We identified 58 consecutive patients who had rod-cone dystrophy and hyperautofluorescent rings on FAF retrospectively from our clinical database and categorized the rings according to disease progression. Three distinct stages of ring were evident; early, mid or late-stage disease ( Figure 2 A-C). An early stage ring is one that extends into the nasal retina and envelops the optic disc. The mid-stage ring is the classic shape seen when photoreceptor density shifts from rod-dominant to cone-dominant and can sometimes adopt a “bullseye” pattern. In late-stage disease, there are diffuse autofluorescence changes within the central cone-rich area. Here, there is significant disruption of the retinal anatomy and structural correlations are more challenging. No cases of unilateral or X-linked rod-cone dystrophy in female patients were included in the study.

(A) Early-stage large hyperautofluorescent ring involving the optic disc, (B) Mid-stage circumferential hyperautofluorescent ring around the fovea, and (C) Late- stage pattern of diffuse, irregular hyperautofluoresence in the central macula. (D, E) Patients excluded from the study as the horizontal or vertical diameters could not be measured due to missing portions of the hyperautofluorescent ring. Arrows demonstrate the poorly defined superior (D) and nasal (E) aspects. (F) Patients excluded due to poorly defined ring. (G) When a double ring was present, the internal ring was measured. (H, I) Vertical and horizontal measurements across the fovea of an elliptical and nonelliptical hyperautofluorescent ring.

For the cross-sectional analysis, we included patients with early-mid stage ring phenotypes in whom the ring borders were measurable (horizontal and/or vertical diameter across the fovea) and with analyzable OCT for EZ/ELM metrics. This yielded a cross-sectional cohort of 29 patients. For the longitudinal analysis, we then restricted the cross-sectional cohort to patients with ≥3 clinic visits with FAF suitable for registered repeat measurements. Twenty patients (mean age 42.2 years; range 11.0-76.0 years) met this requirement and were included in the longitudinal progression analysis, while 9/29 were excluded because they had fewer than 3 follow-up visits.

Baseline and subsequent hyperautofluorescent ring parameters were evaluated. Patients had undergone a full consultative ophthalmic examination at the Oxford Eye Hospital. Clinical data and genetic testing were performed as part of routine clinical care and data and results were collected retrospectively. For this analysis, patients with a clinical diagnosis of a rod-cone dystrophy confirmed on molecular genetic testing and who had images with a distinct macular hyperautofluorescent ring were included. Patients with ocular co-pathologies were excluded from the study.

Patients had undergone a standardized imaging protocol consisting of pseudo-colour fundus Optos images (Optomap P200; Optos plc, Dunfermline, UK) and spectral-domain optical coherence tomography (Spectralis, Heidelberg Engineering, Inc., Heidelberg, Germany). Fundus autofluorescence imaging was performed with a 30° and 55° field of view using a 488-nm excitation Spectralis confocal scanning laser ophthalmoscope (Heidelberg Engineering, Heidelberg, Germany) after pupil dilation with topical 0.5% tropicamide and 2.5% phenylephrine. Each image of a series was checked for optimal image quality. Images were excluded in cases of inhomogeneous illumination, sectorial opacities (e.g., eyelashes, floater), or unstable fixation. OCT follow-up images were co-registered to the baseline scan location using the device’s follow-up functionality. The right eye was used for analysis. If not applicable (for example, due to poor image quality or anatomical abnormalities), the left eye was used instead. The width of the ellipsoid zone (EZ) and external limiting membrane (ELM) were measured from spectral domain optical coherence tomography (OCT), Figure 1 . On the SD-OCT, the nasal and temporal edges of the EZ line were defined as the locations where the EZ line met the RPE. The width of the EZ line was defined as the distance between these 2 locations at the fovea. Patients were excluded from the study if the horizontal or vertical diameters could not be measured due to missing portions of the hyperautofluorescent ring ( Figure 2 D, E) or a very poorly defined ring ( Figure 2 F). When a double ring was present, the internal ring was measured. The FAF images were analyzed using measurements of horizontal hyperautofluorescent ring diameter across the fovea ( Figure 2 H). The vertical diameter was considered to be the line crossing the foveal center and perpendicular to the horizontal diameter even in cases where the ring was nonelliptical ( Figure 2 I). The measurement of the horizontal and vertical autofluorescence ring diameter were demarcated by the external delineable boundary. This was done manually. It should be noted that the ring thickness can be added to the inner ring diameter to get the outer ring diameter. In cases involving the nasal edge of the hyperautofluorescent falling outside the optic nerve, the diameter was still measured across the fovea and extended beyond the optic nerve. All measurements were done by 2 independent graders (MAE and SR).

