Abstract
Purpose:
To characterize choroidal neovascular (CNV) lesions and the corresponding change in visual acuity (VA) in eyes that converted to neovascular age-related macular degeneration (AMD) in the Age-Related Eye Disease Study 2–HOme Monitoring of the Eye Study. (AREDS2–HOME Study).
Design:
Cohort study.
Participants:
A total of 1520 participants at risk of developing CNV were enrolled, each of whom contributed 1 or both study eye(s) that had a best-corrected VA letter score of ≥54 letters (Snellen equivalent 20/60) and ≥1 large (≥125 μm) macular druse in the absence of neovascular AMD or central geographic atrophy.
Methods:
A multicenter clinical trial comparing standard care (SC) versus SC plus ForeseeHome (FH; Notal Vision, Manassas, VA) monitoring strategy in the detection of neovascular AMD. Fluorescein angiograms (FA) and OCT were evaluated by an independent reading center (RC) from the visit in which the ophthalmologist identified progression to CNV (n = 82 eyes).
Main Outcome Measures:
Development of CNV on OCT, FA, or both.
Results:
The RC confirmed CNV in 67 of 82 eyes (82%); lesions were confirmed in 42 of 70 eyes (60%) with FA, 59 of 72 eyes (82%) with OCT, and on both images in 34 of 67 eyes (51%). Among the FA-confirmed cases, the median lesion size was 0.82 disc area (DA); lesions were subfoveal in 40.5%, occult CNV composition was present in 54.8%, and associated hemorrhage in 50%. Median (interquartile range [IQR]) lesion size on FA was 0.23 (0.0–0.91) DA versus 0.70 (0.0–1.50) DA, P = 0.051) in the FH and SC eyes, respectively. Among the OCT-confirmed cases median (IQR) center point thickness was 209 (175–274) μm, retinal pigment epithelial lesion complex was present in 86.4%, and subretinal fluid (SRF) was present in 76.3%. The median change in VA from baseline was −4.0 letters and −10.0 letters in the FH and SC eyes (P = 0.008) confirmed as CNV at the RC.
Conclusions:
Incident CNV lesions were more prevalent on OCT images than on FA; however, the use of both OCT and FA enhanced detection of incident lesions. Lesions were smaller and associated with less vision loss among eyes in the FH group.
Neovascular age-related macular degeneration (AMD) has historically been responsible for the majority of irreversible central vision loss associated with AMD.1 With the advent of anti-vascular endothelial growth factor (anti-VEGF) therapy to manage eyes with neovascular AMD, rates of blindness due to AMD have been markedly reduced.2,3 Pivotal phase 3 trials evaluating these therapies demonstrate avoidance of moderate vision loss (≥3 lines of acuity) in up to 95% of affected eyes through at least 2 years of therapy.4–7 Baseline size of the choroidal neovascular (CNV) lesion and baseline visual acuity (VA) are associated with retention of better levels of VA after anti-VEGF therapy in multiple studies.8–11 As such, there is a strong rationale to detect incident neovascular AMD and initiate therapy before substantial vision loss has occurred.12,13
The HOme Monitoring of the Eye (HOME) study, an ancillary study of the Age-Related Eye Disease Study 2 (AREDS2-HOME study), evaluated the ForeseeHome (FH; Notal Vision, Manassas, VA) monitoring strategy in addition to standard care (SC) monitoring efforts (device arm) compared with SC monitoring alone (SC arm) for eyes at high risk of developing neovascular AMD.14 The primary end point of this multicenter randomized trial was development of neovascular AMD (CNV) as confirmed by the study ophthalmologist investigator, whereas the primary study outcome was the change in VA from baseline when the primary end point occurred. Median loss of VA was significantly less in the device arm relative to the SC arm at the visit in which incident CNV was documented: median −4.0 letters versus −9.0 letters (P = 0.021).9 Furthermore, a larger proportion of eyes in the device arm maintained vision of 20/40 or better at the primary end point compared with the SC group (87% vs. 62%, P = 0.014).9
The primary focus of this article is to describe the lesion characteristics identified in eyes that met the AREDS2-HOME study end point and converted to neovascular lesions. We also explored whether there are any differences in these features between eyes assigned to the monitoring device group and those in the SC group. The relationship between VA loss at the time of reading center (RC)-confirmed conversion from nonneovascular to neovascular AMD and image modality on which the conversion was confirmed are also explored.
Methods
The AREDS2-HOME study was conducted at 44 AREDS2 clinical centers across the United States. The trial design and primary outcomes are described elsewhere.14–16 In brief, 1520 participants were enrolled; for each participant cither 1 or 2 eyes with a best-corrected VA letter score of at least 54 letters (Snellen equivalent 20/60) and at least 1 large (≥125 μm) druse within the macula were studied. A total of 763 participants were randomly assigned to the device arm and 757 participants were assigned to the SC arm. Baseline color photographs were obtained for all participants to assess the Age-Related Eye Disease Study (AREDS) AMD severity level.17 Stereoscopic color photographs, fluorescein angiograms (FA), and OCT were required and were to be forwarded to the central RC (Fundus Photograph Reading Center, Madison, WI) at any study visit that was triggered by the FH device, symptom realization by the participant, or when clinical signs of neovascular AMD were present at an SC visit. The institutional review boards for human subject research from the individual clinical sites approved the protocol, and all participants signed informed consent forms. The research adhered to the tenets of the Declaration of Helsinki and complied with the Health Insurance Portability and Accountability Act.
