Geographic Atrophy in AMD: Prognostic Factors Based on Long-Term Follow-Up

Abstract

Introduction

The aim of this large-scale long-term retrospective study was to show the enlargement rate (ER) of geographic atrophy (GA) in age-related macular degeneration (AMD), defined as complete retinal pigment epithelium and outer retinal atrophy (cRORA), to find predictors of progression in a clinical routine setting and to compare GA evaluation methods.

Methods

All patients available in our database with follow-up of at least 24 months and cRORA in at least one eye, regardless of neovascular AMD being present, were included. SD-OCT and fundus autofluorescence (FAF) evaluations were performed according to a standardized protocol. The cRORA area ER, the cRORA square root area ER, the FAF GA area, and the condition of the outer retina (inner-/outer-segment [IS/OS] line and external limiting membrane [ELM] disruption scores) were determined.

Results

204 eyes of 129 patients were included. Mean follow-up time was 4.2 ± 2.2 (range 2–10) years. 109 of 204 (53.4%) eyes were classified as MNV-associated GA in AMD (initially or during follow-up); 95 of 204 (46.6%) eyes were classified as pure GA in AMD. The primary lesion was unifocal in 146 (72%) eyes and multifocal in 58 (28%) eyes. A strong correlation was observed between the area of cRORA (SD-OCT) and the FAF GA area (r = 0.924; p < 0.001). Mean ER was 1.44 ± 1.2 mm²/year, mean square root ER 0.29 ± 0.19 mm/year. There was no significant difference in mean ER between eyes without (pure GA) and with intravitreal anti-VEGF injections (MNV-associated GA) (0.30 ± 0.19 mm/year vs. 0.28 ± 0.20 mm/year; p = 0.466). Eyes with multifocal atrophy pattern at baseline had a significantly higher mean ER compared to eyes with unifocal pattern (0.34 ± 0.19 mm/year vs. 0.27 ± 1.19 mm/year; p = 0.008). There were moderate significant correlations between ELM and IS/OS disruption scores and visual acuity at baseline, 5 and 7 years (all r values ca. −0.5; p < 0.001). In multivariate regression analysis, a multifocal cRORA pattern at baseline (p = 0.022) and a smaller baseline lesion size (p = 0.036) were associated with a higher mean ER.

Conclusion

SD-OCT-evaluated cRORA area might serve as a GA parameter comparable to traditional FAF measurement in clinical routine. The dispersion pattern and baseline lesion size might be predictors of ER, whereas anti-VEGF treatment seems not to be associated with ER.

Introduction

Age-related macular degeneration (AMD) might show clinically atrophic and neovascular aspects. The advanced dry form presents itself as geographic atrophy (GA) or, according to the international group of experts in the Classification of Atrophy Meetings (CAM) definition, in its most advanced form as complete RPE and outer retinal atrophy (cRORA) [1]. The neovascular form is characterized by the formation of macular neovascularization (MNV) but might also show GA. Intravitreal injections of the anti-vascular endothelial growth factor (anti-VEGF) substances ranibizumab, aflibercept, faricimab, and brolucizumab are now established as the standard treatment for neovascular AMD. This standard is based on the results of the phase III studies MARINA/ANCHOR, VIEW1/2, and HAWK/HARRIER [2–6]. While the neovascular aspect of AMD is treated with intravitreal anti-VEGF injections, newer drugs like pegcetacoplan [7] and avacincaptad pegol [8] showed a reduction in lesion growth with pegcetacoplan being approved as therapy for GA by the FDA very recently. Despite the known strong genetic component [9, 10] and dysfunction of the complement system [11, 12] in the pathophysiology of GA, studies with complement factor inhibitors such as lampalizumab and eculizumab have not yet been able to slow down the progression of GA [13, 14].

Beside subretinal fibrosis, development and progression of GA are strong visual outcome influencing factors in both pure GA [15] and MNV-associated GA in AMD [16–18]. So far, the natural course of GA progression has primarily been investigated with the help of color fundus photography and fundus autofluorescence (FAF) imaging [15, 19]. Recent advances in spectral-domain optical coherence tomography (SD-OCT) technology now provide the opportunity to obtain the earliest diagnosis of GA and to follow a progression as well [1]. Other than color fundus or FAF imaging, SD-OCT imaging in a follow-up mode allows for a constant positioning of measurements.

