Microperimetry Features of Geographic Atrophy Identified With En Face Optical Coherence Tomography

Importance  Progressive geographic atrophy (GA) of the retinal pigment epithelium leads to loss of central vision. To identify GA in age-related macular degeneration and assess treatment, correlation of function observed on microperimetry with structure observed on optical coherence tomographic (OCT) images may be of value.

Objective  To characterize the microperimetric function of GA as identified from en face OCT imaging.

Design, Setting, and Participants  In a case-series study, 20 patients (22 eyes) entered the study at the University of Padova according to preplanned conditions. From March 1 to July 30, 2014, en face OCT images were obtained at the outer retinal layer and choroidal layer levels. The microperimetry sensitivity map was superimposed on the en face OCT images, which had been used to measure GA areas. Relative and dense scotoma rates were calculated in the GA areas. After data collection, the study eyes were divided into 3 groups according to the macular residual mean sensitivity.

Main Outcomes and Measures  Retinal sensitivity measured by microperimetry within areas of GA identified by en face OCT images.

Results  Twenty patients (5 men and 15 women) were included in the study, with a mean (SD) age of 79.5 (7.0) years (range, 69-98 years). Macular residual mean retinal sensitivity was less than 5 dB in 7 eyes (group 1), 5 to 10 dB in 9 eyes (group 2), and greater than 10 dB in 6 eyes (group 3). Mean (SD) GA area differed among the groups at the outer retinal (13.13 [5.03] mm² [range, 5.75-21.04 mm²] in group 1; 7.80 [3.25] mm² [range, 3.31-13.52 mm²] in group 2; and 3.94 [2.35] mm² [range, 1.46-7.90 mm²] in group 3; P = .001) and choroidal (11.83 [5.55] mm² [range, 4.55-22.14 mm²] in group 1; 7.00 [4.29] mm² [range, 0.90-13.83 mm²] in group 2; and 3.27 [2.29] mm² [range, 0.91-7.23 mm²] in group 3; P = .007) layer levels. Mean (SD) GA area imaged at the outer retinal layer level was significantly larger than that imaged at the choroidal level in group 3 (difference, 0.67 mm²; 95% CI, 0.31-1.03 mm²; P = .005), but not in groups 1 or 2. Mean (SD) rate of relative scotoma was significantly higher in the GA area imaged at the outer retinal layer level than at the choroidal level in group 3 (47.70% [31.30%] [range, 13.60%-100%] vs 34.00% [37.30%] [range, 0%-100%]; difference, 13.74%; 95% CI, 3.84%-23.63%; P = .02), but not in groups 1 or 2.

Conclusions and Relevance  In the early stage of GA, when retinal sensitivity is relatively good, these data suggest that the GA area imaged on en face OCT at the outer retinal level correctly detects the wide functional degenerative involvement of the photoreceptors. These findings provide novel data that correlate function and structure, which may be of value when assessing treatments that might prevent or reduce the rate of growth of GA.

Geographic atrophy (GA) represents the atrophic late-stage manifestation of dry age-related macular degeneration (AMD).¹ Geographic atrophy is characterized by the loss of photoreceptors, retinal pigment epithelium (RPE), and choriocapillaris within the macula. The areas of GA slowly progress over time.² Therefore, GA is a significant cause of moderate to severe central visual loss. Geographic atrophy is usually assessed with color fundus photography or, more recently, fundus autofluorescence (FAF), using blue (B-FAF) and near-infrared (NIR-FAF) types.^3-5 Since 2008, several studies^6-9 have shown that GA can be identified (with high reproducibility) and quantified (with high repeatability) by means of en face optical coherence tomography (OCT). As a complement to conventional B-scan OCT, en face OCT is a new approach for fundus imaging. En face OCT provides a coronal full macular view at different depth levels and additional anatomical insights about macular diseases.^10,11 En face OCT images vary according to the different position of the coronal scan in the retina or in the choroid. The increased OCT choroidal signal associated with GA results in the visualization of GA areas on the choroidal side, whereas the loss of the OCT signal in the outer retina allows visualization of GA on the retinal side.⁷ Using enhanced-depth imaging spectral domain OCT, which allows a better visualization of the choroid than standard OCT, the GA area visualized and measured on en face OCT is comparable to the GA area quantified on B-FAF and NIR-FAF.^6,12-15 However, on en face OCT, the GA area is larger when visualized at the outer retinal layer level than at the choroidal layer level.⁶

The correlation of function from microperimetry with structure from OCT images in GA of the RPE from eyes with AMD could be of value when assessing treatment for this condition. The purpose of this study was to characterize the microperimetric function of GA as identified from en face OCT imaging.