The boundaries of the hyperautofluorescent ring were measured using the measuring tool included in the Spectralis Software. This was chosen as it is a simple and repeatable method. Measurements from follow-up visits were transferred onto the baseline image using the internal image registration function of HEYEX (Heidelberg, Germany), thereby overcoming discrepancies arising from small changes in focus, orientation, and edge distortion. Measurements of the horizontal hyperautofluorescent ring diameter were used, as opposed to the vertical axis, for 2 reasons. First, being the greater distance, measurement error and measurement variation between observers would be expected to have a smaller effect. Second, at matched eccentricity, it is observed that visual performance along the horizontal meridian outperforms the vertical visual meridian, suggesting that this meridian provides a more important role in human visual function and behaviour. The presence of polar angle asymmetries has been recognised in various visual tasks, including those involving contrast sensitivity, ^,, spatial resolution, ^, and tasks that engage higher visual areas, such as visual working memory. IRB/Ethics Committee ruled that approval was not required for this study.

Microperimetry was performed using the Macular Integrity Assessment (MAIA) confocal microperimeter (CenterVue, Padova, Italy) using the 10-2 grid and <=30% fixation losses reliability criteria.

Ethics approval was not needed for the retrospective audit of pseudoanoymised images; however, the gene therapy trial ( ClinicalTrials.gov identifier, NCT03116113) was previously approved with NightstaRx Ltd as the sponsor.

Statistical analysis was performed using R 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria) using the lme4 package (Bates, 2015). Normality was assessed using visual inspection of q-q plots and homogeneity of variances was assessed using a residual plot. All statistical tests were two-tailed with an alpha significance level set at 0.05. Pearson correlation coefficient was used as a measure of the strength of correlation of OCT parameters and ring radius.

AF ring total diameter data were plotted across time for all available datapoints. The proportion of variability in ring diameter attributable to time or baseline ring diameter was assessed using the coefficient of determination (R ² value). A linear mixed model, with the dependent variable of total horizontal hyperautofluorescent ring width and independent variables of time, genotype and the interactions was used.

Fifty-eight rod-cone dystrophy patients were analyzed in this study. None had syndromic disease. Analyzing hyperautofluorescent rings, 20 of these patients presented with early to mid-stage disease. The most common disease-causing genes in this cohort were USH2A (31.0%), RPGR (31.0%) and RHO (8.6%). Further details including patient molecular genetic testing can be found in supplementary Table 1 (available at https://www.aaojournal.org )

The 3 distinct parameters of the hyperautofluorescent ring on FAF were strongly correlated with the anatomical measurements on OCT ( P <.001) (Supplementary Table 2, available at https://www.aaojournal.org ). The internal ring diameter, external ring diameter and ring thickness, were strongly correlated with the ellipsoid zone width (r = 0.94), external limiting membrane width (r = 0.97) and loss of the ellipsoid zone (r = 0.92), respectively in a variety of rod-cone dystrophies. The inner border of the ring corresponded to the region where the EZ terminates, whereas the area outside the ring is characterized by the absence of both the EZ and the external limiting membrane ELM, resulting in the outer nuclear layer being directly in contact with the RPE. Consequently, both structure and function remained intact within the hyperautofluorescent ring.

Early and mid-stage rings correlated with inner and outer photoreceptor segment loss and got smaller in diameter as the degeneration progressed ( Figure 3 ). The results (R ² = 0.958) suggested that FAF changes declined very predictably in a logarithmic fashion ( Figure 4 ). Half-lives for RPGR, USH2A and RHO were 5.85, 19.75 and 23.42 years respectively ( P <.05).

Fundus autofluorescence and optical coherence tomography images monitoring progression over time. Dashed lines indicate the width of hyperautofluorescent ring. White lines indicate the width of the ellipsoid zone on OCT. With progression from early to mid-stage disease there is progressive constriction of the hyperautofluorescent ring and the ellipsoid line width.

Hyperautofluorescent ring diameter decay over time for different patients with different genetic mutations. Linear mixed models which included time and the factor in question, using an interaction between time and the potential factor as predictors. The ring diameter fits an exponential decay function (R2 = 0.995; P < 0.001). Age and sex had no statistically significant effect on the exponential decay function.

Stay updated, free articles. Join our Telegram channel

Join