The aim of our study is to show the enlargement rate (ER) of GA, defined as cRORA, in SD-OCT in a clinical routine population. Furthermore, the area of the cRORA in SD-OCT is to be correlated with the area in the FAF to evaluate the usability of cRORA in SD-OCT instead of FAF GA area for clinical practice. Recent studies have already shown high correlations between GA area in SD-OCT and FAF [20, 21]. However, both studies did not include cRORAs associated with MNV-associated GA in AMD. We intentionally included cRORA associated with MNV since neovascularization occurs at significant rates in the clinical setting. The ER will be correlated with morphological and functional parameters to determine reliable predictors of the disease progression in order to provide patients with an individual prognosis regarding the expected course.

Methods

This was a retrospective single-center study between 2009 and 2021 at the Vista Augenklinik Binningen, Switzerland. Approval to conduct this study was obtained from the local Ethics Committee (Ethics Committee Northwestern Switzerland (EKNZ Nr. BASEC 2021-01661) and the study was carried out in accordance with the ICH-GCP guidelines and the principles of the Declaration of Helsinki. Most patients have signed the general consent of the Vista Augenklinik Binningen. If a patient did not sign this general consent so far, the data were not used or the written general consent was obtained retrospectively. In the case of patients who have already died, the data were used without being asked because the effort involved in obtaining the general consent would have been disproportionate. The need for informed consent of the patients who have deceased was waived by the EKNZ (EKNZ Nr. BASEC 2021-01661).

Inclusion and Exclusion Criteria

All patients available in our database were included in the analysis in whom cRORA existed in at least one eye, regardless of neovascular AMD being present initially or in the course of the disease. The follow-up documentation of OCT over at least 24 months as well as a complete and entirely traceable documentation of all treatment periods to be evaluated was assumed. In the case of neovascular AMD diagnosed by fluorescence angiography, comprehensible documentation of the type and frequency of intravitreal anti-VEGF therapy was a prerequisite. All data up to October 2021 were included in the analysis. The evaluation of the data was carried out from September 2021 to November 2021. Exclusion criteria were insufficient imaging documentation and insufficient observation period.

Technical Image Evaluation and Measurement of the cRORA

SD-OCT scans were acquired using an established protocol consisting of volume scans. For the volume scan of 20° × 15°, 19 frames (high-speed mode, 9 frames, 512 A-scans) were acquired in a follow-up setting. Further, a 6-mm star scan (high-speed mode, 9 frames, 512 A-scans) was performed. Quantitative and qualitative evaluations were performed by one rater (LC) according to a standardized protocol.

cRORA Area in SD-OCT and FAF

The cRORA area was recorded using the Heyex software. For this purpose, the area considered as cRORA was circumnavigated with the tool “Draw Region” in the SD-OCT IR image manually with the mouse cursor, while simultaneously comparing the borders of the cRORA area using the position marker at the referring scan position in the B-scan image (Fig. 1). The cRORA area was then calculated by the Heyex software. The measurements were carried out at annual follow-up points. The recorded areas were based on the OCT criteria for cRORA as proposed by the CAM: (1) a hypertransmissive region with a diameter of at least 250 µm, (2) a zone of attenuation or disruption of the RPE with a diameter of at least 250 µm, (3) evidence of an overlying photoreceptor degeneration, and (4) absence of RPE scrolling or other signs of RPE crack [1]. The evaluation also took into account a grading of the dispersion pattern with regard to the uni- or multifocality of the primary lesions in the IR image. To calculate the area of ​​atrophy in a multifocal cRORA, the areas circled in yellow were added together (Fig. 1b).

Fig. 1.

a The area considered as cRORA was circumnavigated with the tool “Draw Region” in the SD-OCT IR image with the help of the position marker in the referring scan position in the B image. b To calculate the area of atrophy in a multifocal cRORA, the areas circled in yellow were added together.