Box Section Ref ID

From March 1 to July 30, 2014, all patients with GA attending our clinic at the Department of Ophthamology, University of Padova, Padova, Italy, were enrolled in this prospective, cross-sectional study according to preplanned conditions. Patients were older than 55 years. The inclusion criteria consisted of the presence of unifocal or multifocal GA secondary to AMD in at least 1 eye at fundus examination. Exclusion criteria consisted of significant media opacities, clinical evidence of choroidal neovascularization, any evidence of diabetic retinopathy, myopia greater than 6 diopters (D), glaucoma, previous macular laser treatment, signs or history of hereditary retinal dystrophy, and GA confluent into the peripapillary atrophy. Patients were also excluded if GA extended outside the central OCT scanning area (a square centered on the fovea of 6 × 6 mm). Fluorescein angiography was performed only if presumed signs of neovascular AMD (extracellular fluid, hemorrages, exudates, or fibrosis) were found, and confirmation of neovascular AMD excluded the eye from the study. The ophthalmologic examination consisted of refraction, determination of best-corrected visual acuity, anterior segment examination, 90-D lens biomicroscopy, OCT, and microperimetry. All imaging modalities were performed with dilated pupils. This study was conducted in accordance with the tenets of the Declaration of Helsinki¹⁶ and with the approval of the institutional ethical committee of University of Padova. After a detailed explanation of the purpose of this study, all enrolled patients provided written informed consent.

En face OCT was performed using a spectral domain device (Spectralis imaging system; Heidelberg Engineering) as previously described.⁶ The device uses a superluminescence diode-emitting scan beam at a wavelength of 870 nm. The retina is scanned at a speed of 40 000 A-scans/s with an axial resolution of 3.9 μm and a transversal resolution of 14 µm. The eye tracking and automated real time of the imaging system were used throughout the study. The eye tracker enables each OCT scan to be registered and locked to a reference image. The automated real time was set at 50 frames.

For the purpose of the study, 20° × 20° (5.90 × 5.90 mm) volume scans centered on the fovea were acquired using the enhanced-depth imaging scan modality. Ninety-seven horizontal scans, 60 μm apart, were obtained. Using the automatic retinal segmentation of the device to define the depth level, en face OCT images were generated for the outer retinal and the choroidal levels. The outer retinal en face image was positioned from a segmentation reference line passing through the outer retinal layers, located 25 μm from the Bruch membrane. The choroidal en face image was positioned from a segmentation line parallel to the previous one, located in the choroid 50 μm from the Bruch membrane. The automatic segmentation line at the level of the Bruch membrane was used as the reference line. This line was manually reviewed if the automatic algorithm failed.

On the outer retinal and choroidal en face images obtained for every eye included in the study, the GA areas were manually outlined and measured using the image analysis software of the instrument (Heidelberg Eye Explorer; Heidelberg Engineering GmbH). The reliability and reproducibility of this measurement has been described previously.⁶

Microperimetry (MP1) (Nidek) was performed in the mesopic condition under mydriasis. For the purpose of this study, the following variables were used: a red ring fixation target 2° in diameter; a white, monochromatic background set at 1.27 cd/m² (4 apostilbs); stimulus size Goldman III, with a 200-millisecond projection time; and a 4-2 double staircase strategy. A personal customized grid covering the central 16° centered on the fovea and including all the areas of GA was planned for each eye.

At the end of the recruitment, using the dedicated manufacturer’s software each microperimetry map was superimposed on the outer retinal layer and choroidal layer level en face OCT images where the GA areas were previously identified. Therefore, for each single study eye, 2 different maps were obtained (Figure). Inside these maps, macular residual mean retinal sensitivity (which excludes dense scotoma) and the rate of relative and dense scotoma were calculated. Dense scotomas were defined as tested loci that elicited no response, even at the highest light intensity stimulus (0 dB).¹⁷

For the statistical analysis, after all the data were collected, study eyes were divided into the following 3 groups of macular residual mean retinal sensitivity levels: less than 5 dB (group 1); 5 to 10 dB (group 2); and greater than 10 dB (group 3). All the evaluated variables have been summarized according to the usual methods of the descriptive statistics (mean [SD] and range). Statistical analysis was performed to verify whether the chosen variables (GA area, residual mean retinal sensitivity, and rate of relative and dense scotoma) showed significant differences between the areas identified at the outer retinal and choroidal layer levels and among the 3 sensitivity level groups. In addition, further investigation has been devoted to verify whether differences between the areas varied their range in relation to the sensitivity level (interaction effect between area and sensitivity level). For this reason, the following 3 analysis of variance (ANOVA) models have been applied: (1) the 1-way ANOVA model within each area (at the outer retinal layer and choroidal layer levels), with sensitivity level as the between-group factor, with the Bonferroni post hoc test for multiple comparisons applied in case of rejection of the null hypothesis; (2) the 1-way ANOVA within each sensitivity-level group, with area as the within-patient factor; and (3) the 2-way ANOVA model with interaction, with area as the within-patient factor and sensitivity level as the between-group factor. The data were analyzed using the PROC MIXED procedure of SAS (version 9.2; SAS Institute Inc).

