Geographic atrophy (GA) is a significant cause of progressive visual impairment characterized by the progressive loss of retinal pigment epithelium (RPE), photoreceptors, and choriocapillaris (CC). GA is the end-stage of age-related macular degeneration (AMD), the leading cause of central vision loss in Western countries in patients older than 65 years.1
The characterization of GA, which was initially based on color fundus photography, has been improved by the well-delineated area obtained by blue- or green-light fundus autofluorescence (FAF) imaging,2 and more recently by optical coherence tomography (OCT), which became essential in the diagnosis and classification of GA.3
Recently, an international group of experts in AMD and imaging (Classification of Atrophy Meetings [CAM]) proposed a consensus definition and nomenclature for OCT-defined atrophy in AMD.3
The experts recommend using OCT as the reference technique to define atrophy phenotypes and stages and to classify the different stages as complete RPE and outer retinal atrophy (cRORA), incomplete RPE and outer retinal atrophy (iRORA), complete outer retinal atrophy (cORA), and incomplete outer retinal atrophy (iORA). The group has also stressed the utility of FAF to define atrophy, and this technique remains the gold standard for measuring the GA area based on the size of hypoautofluorescence in AMD. FAF also allows identifying distinct phenotypic patterns that have an impact on disease progression.4 We will use the term of GA as lesions were diagnosed on fundus and color photos before being documented on other devices.
OCT angiography (OCTA) is a noninvasive tool that allows visualization of vascular flow in various retinal layers. In patients with atrophic AMD, the CC flow within a GA lesion has been shown to be severely impaired.5 We assumed that this area of CC flow impairment on OCTA, which we will refer to as CC nonperfusion area, with direct visualization of the deep choroidal vessels on the CC segmentation, could be used as an indirect indicator of the GA surface.
The aim of this study was to compare the GA area measured on OCTA (CC nonperfusion area) and on the gold standard FAF.
Patients and Methods
A prospective, observational, cross-sectional study was conducted in a tertiary center specializing in imaging and treatment of retinal diseases (Explore Vision Center, Paris, France). All patients diagnosed with GA between January 2017 and July 2017 were included.
This study was conducted in accordance with the tenets of the Declaration of Helsinki, and an informed consent was obtained from all patients. A positive opinion was obtained from the France Macula Federation ethics committee.
Inclusion criteria were all consecutive eyes of patients diagnosed with GA that met the CAM criteria (report 3)3 from January 2017 to May 2017. GA diagnosed on color fundus images was confirmed by blue FAF and OCT B-scan. OCT allowed detection and exclusion of neovascular AMD and classified GA between cRORA, iRORA, cORA, and iORA. All patients presented with cRORA on OCT B-scan.
Exclusion criteria were presence or history of choroidal neovascularization in the included eye.
All patients underwent an ophthalmological examination including FAF (Heidelberg Engineering, Heidelberg, Germany), OCTA (XR Avanti; Optovue, Fremont, CA), and OCT B-scan after pupil dilation (tropicamide and neosynephrine).
Based on these examinations and patient age, one of the retina specialists in the retinal clinical practice made the diagnosis of GA. We used the Angioflow module on the OCTA device (XR Avanti; Optovue, Fremont, CA) with the latest software update, which decreases projection artifact.
We performed 3 mm × 3 mm acquisitions for a better resolution and 6 mm × 6 mm acquisitions (in six patients) when the atrophy area was larger than the 3 mm × 3 mm acquisition area.
The preset parameters were used to obtain the CC segmentation slab. All images were anonymized. One of the investigators (GL) measured manually using calipers the visualization area of the choroidal vessels in the window effect of CC nonperfusion (Figure 1). Two measurements were obtained: the area measured in mm2 and the automatic measurement of the choroidal flow area within the same surface.