All study images were obtained by certified photographers using standardized imaging protocols and image systems certified by the RC. The FA acquisition protocol included stereoscopic views of the optic disc and macula with images obtained during dye transit at 1, 2, 5 and 10 minutes. The OCT scan protocol included an instrument-specific cube scan from any of the following spectral domain instruments: Heidelberg Spectralis (Heidelberg Engineering, Heidelberg, Germany) or Zeiss Cirrus (Carl Zeiss Mcditec, Dublin, CA). If time-domain StratusOCT (Carl Zeiss Meditec) images were performed, the fast macular thickness scan was required.
All images were evaluated in a masked fashion at the RC. FA and OCT images were evaluated independent of each another and without comparison to previous images by a single trained and certified grader. The FA evaluation focused on CNV detection; lesion composition (predominantly CNV, predominantly classic, minimally classic, or occult CNV without classic CNV); and proximity to the fovcal center (subfoveal, juxtafoveal, or extrafoveal). Other CNV lesion components were also recorded, including serous pigment epithelial detachment, fibrous tissue, nonfibrous (atrophic) scar, and blood. The lesion area was measured as a sum of the CNV area and the area of associated lesion components. Total CNV area, total lesion area, and total leakage area were measured using planimetry. Color photographs served as complementary images to confirm the presence of blood, hard exudates, SRF, and fibrosis. When the RC identified CNV leakage on FA or at least one lesion component on FA or color photographs, the primary end point was considered confirmed by the RC.
OCT images were systematically reviewed for the following CNV associated features: presence, thickness, and location of retinal pigment epithelial (RPE) lesion complex (irregular elevation or thickening of RPE indicative of CNV); intrarctinal cysts; and the presence, location, and height of SRF. The eye was deemed to have progressed to CNV by the RC if any one of these features (RPE lesion complex, cysts, or SRF) was present on OCT. Quantitative measurements were collectcd following correction of segmentation errors and decentration when required. The manufacturer-provided software was used to measure thickness in the central subficld. Center point thickness was measured using digital calipers from the internal limiting membrane to the anterior surface of the RPE or the anterior surface of the RPE lesion complex.
Each FA or OCT initially determined as neovascular AMD by the clinical site ophthalmologist was subsequently reviewed by an RC ophthalmologist. This review could include a consensus review of both OCT and FA images simultaneously from the select visit and prior visits. Only eyes in which an RC ophthalmologist confirmed the presence of neovascular AMD were included in this report. Quality control assessments were performed on a random selection of 5% of all image submissions, irrespective of image findings.
Imaging characteristics of CNV lesions are reported by treatment group and in 3 categories: (1) FA and OCT characteristics for eyes that were RC-confirmed CNV events by either FA or OCT or both modalities, (2) FA characteristics of eyes confirmed as incident CNV on FA, and (3) OCT characteristics of eyes confirmed as incident CNV on OCT. The change in VA between baseline and the visit at which the RC confirmed the presence of neovascular AMD, by image modality and treatment group, is presented.
Descriptive statistics are provided. Mean, standard deviation, and median and interquartile range (IQR) were computed for continuous measurements, whereas frequency tabulations for categorical variables are provided. Comparisons between treatment groups were made using nonparamctric Wilcoxon two-sample tests for continuous variables and independent test for proportion for categorical variables. Because the distribution in change in VA from baseline did not follow a normal distribution, the Mann–Whitney test was used. All P values presented are based on one-sided tests and are unadjusted for multiple comparisons.14
Results
Eighty-two (5.4%) of the 1520 AREDS2-HOME participants met the study end point, progression to neovascular AMD (per the study ophthalmologist) in at least one eye during mean (standard deviation) follow-up of 1.4 (0.6) years. This consisted of 51 eyes (62%) among participants in the device arm and 31 eyes (38%) among participants in the SC arm. For study visits triggered by new symptoms or the FH device in the device group, the median time between the trigger and the study visit was 5 days (IQR, 3–8 days), whereas the median time between new symptoms and the study visit was 7.5 days (IQR, 4–18.5 days) in the SC arm.
Among the 82 eyes that met the study end point, 11 (13%) were missing an FA, and 9 (11%) were missing an OCT at the RC, whereas 1 additional eye had ungradablc images (both FA and OCT) at the end point visit (Table 1). As such, a gradablc FA was present in 70 eyes, a gradablc OCT in 72 eyes, and 67 eyes (82%) had both a gradablc FA and OCT at the visit labeled as neovascular AMD by the study ophthalmologist (Table 1).
Table 1.
Comparison of CNV Event Confirmation by the Reading Center Using Fluorescein Angiography versus OCT Images
| Frequency | CNV Confirmed by OCT | |||||
|---|---|---|---|---|---|---|
| Absent | Present | Cannot Grade | Missing | Total | ||
| CNV confirmed by FA | Absent | 7 | 20 | 0 | 1 | 28 |
| Present | 6 | 34 | 0 | 2 | 42 | |
| Cannot grade | 0 | 0 | 1 | 0 | 1 | |
| Missing | 0 | 5 | 0 | 6 | 11 | |
| Total | 13 | 59 | 1 | 9 | 82 | |
CNV = choroidal neovascular macular degeneration; FA = fluorescein angiogram.
Among the 70 eyes with a gradablc FA, neovascular AMD was confirmed by the RC on the FA in 42 (60%) eyes; among the 72 eyes with a gradable OCT image, neovascular AMD was confirmed by the RC on the OCT in 59 eyes (82%). When considering all available images (FA or OCT), the RC review confirmed neovascular AMD was present in 67 (82%) of the 82 eyes identified by the investigators. Confirmation on both image modalities was observed in 34 (51%) of the 67 eyes that had both modalities available for review. An additional 25 (30%) eyes meeting end point, 5 of which did not have an FA for review, were identified on OCT alone; an additional 8 (10%) eyes meeting end point, 2 of which did not have an OCT for review, were confirmed on FA only.