Twenty consecutive patients (22 eyes; 5 men and 15 women) with GA were included in the study. Mean (SD) age was 79.5 (7.0) years (range, 69-98 years). Macular residual mean sensitivity was less than 5 dB in 7 eyes (group 1), 5 to 10 dB in 9 eyes (group 2), and greater than 10 dB in 6 eyes (group 3). Mean age did not differ among the groups.

Mean GA area at the outer retinal layer level was significantly different among the groups (13.13 [5.03] mm² [range, 5.75-21.04 mm²] in group 1; 7.80 [3.25] mm² [range, 3.31-13.52 mm²] in group 2; and 3.94 [2.35] mm² [range, 1.46-7.90 mm²] in group 3; P = .001) (Table 1). The mean difference was 5.33 (95% CI, 0.56-10.11) mm² between groups 1 and 2 (P = .03), 9.19 (95% CI, 3.92-14.46) mm² between groups 1 and 3 (P < .01), and 3.86 (95% CI, −1.13-8.85) mm² between groups 2 and 3 (P = .15) (Table 2). Mean GA area at the choroidal level was also different among the groups (11.83 [5.55] mm² [range, 4.55-22.14 mm²] in group 1; 7.00 [4.29] mm² [range, 0.90-13.83 mm²] in group 2; and 3.27 [2.29] mm² [range, 0.91-7.23 mm²] in group 3; P = .007) (Table 1). The mean difference was 4.83 (95% CI, −0.73 to 10.39) mm² between groups 1 and 2 (P = .11), 8.56 (95% CI, 2.43-14.70) mm² between groups 1 and 3 (P = .01), and 3.37 (95% CI, −2.09 to 9.54) mm² between groups 2 and 3 (P = .27) (Table 2). Mean GA area was larger at the outer retinal level than at the choroidal level in group 3 (difference, 0.67 [95% CI, 0.31-1.03] mm²; P = .005). A significant difference was not identified between layer levels in groups 1 or 2 (P = .21 and P = .34, respectively) (Table 1).

The mean rate of relative scotoma in the GA area was not different among the 3 groups for imaging at the outer retinal layer level or the choroidal layer level (P = .06 and P = .12, respectively). The mean differences (95% CI) between the groups are summarized in Table 2. In group 3, the mean rate of relative scotoma was higher when quantified at the outer retinal level than at the choroidal level (47.70% [31.30%] [13.6%-100%] vs 34.00% [37.30%] [0%-100%]; mean difference, 13.74% [95% CI, 3.84%-23.63%]; P = .02); not in groups 1 and 2 (P = .38 and P = .63, respectively) (Table 1).

Residual mean sensitivity in the GA area imaged at the outer retinal level was significantly different among the 3 groups (1.17 [1.10] dB [range, 0.00-3.00 dB] in group 1; 3.06 [2.99] dB [range, 0.00-3.00 dB] in group 2; and 8.62 [2.87] dB [range, 6.30-13.00 dB] in group 3; P < .001) (Table 1). The mean difference was −1.88 (95% CI, −5.10 to −1.34) dB between groups 1 and 2 (P = .32); −7.45 (95% CI, −11.00 to −3.89) dB between groups 1 and 3 (P < .01); and −5.56 (95% CI, −8.93 to −2.20) dB between groups 2 and 3 (P < .01) (Table 2). Residual mean sensitivity in the GA area imaged at the choroidal level was also different among the groups (1.00 [1.19] dB [range, 0.00-3.20 dB] in group 1; 3.96 [2.02] dB [range, 0.80-7.10 dB] in group 2; and 7.82 [3.85] dB [range, 0.00-14.00 dB] in group 3; P = .02) (Table 1). The mean differences were −2.96 (95% CI, −6.68 to 0.77) dB between groups 1 and 2 (P = .14); −6.82 (95% CI, −10.93 to −2.70) dB between groups 1 and 3 (P < .01); and −3.86 (95% CI, −7.76 to 0.04) dB between groups 2 and 3 (P = .05) (Table 2).