Measurement methods for choriocapillaris (CC) nonperfusion and geographic atrophy (GA) areas. Optical coherence tomography angiography, 3 mm × 3 mm acquisition, CC segmentation, with the CC nonperfusion area delimited manually by a yellow line using calipers for surface size quantification (A). Within the limits of the CC nonperfusion area, a choroidal flow is detected (yellow) and automatically quantified (B). Autofluorescence imaging identifying the GA area (hypoautofluorescence) (C). This area is manually delimited by a yellow line using calipers for surface size quantification (D).
One of the investigators (GL) also measured manually the area of hypoautofluorescence on blue FAF images (Figure 1).6 Once the area manually delimited, the total surface was obtained using Region Finder software (Heidelberg Engineering, Heidelberg, Germany). OCTA was compared with the OCT B-scan to verify that the CC nonperfusion area was superimposable over the GA area (Figures 2 to 4).
Example of a patient with geographic atrophy (GA). Choriocapillaris (CC) nonperfusion is visible within the margin of GA (complete retinal pigment epithelium (RPE) and outer retinal atrophy). (A) 3 mm × 3 mm optical coherence tomography (OCT) angiography acquisition of the posterior pole. CC segmentation showing the area of CC nonperfusion corresponding to the area of deep choroidal vessels visualization by hypertransmission (yellow arrow). (B) En face OCT, same segmentation as A, showing hypertransmission corresponding to GA. The horizontal green line shows the location of B-scans (C, D). (C) Horizontal B-scan showing flow area in red. The white arrows show the termination of the RPE, the interrupted ellipsoid zone (EZ) and the termination of external limiting membrane (ELM), and EZ and the location of ELM descent. The yellow dotted lines show the margin of RPE atrophy. (D) Horizontal B-scan, same location as C.
Example of a patient with geographic atrophy (GA). Choriocapillaris (CC) nonperfusion is visible within the margin of GA (complete retinal pigment epithelium (RPE) and outer retinal atrophy). (A) 6 mm × 6 mm optical coherence tomography (OCT) angiography acquisition of the posterior pole. CC segmentation showing the area of CC nonperfusion corresponding to the area of deep choroidal vessels visualization by hypertransmission. (B) En face OCT, same segmentation as A, showing hypertransmission corresponding to GA. The horizontal green line shows the location of B-scans (C, D). (C) Horizontal B-scan showing flow area in red. The white arrows show the termination of the RPE and the interrupted ellipsoid zone. The red line shows the temporal margin of RPE atrophy, and the yellow dotted line shows the nasal margin of RPE atrophy. (D) Horizontal B-scan, same location as C.
Example of a patient with geographic atrophy (GA). Choriocapillaris (CC) nonperfusion is visible within the margin of GA (complete retinal pigment epithelium and outer retinal atrophy). (A) 3 mm × 3 mm optical coherence tomography (OCT) angiography acquisition of the posterior pole. CC segmentation showing the area of CC nonperfusion corresponding to the area of deep choroidal vessels visualization by hypertransmission. (B) En face OCT, same segmentation as A, showing hypertransmission corresponding to GA. The horizontal green line shows the location of B-scans (C, D). (C) Horizontal B-scan showing flow area in red. The white arrows show the termination of the retinal pigment epithelium (RPE) and the interrupted ellipsoid zone. The yellow dotted lines show the margin of RPE atrophy. (D) Horizontal B-scan, same location as C.
The primary endpoint was the GA area based on the CC nonperfusion area measured on OCTA compared with the GA area measured on the gold standard method of FAF.
Secondary endpoint was the choroidal flow area measured on OCTA within the margin of the CC nonperfusion (Figure 1).
Statistical analysis was performed using a paired t-test with PRISM 7 software. A Pearson test was used to assess correlations. A P value of less than .05 was considered statistically significant.