In the FAF, clearly circumscribed hypofluorescent areas were rated as cRORA and manually circumnavigated with the mouse cursor while the area was then calculated by the Heyex software (Fig. 2). A square root transformation was performed as proposed by Feuer et al. [22] in order to eliminate the dependence of the ER on the baseline lesion size.

Fig. 2.

The area considered as cRORA was circumnavigated with the tool “Draw Region” in the FAF image.

ELM, IS/OS, PE Scores

Photoreceptor inner segments and outer segments (ISs/OSs) impairment and external limiting membrane (ELM) impairment, as well as pigment epithelium impairment, were evaluated in the macular star scan by grading the disruption in the central 1 mm between values from 0 (intact) to 3 (severe) as first described by Hatz et al. [18] and recently pictured by Hoffmann and Hatz [23, 24] (Fig. 3ab).

Fig. 3.

a, b This example illustrates severely impaired ELM, IS/OS and PE (>3/4 disruption within 1-mm center) in the macular star scan.

Statistical Analysis

Statistical analyses were performed with SPSS version 21 (SPSS; Chicago, IL, USA). Descriptive data were presented as mean ± standard deviation or percentages. To compare baseline and follow-up values and different subgroups, paired and unpaired t tests were used. Pearson correlation coefficients were calculated for correlation analyses. To remove the dependence of cRORA growth rates on baseline cRORA area, a square root transformation strategy was used for cRORA area measurements [22]. Possible predictors of cRORA ER were subsequently entered in linear multivariate regression analysis with a backward selection procedure. p < 0.05 was considered statistically significant.

Results

Baseline Characteristics

A total of 204 eyes from 129 participants (64% females [n = 83] and 36% males [n = 46]) who met the inclusion criteria were recruited. The overall mean age was 80.6 ± 7.8 years (range from 64 to 94 years). The mean follow-up time was 4.2 ± 2.2 (range 2–10 years) years with a median of 4 years. 109 of 204 (53.4%) eyes were classified as MNV-associated GA in AMD (neovascular aspect being present before BL, at BL, or during follow-up period); 95 of 204 (46.6%) eyes were classified as pure GA in AMD. In 75 of 129 participants (58%), both eyes were affected; in 54 of 129 participants (42%), only one eye was affected. On average, the area of ​​the cRORA in the first SD-OCT measurement was 4.4 ± 4.5 mm² (range 0.23 mm²–23.0 mm²). Morphologically, the primary lesion was judged to be unifocal in 146 (72%) eyes and multifocal in 58 (28%) eyes. For baseline characteristics, see Table 1.

Table 1.

Baseline characteristics of study population

	Years	SD	Range
Overall mean age at BL	80.6	±7.8	64–94
Overall mean follow-up time	4.2	±2.2	2–10
		n	%
AMD type at BL	MNV-associated atrophy	109	53
	Pure GA	95	47
	Total	204	100
Atrophy pattern at BL	Unifocal	146	72
	Multifocal	58	28
	Total	204	100

SD, standard deviation.

Correlation between SD-OCT and FAF Area Measurement

A strong correlation was observed between the area of cRORA as measured on infrared (IR) SD-OCT and the area of hypoautofluorescence as measured by FAF in all available FAF measurements (n = 281; r = 0.924; p < 0.001). Dividing the correlation analysis into subgroups, there were very strong SD-OCT and FAF area correlations between unifocal and pure GA (r = 0.957; p < 0.001), unifocal and MNV-associated GA (r = 0.964; p < 0.001), multifocal and pure GA (r = 0.961; p < 0.001), as well as for the multifocal MNV-associated GA group (r = 0.884; p < 0.001).