Confirmation occurred in 41 of 51 (80.4%) of FH eyes, and 26 of 31 (83.9%) of the SC eyes that met the primary end point. There was no difference in the number of missing images between the 2 groups: 10 (16.1%) in the FH group and 5 (16.1%) in the SC group (P = 0.69). Intergrader agreement for identification of FA and OCT features is available in Supplementary Table 1 (available at www.aaojoumal.org). Figures 1 through 3 show examples of eyes reaching the study end point confirmed by both image modalities (Fig 1), FA alone (Fig 2), and OCT alone (Fig 3).
Figure 1.
Choroidal neovascular macular degeneration (CNV) visible on fluorescein angiogram (FA) and OCT. The top row shows early and late-phase FA images with the arrow depicting CNV. The bottom row shows an OCT image with CNV (arrow) and corresponding line scan on the color photograph.
Figure 3.
Choroidal neovascular macular degeneration (CNV) visible on OCT only. Early and late-phase fluorescein angiogram images (top row) show fluorescein staining and no evidence of CNV. OCT shows active CNV with subretinal fluid at corresponding line scan depicted on the color photograph.
Figure 2.
Choroidal neovascular macular degeneration (CNV) on fluorescein angiogram (FA) only. Early and late-phase of FA showing CNV (arrow in top row). OCT in the corresponding area does not show any signs of CNV (bottom).
Specific Lesion Characteristics among Eyes with RC-Confirmed Neovascular AMD on FA and/or OCT
Table 2 shows specific neovascular lesion characteristics, by image modality, among the eyes of participants for whom the RC confirmed neovascular AMD was present on either FA, OCT, or both image modalities. The median (IQR) disc area (DA) values for CNV area, CNV lesion area, and CNV leakage areas for the FH arm/SC arm were 0.17 (0.0–0.69)/0.60 (0–1.50) DA; 0.23 (0–0.91)/0.70 (0–1.50) DA; and 0.55 (0.1–1.15)70.90 (0–2.20). Each variable indicated smaller areas of pathology in the device arm (P = 0.048 for CV area, P = 0.05 for CNV lesion area, and P = 0.032 for CNV lesion leakage). Blood was less commonly associated with CNV lesions in the device arm compared with the SC arm (26.8% vs. 46.2%, P = 0.054). Other features, such as fibrosis, atrophy, and hard exudates, were seldom present (n = 1 eye).
Table 2.
Specific Lesion Characteristics among Eyes Identified by the Ophthalmologist as Progressed to Neovascular AMD*
| FA Characteristics (N = 67) | Device Arm (n = 41) N (%) |
SC Arm (n = 26) N (%) |
Total (n = 67) N (%) |
P Value† |
|---|---|---|---|---|
| CNV absent or <50% of lesion | 19 (46.3) | 9 (34.6) | 28 (41.8) | 0.694 |
| Predominantly classic | 3 (7.3) | 4 (15.4) | 7 (10.5) | |
| Minimally classic | 3 (7.3) | 1 (3.9) | 4 (6.0) | |
| Occult CNV without a classic component | 14 (34.2) | 9 (34.6) | 23 (34.3) | |
| CNV location | ||||
| Subfoveal‡ | 10 (24.4) | 7 (26.9) | 17 (25.4) | 0.732 |
| Juxtafoveal‡ | 2 (4.9) | 3 (11.5) | 5 (7.5) | |
| Extrafoveal‡ | 11 (26.8) | 6 (23.1) | 17 (25.4) | |
| Missing‡ | 18 (44.0) | 10 (38.5) | 28 (41.8) | |
| Hemorrhage | ||||
| Absent | 28 (68.3) | 11 (42.3) | 39 (58.2) | 0.054 |
| Present | 11 (26.8) | 12 (46.2) | 23 (34.3) | |
| Subretinal fluid | ||||
| Absent | 28 (68.3) | 12 (46.2) | 40 (59.7) | 0.100 |
| Presence | 11 (26.8) | 11 (42.3) | 22 (32.8) | |
| CNV area (DA)§ | ||||
| N, Mean (SD) | 39, 0.49 (0.73) | 23, 1.11 (1.48) | 62, 0.72 (1.10) | |
| Median (IQR) | 0.17 (0–0.69) | 0.60 (0–1.50) | 0.23 (0–0.95) | 0.048 |
| Lesion area (DA) | ||||
| N, Mean (SD) | 39, 0.64 (1.04) | 23, 1.47 (2.53) | 62, 0.95 (1.77) | |
| Median (IQR) | 0.23 (0–0.91) | 0.70 (0–1.50) | 0.34 (0–1.25) | 0.051 |
| Leakage area from CNV (DA) | ||||
| N, Mean (SD) | 39, 0.72 (0.84) | 23, 1.49 (1.61) | 62, 1.01 (1.23) | |
| Median (IQR) | 0.55 (0–1.15) | 0.90 (0–2.20) | 0.62 (0–1.60) | 0.032 |
| Hemorrhage area (DA) | ||||
| N, Mean (SD) | 39, 0.11 (0.44) | 23, 0.44 (0.87) | 62, 0.23 (0.65) | |
| Median (IQR) | 0 (0–0.03) | 0.01 (0–0.46) | 0 (0–0.08) | 0.018 |
| OCT Characteristics (n = 67) | Device (n = 41) N (%) |
SC (n = 26) N (%) |
Total (n = 67) N (%) |
P Value |
| RPE lesion complex | ||||
| Absent | 9 (22.0) | 5 (19.2) | 14 (20.9) | 0.514 |
| Present | 31 (75.6) | 20 (76.9) | 51 (70.8) | |
| Location of RPE lesion complex | ||||