Forty-two eyes were screened, and two eyes with choroidal neovascularization seen on OCTA despite the absence of history of neovascular AMD and the absence of fluid on the OCT B-scan were excluded. Forty eyes of 34 patients (27 women and seven men) with a mean age of 82.63 years ± 9.21 years (range: 66 years to 100 years) were included. There were 23 right eyes and 17 left eyes. Thirty-four eyes were examined using 3 mm × 3 mm acquisitions. However, in six patients, the GA size exceeded the 3 mm × 3 mm area, and 6 mm × 6 mm acquisitions were performed and used for measurements.
In all cases, we observed a CC nonperfusion on OCTA within the edges of the GA area (Figures 2 to 4). The mean GA surface on FAF images was 2.184 ± 3.045 mm2. The mean CC nonperfusion area measured on OCTA images was 2.349 ± 3.237 mm2. The CC nonperfusion area measured on OCTA was significantly larger (P = .035) than the GA area measured on FAF. All of the CC nonperfusion areas (Figure 1A) were larger except for five out of the 40 eyes. The mean difference between the CC nonperfusion area and the GA area on FAF was −0.165 ± 0.290 mm2 (min/max: −2.848/1.308 mm2). We found a strong linear correlation between measurements made using both techniques (r = 0.97; P < .0001; confidence interval = 0.98–0.99) (Figure 5).
Correlation curve based on a Person correlation test. A linear correlation is observed between the choriocapillaris nonperfusion area assessed on optical coherence tomography angiography (OCTA) and the geographic atrophy (GA) area measured on fundus autofluorescence (FAF).
A choroidal flow (Figure 1B) was present and measurable in all cases in the CC segmentation within the limits of the CC nonperfusion area. The mean choroidal flow area within the CC nonperfusion area was 1.427 ± 1.427 mm2. This measurement allowed assessing the flow density within the margins of the CC nonperfusion area at 61%.
In this prospective, observational, cross-sectional study, we compared the CC nonperfusion area measured on OCTA and the GA area measured on FAF. We found a strong correlation between the GA area measured on FAF and the CC nonperfusion area measured on OCTA. Our study revealed that the CC nonperfusion area measured on OCTA was slightly but significantly larger than the GA area measured on FAF.
To the best of our knowledge, there is to date no report of the use of OCTA to assess the GA area in atrophic AMD based on the CC nonperfusion area measurement. However, it has been previously shown that the CC nonperfusion area measured on OCTA was larger than the macular atrophy area measured on FAF in case of neovascular AMD.5 As suggested by Takasago et al.,5 these results support that the CC nonperfusion occurs at the same time or earlier than GA.
GA results from the degeneration of photoreceptors, RPE and CC. FAF is based on the RPE function. Areas of RPE cell atrophy or loss correspond to areas of hypoautofluorescence on FAF. However, in the case of GA, there is also a surrounding hyperautofluorescence area corresponding to a suffering RPE, and this area is not taken into account in the quantification of the GA area. This could partly explain the difference in area observed between the measurements made on FAF and on OCTA. However, OCTA allows measuring the CC nonperfusion area, which could be larger and more accurate in the follow-up of GA or at least could complete the assessment made by FAF and OCT. The OCTA measurement could thus be used as a surrogate marker in clinical studies assessing GA.
In a previous study,7 it has been shown that there was a CC flow impairment in the GA area in atrophic AMD and a less severe CC flow impairment beyond the GA area.8 In our study, we confirmed the absence of CC perfusion on OCTA within the GA area assessed by FAF, and we also confirmed by a larger measurement of the CC nonperfusion area compared with the GA area that CC nonperfusion areas were present outside the GA area. This is consistent with previous studies showing a focal CC dropout outside the GA limits.7–9 These results suggest a role of the CC in the pathophysiology of GA. In addition, although the chronology of the disease is incompletely known, the fact that the CC nonperfusion area was larger than the GA area supports an earlier dysfunction of the CC or at least that the RPE and photoreceptors could survive despite the presence of a CC nonperfusion.