Results after Square Root Transformation

After square root transformation, the overall mean ER of the cRORA area was 0.29 ± 0.19 mm/year (range 0.19–1.16 mm/year) (Table 2). Eyes with a follow-up of at least 5 years (n = 53) had a significantly lower mean ER compared to eyes with a shorter follow-up (n = 149) (0.25 ± 0.13 mm/year vs. 0.31 ± 0.21 mm/year; p = 0.041). In eyes which had received anti-VEGF injections, there was no significant difference in mean ER between eyes without and with intravitreal anti-VEGF injections (0.30 ± 0.19 mm/year vs. 0.28 ± 0.20 mm/year; p = 0.466). There was a weak negative correlation between the number of intravitreal anti-VEGF injections and mean ER (r = −0.140, p = 0.044). Eyes with a multifocal atrophy pattern at baseline (n = 58) had a significantly higher mean ER compared to eyes with a unifocal pattern (0.34 ± 0.19 mm/year vs. 0.27 ± 1.19 mm/year; p = 0.008) (Table 2). The mean ER for both the unifocal and the multifocal cRORA subgroup was higher at baseline and decreased over time. Only 2 out of 95 eyes (2.1%) with pure GA at BL developed MNV, which required intravitreal anti-VEGF treatment, during the course of the disease. A comprehensive analysis of ER pre- versus post-intravitreal anti-VEGF treatment was therefore not possible.

Table 2.

cRORA enlargement rates

			ER	SD	Unit
Overall mean ER	Before sqrt		1.44	±1.2	mm2/year
	After sqrt		0.29	±0.19	mm/year

	Yes/no	N	ER (mm/year)	SD	p values
Comparison of cRORA ER (after square root transformation)
Intravitreal anti-VEGF injections	Yes	109	0.28	±0.20	0.466
	No	95	0.30	±0.19
Unifocal atrophy pattern at BL	Yes	146	0.27	±1.19	0.008
	No	58	0.34	±0.19
Follow-up time ≥5 years	Yes	53	0.25	±0.13	0.041
	No	149	0.31	±0.21

SD, standard deviation.

Results without Square Root Transformation

Mean ER without square root transformation was 1.44 ± 1.2 mm²/year (Table 2) and revealed similar results with no significant difference between eyes receiving intravitreal anti-VEGF injections or not (1.47 ± 1.17 mm²/year vs. 1.39 ± 1.16 mm²/year; p = 0.602). For eyes receiving anti-VEGF treatments, there was also a significant weak correlation between the number of intravitreal injections and mean ER (r = −0.189; p = 0.006).

Correlation between ELM, IS/OS, PE Scores and BCVA

Correlation between ELM score (central 1 mm) and BCVA including all eyes was moderate yet highly significant at baseline, 1 year, 2 years, 3 years, 4 years, 5 years, and 7 years (r = −0.512, r = −0.569, r = −0.648, r = −0.558, r = −0.478, r = −0.478, r = −0.560; p < 0.001). Likewise, IS/OS and RPE scores (central 1 mm) showed a moderate correlation to BCVA at baseline, 1 year, 2 years, 3 years, 4 years, 5 years, and 7 years (r = −0.515, r = −0.568, r = −0.644, r = −0.547, r = −0.79, r = −0.558, r = −0.562; p < 0.001 and r = −0.473, r = −0.575, r = −0.644, r = −0.573, r = −0.492, r = −0.535, r = −0.562; p < 0.001). At baseline, eyes with cRORA only in the central 3 mm ETDRS ring had no significant difference in BCVA compared to eyes with cRORA in the central 6 mm ETDRS ring (62.51 ± 19.26 vs. 57.81 ± 22.40 ETDRS letters, p = 0.122).

Multivariate Regression Analysis

In multivariate regression analysis, the dispersion pattern and baseline lesion size of the cRORA were significant predictors of mean ER, whereas neither the necessity nor the number of intravitreal anti-VEGF injections predicted mean ER. A multifocal cRORA pattern at baseline (p = 0.022) and a smaller baseline lesion size (p = 0.036) were significantly associated with a higher mean ER.