| Outside central subfield | 2 (4.9) | 2 (7.7) | 4 (6.0) | 0.876 |
| Within and outside central subfield | 29 (70.7) | 18 (69.2) | 47 (70.2) | |
| No lesion complex | 9 (22.0) | 5 (19.2) | 14 (20.9) | |
| Subretinal fluid | ||||
| Absent | 13 (31.7) | 7 (26.9) | 20 (29.9) | |
| Definite, outside central subfield only | 7 (17.1) | 6 (23.1) | 13 (19.4) | |
| Definite, involving central subfield | 20 (48.8) | 12 (46.2) | 32 (47.7) | |
| Intraretinal cysts | ||||
| Absent | 24 (58.5) | 16 (61.5) | 40 (59.7) | 0.538 |
| Definite, outside central subfield only | 1 (2.4) | 2 (7.7) | 3 (4.5) | |
| Definite, involving central subfield | 14 (34.1) | 7 (26.9) | 21 (31.3) | |
| Cannot grade | 1 (2.4) | 0 (0) | 1 (1.5) | |
| Serous pigment epithelial detachment | ||||
| Absent | 35 (85.4) | 22 (84.6) | 57 (85.1) | 0.949 |
| Definite, outside central subfield only | 1 (2.4) | 1 (3.9) | 2 (3.0) | |
| Definite, involving central subfield | 3 (7.3) | 2 (7.7) | 5 (7.5) | |
| Cannot grade | 1 (2.4) | 0 (0) | 1 (1.5) | |
| Center point thickness (μm) | ||||
| N, Mean (SD) | 40, 223.15 (92.31) | 25, 245.36 (117.91) | 65, 231.69 (102.59) | |
| Median (IQR) | 207.5 (179–241.5) | 211 (175–278) | 209 (179–260) | 0.295 |
| Subfoveal Height of Subretinal Fluid (μm) | ||||
| N, Mean (SD) | 20, 85.2 (82.44) | 12, 105.75 (109.73) | 32, 92.91 (92.41) | |
| Median (IQR) | 76 (0–118.5) | 77.5 (19–167) | 76 (0–149) | 0.414 |
| Subfoveal Height of RPE lesion complex (μm) | ||||
| N, Mean (SD) | 29, 108.17 (120.55) | 18, 169.56 (142.12) | 47, 131.68 (131.23) | |
| Median (IQR) | 76 (0–142) | 154.5 (64–234) | 94 (47–200) | 0.0444 |
| Maximum height of RPE lesion complex in the entire scan (μm) | ||||
| N, Mean (SD) | 31, 226.68 (113.14) | 19, 238.32 (149.52) | 50, 231.1 (126.81) | |
| Median (IQR) | 213 (142–294) | 189 (118–326) | 195.5 (142–304) | 0.492 |
AMD = age-related macular degeneration; CNV = choroidal neovascular macular degeneration; DA = disc area; FA = fluorescein angiogram; FH = ForeseeHome monitoring; IQR = interquartile range; RPE = retinal pigment epithelial; SC = standard care; SD = standard deviation.
Characteristics from FA and OCT on the 67 eyes confirmed as neovascular AMD on FA, OCT, or both at the reading center.
P-value calculation excludes missing data; 5 FAs (2 FH/3 SC) and 2 OCTs (1 FH/1 SC) are missing.
Disc area (DA); 1 DA = 2.54 mm2.
Subfoveal: most posterior border is 0 μm from the foveal center; juxtafoveal: most posterior border is 1 to 199 μm from the foveal center; extrafoveal: most posterior border is >200 μm from the foveal center; missing: distance to macula not measured for repeated visits.
Of the 67 OCT scans in this subgroup, an RPE lesion complex was identified on the OCT images in 51 (76.1%), SRF in 45(67.1%), intraretinal cysts in 24 (35.8%), and serous pigment epithelial detachment in 7 (10.5%) patients. Median (IQR) center point thickness of the macula was 209 μm (179–260 μm). No differences were detected between the treatment groups for any of the qualitative or quantitative variables evaluated on OCT.
Specific Neovascular Lesion Characteristics among Eyes with RC-Confirmed Neovascular AMD on FA
Among the 42 eyes in which the RC confirmed the presence of neovascular AMD on an FA, the median (IQR) CNV area, lesion area, and leakage area were 0.61 (0.23–1.50) DA, 0.82 (0.30–1.60), and 1.13 (0.61 – 1.92), respectively. Although each quantitative measure of CNV area, CNV lesion area, and leakage area favored smaller lesions in the device arm, the comparisons were not statistically significant (Table 3). Irrespective of treatment assignment, the majority of lesions were predominantly CNV, among which occult CNV without a classic component was the most frequent lesion composition, accounting for 54.8% of all events. Lesion location tended to be either subfoveal (40.5%) or cxtrafovcal (40.5%) without treatment group differences. Blood or SRF was commonly seen (50% and 52.4%, respectively) among these eyes, although the area occupied by blood was limited (median 0.005 DA). Other lesion components, such as fibrosis, atrophy, and serous pigment epithelial detachments, were rare (n = 1). No eyes were found to have retinal angiomatous proliferation or an RPE rip.
