted curves correspond to the 95% confidence interval of the fit. Pearson correlation coefficient = 0.77. The data from participant 13286 (gray circle) was determined to be an outlier by Q test and therefore was excluded from the fit.