Discussion

This study including a large number of eyes in a clinical routine setting showed a strong and highly significant correlation between the measurements of cRORA in FAF and SD-OCT. This indicates a comparability of these measurement modalities. Despite the area measurement of GA in the FAF being easier to perform due to higher contrast differences, evaluating cRORA in the IR image under simultaneous control in the OCT B-scan might be a good alternative. Since slight contrasts are more easily visible than in the SD-OCT IR image, FAF evaluation has the added benefit of allowing a more detailed assessment of the perilesional area and the dispersion pattern as it is the strongest prognostic factor in relation to the expected ER. FAF imaging has to be performed as an additional examination beside IR including OCT follow-up and is much more burden for the patients due to intensive flickering light in this examination. Further opacities of the ocular media might have more impact on FAF quality than IR/OCT quality. Another important issue favoring IR/OCT measurement is the limited availability of FAF imaging devices, especially in less retina-specialized settings like general ophthalmologists offices. Several authors have demonstrated good comparability of GA lesion area measurements comparing enface OCT, OCTA, and FAF in smaller settings than our study [25–27]. Based on cRORA measurements at OCT, in the future automated artificial intelligence-based tools might provide reliable guidance for the management of GA in clinical practice [28].

Previous work has already shown the importance of the square root transformation for evaluating the ER [22]. In our study, too, we were able to demonstrate linear growth of 0.29 ± 0.19 mm/year (range 0.19–1.16 mm/year) after the square root transformation with a mean follow-up time of 4.2 ± 2.2 years. This growth rate was comparable to Yehoshua et al. [29] who reported similar ERs for pure GA in AMD in a smaller study with shorter follow-up (mean 1.2 years). In their study, the mean ER for the square root of lesion area did not correlate with baseline size and was found to be 0.28 ± 0.17 mm/year. Other comparable results for ERs were provided by Keenan et al. [30] in AREDS2 Report Number 16 (ER = 0.28 ± 0.21 mm/year; follow-up = 4.5 ± 1.1 years) and Grassmann et al. [31] (ER = 0.30 ± 0.25 mm/year; follow-up = 4.5 ± 2.9 years) in both pure GA and MNV-associated GA in AMD.

Comparing our result for the mean ER of 1.44 ± 1.2 mm²/year without using the square root transformation, we see a similarity within a spread from 1.28 to 2.6 mm²/year in terms of ER reported by other authors [9, 19, 32–35]. However, after applying the square root transformation, the reported ERs were much more similar. Applying the square root transformation turned out to be crucial for the interpretation and analysis of our data and for better comparability with data from other studies.

One of the strengths of our study is the complete documentation of the anti-VEGF treatment over the entire follow-up period. Neither the total number nor the necessity of anti-VEGF therapy had a statistically significant influence on the ER. There was no significant difference in the ER between eyes that were treated with intravitreal anti-VEGF and untreated eyes. In eyes which had received anti-VEGF treatment, only a very weak negative correlation could be detected regarding the number of injections. Similarly, the CATT study did not observe a significant association between the number of injections and cRORA growth rate at 2- and 5-year follow-up [36]. The issue of cRORA progression under anti-VEGF therapy has been discussed controversially during the last years. Schütze et al. [37] performed a small study without control group showing a progressive RPE atrophy in eyes with nAMD under anti-VEGF therapy. The extension studies of big phase 3 trials also showed progressing macular atrophy in the late disease stages, eventually affecting nearly all study eyes. A progression of macular atrophy was associated with vision decline over that period [16, 38]. Due to lack of a nonanti-VEGF-treated control group, these studies were not able to distinguish between cRORA as a potential complication of VEGF inhibition at the RPE level and the natural disease course of nAMD causing RPE damage. In contrast to these previous studies, we were able to compare courses of disease in eyes with and without anti-VEGF therapy.

Our data showed that multifocal cRORAs grew significantly faster than unifocal cRORAs in both pure and MNV-associated GA in AMD (Fig. 4). This finding was consistent with Holz et al. [39], who showed that in pure GA in AMD the phenotypic features of FAF abnormalities had a much stronger impact on atrophy progression than any other risk factor that has been addressed in previous studies on progression of GA attributable to AMD. They introduced the “diffuse trickling” pattern that is associated with an extremely rapid progression of atrophy. We were able to support this observation in both pure GA and MNV-associated GA, and, with regard to our data, it represents the most important prognostic factor for the ER.

Fig. 4.

a Annual sequence of IR images showing the progression of unifocal cRORA. b Annual sequence of IR images showing the progression of multifocal cRORA.