Table 3.
FA Characteristics of Neovascular AMD Lesions when the Event was Confirmed on FA
| Characteristic | Device Arm (n = 23) N (%) |
SC Arm (n = 19) N (%) |
Total (n = 42) N (%) |
P Value* |
|---|---|---|---|---|
| Composition of CNV lesion | ||||
| CNV <50% of lesion | 3 (13.0) | 5 (26.3) | 8 (19.1) | 0.561 |
| Predominantly classic | 3 (13.0) | 4 (21.1) | 7 (16.7) | |
| Minimally classic | 3 (13.0) | 1 (5.3) | 4 (9.5) | |
| Occult only | 14 (60.9) | 9 (47.4) | 23 (54.8) | |
| CNV location* | ||||
| Subfoveal | 10 (43.5) | 7 (36.8) | 17 (40.5) | 0.741 |
| Juxtafoveal | 2 (8.7) | 3 (15.8) | 5 (11.9) | |
| Extrafoveal | 11 (47.8) | 6 (31.6) | 17 (40.5) | |
| Hemorrhage | ||||
| Absent | 13 (56.5%) | 8 (42.1) | 21 (50) | 0.352 |
| Present | 10 (43.5%) | 11 (57.9) | 21 (50) | |
| Subretinal fluid | ||||
| Absent | 12 (52.2%) | 8 (42.1) | 20 (47.6) | 0.516 |
| Present | 11 (47.8%) | 11 (57.9) | 22 (52.4) | |
| CNV area (DA) | ||||
| Mean (SD) | 0.83 (0.79) | 1.35 (1.53) | 1.06 (1.19) | |
| Median (IQR) | 0.48 (0.23–1.31) | 0.65 (0.23–2.2) | 0.61 (0.23–1.5) | 0.232 |
| Lesion area (DA) | ||||
| Mean (SD) | 1.09 (1.16) | 1.77 (2.69) | 1.40 (2.00) | |
| Median (IQR) | 0.69 (0.3–1.6) | 0.99 (0.23–2.2) | 0.82 (0.3–1.6) | 0.307 |
| Leakage area (DA) | ||||
| Mean (SD) | 1.22 (0.77) | 1.81 (1.60) | 1.48 (1.24) | |
| Median (IQR) | 1 (0.6–1.72) | 1.47 (0.7–2.5) | 1.13 (0.61–1.92) | 0.144 |
| Hemorrhage area (DA) | ||||
| Mean (SD) | 0.19 (0.56) | 0.53 (0.94) | 0.34 (0.76) | |
| Median (IQR) | 0 (0–0.09) | 0.02 (0–0.67) | 0.005 (0–0.12) | 0.117 |
AMD = age-related macular degeneration; CNV = choroidal neovascular macular degeneration; DA = disc area; FA = fluorescein angiogram; IQR = interquartile range; SC = standard care; SD = standard deviation.
n = 3 missing in the SC arm for assessment of CNV location.
Specific Neovascular Lesion Characteristics among Eyes with RC-Confirmed Neovascular AMD on OCT
RC confirmation of neovascular AMD occurred in 59 eyes based on Cirrus OCT (n = 19), Spectralis OCT (n = 36), or StratusOCT (n = 4). As seen in Table 4, when neovascular AMD was confirmed on OCT, an RPE lesion complex was found in 51 (86.4%) eyes, the complex tended to be centrally located in 47 (79.7%) eyes, SRF was present in 45 (76.3%) eyes, and intraretinal cysts were found in 24 (40.7%) eyes. The lesion morphology appeared similar between the treatment groups with the exception of the thickness of the subfoveal RPE lesion complex, which was thinner in the device group (median [IQR] 76 fxm [0–1421 FH vs. 155 μm [64–234] SC, P = 0.044). The median center point thickness was 209 μm (IQR 175–274).
Table 4.
OCT Characteristics of Neovascular AMD Lesions when the Event was Confirmed on OCT
| Characteristics | Device Arm (n=37) N (%) |
SC Arm (n=22) N (%) |
Total (n=59) N (%) |
P Value |
|---|---|---|---|---|
| RPE lesion complex | ||||
| Absent | 6 (16.2) | 2 (9.1) | 8 (13.6) | 0.362 |
| Present | 31 (83.8) | 20 (90.9) | 51 (86.4) | |
| Location of RPE lesion complex | ||||
| Outside central subfield only | 2 (5.4) | 2 (9.1) | 4 (6.8) | 0.514 |
| Involving central subfield | 29 (78.4) | 18 (81.8) | 47 (79.7) | |
| Subretinal fluid | ||||
| Absent | 10 (27.0) | 4 (18.2) | 14 (23.7) | 0.732 |
| Outside central subfield only | 7 (18.9) | 6 (27.3) | 13 (22.0) | |
| Involving central subfield | 20 (54.0) | 12 (54.5) | 32 (54.3) | |
| Intraretinal cysts | ||||
| Absent | 21 (56.8) | 13 (59.1) | 34 (57.6) | 0.397 |
| Outside central subfield only | 1 (2.7) | 2 (9.1) | 3 (5.1) | |
| Involving central subfield | 14 (37.8) | 7 (31.8) | 21 (35.6) | |
| Cannot grade | 1 (2.7) | 0 (0) | 1 (1.7) | |
| Serous pigment epithelial detachment | ||||
| Absent | 32 (86.5) | 19 (86.4) | 51 (86.4) | 1.00 |
| Present | 5 (13.5) | 3 (13.6) | 8 (13.6) | |
| Center point thickness (μm) | ||||
| Mean (SD) | 225 (95) | 253 (124) | 236 (107) | |
| Median (IQR) | 209 (179–251) | 229 (155–310) | 209 (175–274) | 0.235 |
| Subfoveal height of subretinal fluid (μm) | ||||
| Mean (SD) | 85 (82) | 106 (110) | 93 (92) | |
| Median (IQR) | 76 (0–118.5) | 78 (19–167) | 76 (0–149) | 0.414 |
| Subfoveal height of RPE lesion complex (μm) | ||||
| Mean (SD) | 108 (121) | 170 (142) | 132 (131) | |
| Median (IQR) | 76 (0–142) | 155 (64–234) | 94 (47–200) | 0.044 |
| Maximum height of RPE lesion complex in the entire scan (μm) | ||||
| Mean (SD) | 227 (113) | 238 (150) | 231 (127) | |
| Median (IQR) | 213 (142–294) | 189 (118–326) | 196 (142–304) | 0.492 |
AMD = age-related macular degeneration; CNV = choroidal neovascular macular degeneration; DA = disc area; FA = fluorescein angiogram; 1QR = interquartile range; SC = standard care; SD = standard deviation.