As a complement to conventional B-scan OCT, en face OCT is a new imaging modality that allows a full-extent 2-D view of the posterior pole, a topographical analysis, and a close comparison with other 2-D imaging modalities. Using this imaging modality, the GA area may be easily visualized and quantified, by means of different OCT devices, and the growth of the GA can be evaluated.^6,8,9,18 However, the en face OCT images differ according to the different position of the coronal scan in the retinal or choroidal layer. A good correlation has been found between en face OCT and FAF at the outer retinal level, but not at the choroidal level.⁶ Moreover, the mean GA area has been shown to be larger at the outer retinal level than at the choroidal level.⁶ In the present study, we measured the GA area on en face OCT in 3 different groups according to the residual mean retinal sensitivity of the whole macula. We found that the mean GA area was still larger at the outer retinal level than at the choroidal level, but only in group 3, which is characterized by better macular residual sensitivity (>10 dB). In this group, the GA area was smaller on both en face images compared with groups 1 and 2.

Sayegh et al¹⁹ showed that the larger morphologic defect in the GA area, graded in the analysis of B-scan spectral domain–OCT, was in the photoreceptor layer and not within the RPE layer. The signal enhancement in the choroid, using conventional B-scan OCT, was larger than the area of complete RPE alteration, suggesting that complete RPE loss may not be the only variable responsible for this characteristic OCT feature.¹⁹ These results agree with our findings. Therefore, the outer retinal layer loss or disruption, easily detectable using en face OCT, may precede the complete RPE loss.

In the present study, we also functionally characterized GA areas using microperimetry. We found that GA areas, identified using en face OCT, differ not only for extension, but also for retinal sensitivity. In group 3 (eyes with better residual sensitivity), the mean rate of relative or dense scotoma was different between GA visualized with the 2 en face images. In the choroidal image, 66.0% and 34.0% of the stimulated points were defined as dense and relative scotoma, respectively, whereas in the outer retinal layer image, the rates of dense and relative scotoma were 52.3% and 47.7%, respectively. These findings suggest that the GA area visualized at the choroidal layer level better represents the worst functioning retinal layer zones (dense scotomas), whereas the larger GA area visualized at the outer retinal level corresponds to the wider dysfunctioning retina. Sayegh et al¹⁹ found that 38% of the light stimulation points were perceived in an area of presumed atrophy, mainly along the borders of the GA. Moreover, the early functional impairment of photoreceptors at the GA boundaries has been recently detected by microperimetry, also in areas of abnormal NIR-FAF findings.^14,20-24 The sequence of events in GA has different pathophysiologic hypotheses.^25-30 Our findings seem to confirm the hypothesis that in GA, photoreceptor impairment and RPE changes, better visualized by en face OCT imaging at the outer retinal layer level, occur before choriocapillaris loss.^31,32

The small sample size is a limitation of this study. Given the likely broad biological variability of GA and the strict enrolment criteria, the number of the studied eyes was limited. A follow-up study with a wider sample size may be useful to confirm (or refute) whether these findings have additional predictive value in the growth of GA toward legal blindness.

In the early stages of GA, areas identified by en face OCT at the outer retinal level correctly mirror the functional degenerative involvement of the photoreceptors, which suggests an earlier degenerative involvement of the outer retina and RPE than the underlying and longer surviving choriocapillaris. En face OCT and microperimetry may be beneficial in GA detection, in addition to conventional FAF, which simply shows RPE changes. En face OCT and microperimetry may identify the morphologic and functional aspects of GA. These findings provide novel data that correlate function through microperimetry and structure through OCT imaging, which may be of value when assessing treatments that might prevent or reduce the rate of growth of GA.

Corresponding Author: Edoardo Midena, MD, PhD, Department of Ophthalmology, University of Padova, via Giustiniani 2, 35128 Padova, Italy (edoardo.midena@unipd.it).

Submitted for Publication: January 21, 2016; final revision received April 6, 2016; accepted April 10, 2016.

Published Online: June 2, 2016. doi:10.1001/jamaophthalmol.2016.1535.

Author Contributions: Drs Midena and Pilotto had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Pilotto, Midena.

Acquisition, analysis, or interpretation of data: Pilotto, Convento, Guidolin, Abalsamo, Longhin, Parrozzani.

Drafting of the manuscript: Pilotto, Guidolin, Abalsamo, Longhin, Parrozzani.

Critical revision of the manuscript for important intellectual content: Pilotto, Convento, Midena.

Statistical analysis: Pilotto.

Administrative, technical, or material support: Pilotto, Convento, Guidolin, Abalsamo, Parrozzani.

Study supervision: Pilotto, Guidolin, Parrozzani, Midena.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Funding/Support: This study was supported by Fondazione Roma and the Ministry of Health (G. B. Bietti Foundation).

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.