ELM, IS/OS, pigment epithelium scores rated the extent of the lesion in the central mm² of the ETDRS grid. There was a statistically significant, moderate correlation between the BCVA and the extent of the damage to the layers mentioned. Previous work described a paracentral fixation [40], which could explain why with a score of 3 in all layers within the fovea (cRORA) there was not necessarily a complete drop in standard 4 m BCVA score. In addition, Sadda et al. [41] already expressed criticism of this measurement method since cRORA develops gradually with the formation of scotomas that spare the fovea initially, expanding into the visual field late in the course of the disease. Alternatively, they suggest functional assessments beyond BCVA including multifocal electroretinography, microperimetry, low-luminance visual acuity, reading speed, and contrast sensitivity to more fully assess visual impairment. A further argument for using alternative measures as study endpoints is supported by a study showing much better correlation of vision-related quality of life with contrast sensitivity and reading speed than with standard 4 m BCVA in patients with bilateral nAMD [42]. In these patients, the strongest correlations between the total area of macular GA as well as the percentage of GA in the central 1 mm and contrast sensitivity as functional measure were found [18].

Despite advances in the treatment of MNV-associated GA in AMD, the treatment of pure GA in AMD remains elusive. Nevertheless, various treatment options are currently under observation. Complement inhibition has been the focus of several recent clinical studies. Despite the lack of efficacy of lampalizumab [13] and eculizumab [14], newer drugs like pegcetacoplan [7], which very recently became FDA approved, and avacincaptad pegol [8] showed a reduction in lesion growth and therefore might have the potential to become promising treatments for GA. Other investigations aiming to understand the mechanisms of lipofuscin accumulation [43], treating a suspected mitochondrial dysfunction [44], involving gene therapy targeting the regulation of an overactive complement system [45], involving stem cell therapy [46] and neuroprotection [47] have been or are currently performed to better understand mechanisms of the disease and explore treatment targets. Retrospective analyses like our study might help to understand the natural course and characteristics of the disease in a setting resembling the real-life population we aim to treat in the future.

In conclusion, the results of our study could be relevant for further research in this area, especially with regard to the applicability of measuring cRORA using SD-OCT, since it represents a valid alternative to measuring cRORA in the FAF, but is significantly more widespread. The definition of the GA as cRORA turned out to be very helpful since there we found a strong correlation between the area measurements in the SD-OCT IR image and FAF. Therefore, we recommend using this definition for further studies and patient follow-ups in daily practice.

As a further implication for practice, our results could contribute to better informing our patients about the course of their disease. We did not find any evidence concerning higher ER of cRORA in patients who were treated with intravitreal anti-VEGF therapy. The best prognostic factor is the spreading pattern of the cRORA which should be assessed with FAF.

Statement of Ethics

This study protocol was reviewed and approved by the Ethics Committee Northwestern Switzerland, approval number BASEC 2021-01661. The study was carried out in accordance with the ICH-GCP guidelines and the principles of the Declaration of Helsinki. Written informed consent of the patients for use of retrospective data was obtained except in case of patients who have already died. The need for informed consent of the patients who have deceased was waived by the EKNZ (EKNZ Nr. BASEC 2021-01661). For details, see Methods section.

Conflict of Interest Statement

Vista Augenklinik Binningen received reimbursement for contract research from Novartis, Bayer, Roche, and Allergan/Abbvie. K.H. received fees for consultancies and advisory board participations from Novartis, Bayer, Roche, and Allergan/Abbvie. The authors L.C. and L.H. have no conflicts of interest to declare.

Funding Sources

The Vista Augenklinik is solely responsible for the financing of this retrospective evaluation.

Author Contributions

L.C.: acquisition and interpretation of data and writing of manuscript. L.H.: interpretation of data, statistics, and critical review of manuscript. K.H.: design of the study, supervision, and writing of manuscript.

Funding Statement

The Vista Augenklinik is solely responsible for the financing of this retrospective evaluation.

Data Availability Statement

The research data are not publicly available due to legal/ethical reasons (not covered by general consent). On special request to the corresponding author, a data sharing inquiry might be reached out to the local Ethics Committee.