Relationship between Vision Loss and Presence of RC-Confirmed Neovascular AMD
Table 5 shows the change in VA between baseline and the visit in which incident neovascular AMD was identified stratified by the method used to determine presence of neovascular AMD. Whether the neovascular event was confirmed by the ophthalmologist investigator or the RC, there was less vision loss among eyes assigned to the monitoring device than eyes using SC monitoring. Among the 67 eyes with RC-confirmed neovascular AMD, the median (IQR) change in VA was −4.0 (−10.0 to −2.0) letters for the device group and −10.0 (−14.0 to −4.0) letters for the SC group (P = 0.008).
Table 5.
Change in Visual Acuity between Baseline and Visit at which Neovascular AMD was Identified, Stratified by Method Confirming the Event
| Confirmation of Neovascular AMD | Device Arm Median (IQR) Change in VA from Baseline (letter score) | Standard Care Arm Median (IQR) Change in VA from Baseline (letter score) | Total Median (IQR) Change in VA from Baseline (letter score) | P Value |
|---|---|---|---|---|
| Investigator-identified event | N = 51 −4.0 (−11.0 to −1.0) |
N = 30* −9.0 (−14.0 to −4.0) |
N = 81* −7.0 (−12.0 to −2.0) |
0.021 |
| Reading center confirmation on FA and/or OCT | N = 41 −4.0 (−10.0 to −2.0) |
N = 6 −10.0 (−14.0 to −4.0) |
N = 67 −7.0 (−12.0 to −3.0) |
0.008 |
| Reading center confirmation on OCT | N = 37 −3.0 (−8.0 to −2.0) |
N = 22 −9.0 (−13.0 to −4.0) |
N = 59 −5.0 (−11.0 to −2.0) |
0.005 |
| Reading center confirmation on FA | N = 23 −4.0 (−11.0 to −3.0) |
N = 19, −12.0 (−25.0 to −4.0) |
N = 42 −8.0 (−13.0 to −3.0) |
0.006 |
AMD = age-related macular degeneration; FA = fluorescein angiogram; IQR = interquartile range; VA = visual acuity.
Visual acuity not available in 1 eye.
Discussion
The design of the AREDS2-HOME study allows characterization of incident neovascular AMD among a cohort of patients with nonneovascular AMD who were being monitored for disease progression. The dataset from this study is unique in that the study concentrated on the timely diagnosis of neovascular AMD as it prospectively tested a new monitoring paradigm to facilitate early detection of conversion from nonneovascular to neovascular disease. Direct comparison of the features describing new-onset CNV lesions in the AREDS2-HOME study with other reports describing baseline features of participants in clinical trials evaluating treatment for neovascular AMD is problematic. Clinical trials frequently specify eligibility criteria for entry VA; angiographic confirmation of CNV, including lesion location, lesion composition, lesion size, and potentially other CNV-associated features; and evidence of activity using FA/OCT. Furthermore, the lesion present at a baseline study visit may not be indicative of the lesion characteristics at initial presentation. Inclusion criteria for trials may bias against the earliest onset lesions, such as may be imposed by a requirement for VA impairment, subfoveal location, or presence of SRF. Note that in the AREDS2-HOME study, as eyes were prospectively monitored for conversion to CNV, incident CNV was often identified in the absence of SRF or foveal involvement (30% and 60%, respectively). In addition, the lesion size as measured on FA among AREDS2-HOME study participants is considerably smaller compared with those enrolled in the Comparison of Age-Related Macular Degeneration Treatment Trial (CATT) study18 (mean 0.95 DA HOME study vs. mean 1.71 DA CATT) and much smaller than the lesions in the Inhibition of VEGF in Age-related Choroidal Neovascularization (IVAN) study (median 0.34 DA HOME study vs. median 3.71 DA IVAN study).18
Early diagnosis of CNV lesions can be associated with identification of lesions with features associated with better retention of V A at diagnosis and preservation of a better level of VA 1 or more years after initiation of anti-VEGF therapy. The current standard therapy for neovascular AMD—repeated intravitreal injections of anti-VEGF agents—can result in 90% of affected patients avoiding moderate vision loss through 2 years of treatment.6,7 Identifying lesions before vision is lost remains critical to preserving the highest level of function. An increasing number of recent publications also indicate that lesion size at presentation is an important predictor for treatment response.9 Analysis of baseline characteristics in phase 3 clinical trials for ranibizumab (Lucentis, Genentech, San Francisco, CA) showed that better baseline VA and smaller lesion size are associated with more favorable vision outcomes, and CATT confirmed these observations.19 Our report indicates that new-onset neovascular AMD can be identified when minimal vision loss has occurred, the fovea is spared, lesion composition is less aggressive, and the lesion size is small, particularly when using the FH strategy. Eyes assigned to the FH strategy had less vision loss at diagnosis of CNV whether the progression was identified by the investigator or confirmed by an independent RC masked to the clinical circumstances, and these eyes had smaller lesions than those using SC monitoring methods alone.