However, our study has some limitations. We used blue FAF in this study because it was the only FAF available, whereas it has been shown that there is a slightly better inter-reader agreement using SLO green FAF, suggesting that its use may be preferable in clinical trials assessing changes in lesion size as a clinical endpoint.6
Visualization of CC on OCTA can be subject to several biases, including shadowing by drusen or other hyperreflective structures and low signal-to-noise ratio, making analysis of details in standard images inaccurate. However, we used here a robust measurement that relied on no visible CC with increased visualization of deeper large vessels by hypertransmission. We then suggest this measurement as an interesting and convenient method. However, we acknowledge that before using such a method in trials, other parameters such as intra- and inter-rating reproducibility have to be validated.
In conclusion, in this study we found that in GA, the CC nonperfusion area correlates well with the GA area assessed by FAF. Currently, the CC on OCTA is not taken into account in the measurement of GA and its classification. However, it could be a convenient method of monitoring the GA area. As the CC nonperfusion area in GA is larger than the GA area assessed by FAF, this suggests a possible early degeneration of the CC before the impairment of the RPE and photoreceptors. Thus, assessing the CC nonperfusion area could even improve the accuracy when monitoring the GA area in clinical trials testing drugs for GA.
- Lindekleiv H, Erke MG. Projected prevalence of age-related macular degeneration in Scandinavia 2012–2040. Acta Ophthalmol. 2013;91(4):307–311. doi:10.1111/j.1755-3768.2012.02399.x [CrossRef]
- von Rückmann A, Fitzke FW, Bird AC. Distribution of fundus autofluorescence with a scanning laser ophthalmoscope. Br J Ophthalmol. 1995;79(5):407–412. doi:10.1136/bjo.79.5.407 [CrossRef]7612549
- Sadda SR, Guymer R, Holz FG, et al. Consensus definition for atrophy associated with age-related macular degeneration on OCT: Classification of atrophy report 3. Ophthalmology. 2018;125(4):537–548. doi:10.1016/j.ophtha.2017.09.028 [CrossRef]
- Holz FG, Bindewald-Wittich A, Fleckenstein M, Dreyhaupt J, Scholl HP, Schmitz-Valckenberg SFAM-Study Group. Progression of geographic atrophy and impact of fundus autofluorescence patterns in age-related macular degeneration. Am J Ophthalmol. 2007;143(3):463–472. doi:10.1016/j.ajo.2006.11.041 [CrossRef]17239336
- Takasago Y, Shiragami C, Kobayashi M, et al. Macular atrophy findings by optical coherence tomography angiography compared with fundus autofluorescence in treated exudative age-related macular degeneration. Retina. 2019;39(2):296–302. doi:10.1097/IAE.0000000000001980 [CrossRef]
- Pfau M, Goerdt L, Schmitz-Valckenberg S, et al. Green-light autofluorescence versus combined blue-light autofluorescence and near-infrared reflectance imaging in geographic atrophy secondary to age-related macular degeneration. Invest Ophthalmol Vis Sci. 2017;58(6):BIO121–BIO130. doi:10.1167/iovs.17-21764 [CrossRef]28632841
- Kvanta A, Casselholm de Salles M, Amrén U, Bartuma H. Optical coherence tomography angiography of the foveal microvasculature in geographic atrophy. Retina. 2017;37(5):936–942. doi:10.1097/IAE.0000000000001248 [CrossRef]
- Choi W, Waheed NK, Moult EM, et al. Ultrahigh speed swept source optical coherence tomography angiography of retinal and choriocapillaris alterations in diabetic patients with and without retinopathy. Retina. 2017;37(1):11–21. doi:10.1097/IAE.0000000000001250 [CrossRef]
- Kim DY, Fingler J, Zawadzki RJ, et al. Optical imaging of the chorioretinal vasculature in the living human eye. Proc Natl Acad Sci U S A. 2013;110(35):14354–14359. doi:10.1073/pnas.1307315110 [CrossRef]23918361