The sensitivity and specificity of diagnosing CNV with OCT has been shown to be comparable to FA in patients clinically presenting with CNV.20,21 In the current study with preclinical neovascular disease, onset was confirmed simultaneously on both image modalities in about half of the eyes that had gradable FA and OCT images from the visit at which the eye met the study end point. Among the 82 eyes that met the ophthalmologist-determined study end point, the RC was more likely to confirm this on OCT compared with FA (80.5% vs. 59%, respectively). In eyes with both image modalities available for RC review, there was 61% agreement between the 2 image modalities for the RC determination of CNV presence or absence. A larger number of eyes identified by only 1 modality were documented by OCT compared with FA (20 vs. 6 cases), suggesting OCT may be more sensitive to early disease diagnosis. However, the availability of both forms of imaging led to more eyes receiving RC confirmation of a CNV event, highlighting the advantage of obtaining both OCT and FA images when there is suspicion of early-onset CNV. The observation that CNV was identified in only 1 image modality among 39% of the RC-confirmed end point eyes with gradable pairs of imaging is consistent with the early nature of these lesions. Early lesions may have subtle presenting features that may be visible on either FA or OCT, whereas as the disease progresses, more severe findings may be readily apparent on both image modalities.
There were 15 investigator-determined end point eyes (18%) that were not confirmed as events by the RC, among which 7 eyes had missing or ungradable images in at least one of the image modalities. Disagreement between physicians on the presence of CNV may occur; therefore, it should be expected that there would be some differences between study investigators and the RC. Factors that may have contributed to study ophthalmologists diagnosing neovascular AMD more often than the RC may include the ophthalmologists’ ability to simultaneously review FA and OCT images (and possibly compare images from previous visits), and the ability to perform clinical stereoscopic fundus examination and to factor VA measurements or participant symptoms into their decision-making process.
This study suggests that CNV lesions can be identified in clinical practice at an early stage when they occupy a small area, spare the fovea, lack a classic CNV component, and are associated with limited signs of exudation (fluid, blood, and lipid). Early lesions such as these may not necessarily qualify for treatment with anti-VEGF agents. Natural history studies are required to understand if early lesions require treatment or close observation. Efforts to facilitate early diagnosis include regular reminders to routinely monitor for symptoms, enhancement of self-monitoring with the ForeseeHome device between routine examinations, reminders to contact the ophthalmologist promptly whenever there is concern for disease progression, and investigation of new signs/symptoms with FA and OCT. In the future, advances in OCT angiographic analysis may further enhance earlier detection of neovascular AMD. At present, findings from anatomic image analysis in the AREDS2-HOME study demonstrate favorable lesion characteristics at the time of CNV development in patients using FH home monitoring in addition to SC. These more favorable anatomic attributes and the association with minimal VA loss at presentation represent positive predictors of better VA outcomes with proper therapy.
Supplementary Material
Acknowledgments
Financial support was provided by Notal Vision Ltd. through a clinical trial agreement with the National Eye Institute (No. CTA-00833) and a service agreement with EMMES Corporation. The AREDS-2 study was supported by the intramural program funds and contracts from the National Eye Institute/National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland (contract no. HHS-N-260-2005-00007-C; ADB contract no. N01-EY-5-0007.) This work was also supported in part by an unrestricted grant from Research to Prevent Blindness, Inc to the University of Wisconsin-Madison Department of Ophthalmology and Visual Sciences.
Abbreviations and Acronyms:
- AMD
age-related macular degeneration
- AREDS-2
Age-Related Eye Disease Study 2
- CNV
choroidal neovascular macular degeneration
- DA
disc area
- FA
fluorescein angiogram
- FH
Foresee Home monitoring device
- HOME
Home Monitoring of the Eye
- IQR
interquartile range
- RC
reading center
- RPE
retinal pigment epithelial
- SC
standard care
- SD
standard deviation
- SRF
subretinal fluid
- VA
visual acuity
- VEGF
vascular endothelial growth factor
Footnotes
Financial Disclosure(s):
The author(s) have made the following disclosure(s): A.D.: Research support – University of Wisconsin.
T.E.C.: Research support – The Emmes Corporation.
S.B.: Employee –Johns Hopkins University; Grant support – Notal Vision Ltd, Boehringer-Ingelheim, Biophytis, Genentech, Mylan Inc, Novartis, and Bayer.
R.P.D.: Research support – University of Wisconsin.
M.E.: Consultant – Genentech; Research support – Genentech, Alcon/Novartis, Thrombogenics, Apellis Pharmaceuticals and Neurotech, YD Global Life Science.
J.E.K. Consultant – Notal Vision, Genentech, and Alimera Science; Research support – Notal Vision and Optos.
D.M.B.: Research support – Notal Vision, Genentech/Roche, Regeneron/Bayer, and Alcon/Novartis; Consultant – Notal Vision, Genentech/Roche, Regeneron/Bayer, and Alcon/Novartis.
HUMAN SUBJECTS: Human subjects were included in this study. Institutional review boards for human subject research from the individual clinical sites approved the protocol. The study was performed in accordance with the tenets of the Declaration of Helsinki. Written, informed consent was obtained from all participants.
No animal subjects were used in this study.
Supplemental material available at www.ophthalmologyretina.org.
References
- 1.Ferris FL 3rd, Fine SL, Hyman L. Age-related macular degeneration and blindness due to neovascular maculopathy. Arch. Ophthalmol. 1984;102:1640–1642. [DOI] [PubMed] [Google Scholar]
- 2.Varma R, Bressler NM, Doan QV, et al. Visual impairment and blindness avoided with ranibizumab in Hispanic and non-Hispanic whites with diabetic macular edema in the United States. Ophthalmology. 2015;122:982–989. [DOI] [PubMed] [Google Scholar]
- 3.Lee AY, Lee CS, Butt T, et al. UK AMD EMR Users Group Report V: benefits of initiating ranibizumab therapy for neovascular AMD in eyes with vision better than 6/12. Br J Ophthalmol. 2015;99:1045–1050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Rosenfeld PJ, Brown DM, Heier JS, et al. Ranibizumab for neovascular age-related macular degeneration. N Engl J Med. 2006;355:1419–1431. [DOI] [PubMed] [Google Scholar]
- 5.Brown DM, Michels M, Kaiser PK, et al. Ranibizumab versus verteporfin photodynamic therapy for neovascular age-related macular degeneration: two-year results of the ANCHOR study. Ophthalmology. 2009; 116:57–65.e55. [DOI] [PubMed] [Google Scholar]
- 6.Martin DF, Maguire MG, Fine SL, et al. Ranibizumab and bevacizumab for treatment of neovascular age-related macular degeneration: two-year results. Ophthalmology. 2012;119:1388–1398. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chakravarthy U, Harding SP, Rogers CA, et al. Alternative treatments to inhibit VEGF in age-related choroidal neo-vascularisation: 2-year findings of the IVAN randomised controlled trial. Lancet. 2013;382:1258–1267. [DOI] [PubMed] [Google Scholar]
- 8.Ying GS, Maguire MG, Daniel E, et al. Association of baseline characteristics and early vision response with two-year vision outcomes in the Comparison of AMD Treatments Trials (CATT). Ophthalmology. 2015; 122:2523–253l.e2521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Finger RP, Wickremasinghe SS, Baird PN, et al. Predictors of anti-VEGF treatment response in neovascular age-related macular degeneration. Surv Ophthalmol. 2014;59:1–18. [DOI] [PubMed] [Google Scholar]
- 10.Talks JS, Lotery AJ, Ghanchi F, et al. First-year visual acuity outcomes of providing aflibercept according to the VIEW study protocol for age-related macular degeneration. Ophthalmology. 2016;123:337–343. [DOI] [PubMed] [Google Scholar]
- 11.Regillo CD, Busbee BG, Ho AC, et al. Baseline predictors of 12-month treatment response to ranibizumab in patients with wet age-related macular degeneration. Am J Ophthalmol. 2015;160:1014–1023.el012. [DOI] [PubMed] [Google Scholar]
- 12.Ho AC, Albini TA, Brown DM, et al. The potential importance of detection of neovascular age-related macular degeneration when visual acuity is relatively good. JAMA Ophthalmol. 2017;135:268–273. [DOI] [PubMed] [Google Scholar]
- 13.Keane PA, de Salvo G, Sim DA, et al. Strategies for improving early detection and diagnosis of neovascular age-related macular degeneration. Clin Ophthalmol. 2015;9:353–366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Chew EY, Clemons TE, Bressler SB, et al. Randomized trial of a home monitoring system for early detection of choroidal neovascularization home monitoring of the eye (HOME) study. Ophthalmology. 2014;121:535–544. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chew EY, Clemons TE, Bressler SB, et al. Randomized trial of the ForeseeHome monitoring device for early detection of neovascular age-related macular degeneration. The HOme Monitoring of the Eye (HOME) study design—HOME Study report number 1. Contemp. Clin. Trials. 2014;37:294–300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Chew EY, Clemons TE, Harrington M, et al. Effectiveness of different monitoring modalities in the detection of neovascular age-related macular degeneration: The HOME Study.report number 3. Retina. 2016;36:1542–1547. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Davis MD, Gangnon RE, Lee LY, et al. The Age-Related Eye Disease Study severity scale for age-related macular degeneration: AREDS report No. 17. Arch Ophthalmol. 2005;123:1484–1498. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Chakravarthy U, Harding SP, Rogers CA, et al. Ranibizumab versus bevacizumab to treat neovascular age-related macular degeneration: one-year findings from the IVAN randomized trial. Ophthalmology. 2012;119:1399–1411. [DOI] [PubMed] [Google Scholar]
- 19.Rosenfeld PJ, Shapiro H, Tuomi L, et al. Characteristics of patients losing vision after 2 years of monthly dosing in the phase III ranibizumab clinical trials. Ophthalmology. 2011;118:523–530. [DOI] [PubMed] [Google Scholar]
- 20.Wilde C, Patel M, Lakshmanan A, et al. The diagnostic accuracy of spectral-domain optical coherence tomography for neovascular age-related macular degeneration: a comparison with fundus fluorescein angiography. Eye. 2015;29: 602–610. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Gualino V, Tadayoni R, Cohen SY, et al. Optical coherence tomography, fluorescein angiography, and diagnosis of choroidal neovascularization in age-related macular degeneration. Retina. 2018. In Press. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.