Purpose
To analyze total thickness, morphology and individual vascular layers of the choroid in eyes with Stargardt disease using spectral-domain optical coherence tomography (SD OCT).
Design
Cross-sectional retrospective review.
Methods
Twenty-eight patients with Stargardt disease (53 eyes) with a mean age of 46 (15–79) years and 30 healthy subjects (30 eyes) with a mean age of 49 (22–79) years who underwent 1-line raster scanning with SD OCT were identified. Diagnosis of Stargardt disease was based on ophthalmic history and complete ophthalmic evaluation. The healthy subjects had best-corrected visual acuity of 20/20 or better with no chorioretinal pathology. Two independent raters assessed the total thickness, morphology, and the individual vascular layers of the choroid.
Results
The choroid was irregularly shaped in 26 of 41 eyes (64%) with Stargardt disease when compared to 0 of 30 healthy eyes (0%). Mean subfoveal total choroidal thickness and mean subfoveal large choroidal vessel layer thickness were significantly reduced in eyes with Stargardt disease when compared to healthy eyes (272.8 ± 32.8 μm vs 225.4 ± 69.9 μm; P = .03 and 219.5 ± 30.6 vs169.2 ± 70.1; P = .04, respectively). The maximal choroidal thickness was subfoveal in 9 of 41 eyes (22%), focal choroidal thinning was observed in 21 of 41 eyes (51%), and attenuation of large choroidal vessel layer was observed in 8 of 41 eyes (20%) with Stargardt disease. There was no association of the best-corrected visual acuity with any choroidal morphologic feature, except that it was better by a mean of 0.61 ± 0.21 in eyes that had preservation of large choroidal vessel layer (33 of 41, 80%) when compared to those that had attenuation of large choroidal vessel layer ( P = .007).
Conclusion
This study shows alterations in the total thickness, morphology, and the individual vascular layers of the choroid in eyes with Stargardt disease on SD OCT. These findings may potentially contribute to the clinical staging and monitoring of Stargardt disease.
Stargardt disease, also known as Stargardt macular dystrophy, is one of the most common inherited retinal diseases that is caused by disease-causing mutations in the ABCA4 gene. Typically, it presents in childhood or teenage years and involves an accumulation of lipofuscin in the retinal pigment epithelium (RPE) that causes its degeneration, and leads to a secondary photoreceptor cell and choriocapillaris atrophy. Affected individuals have significant variability in the phenotype and severity of disease. In advanced clinical stages, localized areas of chorioretinal degeneration involve the fovea and compromise vision.
Studies in animal models of retinal degeneration suggest that the loss of RPE precedes atrophy of the choriocapillaris and that reduced levels of vascular endothelial growth factor derived from the RPE may be a causative factor for their degeneration. This suggests that RPE modulates choriocapillaris structure and function in retinal degeneration. The choroidal vessels are responsible for the vascular support of the outer retinal layers and studies in eyes with Stargardt disease suggest that choroidal angiopathy may be involved in its pathogenesis.
Recent advances in spectral-domain optical coherence tomography (SD OCT), such as higher acquisition speed, better penetration, increased axial resolution, image averaging, and enhanced depth imaging, not only permit the visualization of precise anatomic detail of the individual retinal layers but also allow for an efficient visualization of the choroid. Previous studies have analyzed the total choroidal thickness in healthy and diseased states. More recently, there has been growing interest in analysis of the morphology of choroid and its individual vascular layers and assessment of choroidal vascular density with application of different techniques applied to SD OCT images.
Previous published studies have analyzed the total choroidal thickness using SD OCT in diseases such as age-related macular degeneration, diabetic retinopathy, and central serous chorioretinopathy. However, there is scarcity of such data for inherited retinal diseases such as Stargardt disease. A recent study analyzed the total choroidal thickness using SD OCT in various inherited retinal diseases including Stargardt disease, but it was limited by a small number of eyes (6 eyes). This study found varying degrees of choroidal thinning in half the eyes (3 of 6) with Stargardt disease, and the degree of choroidal thinning was dependent on the stage of disease. Given the evidence suggesting involvement of choroidal angiopathy in eyes with Stargardt disease, the present study aimed to analyze the choroid in a larger number of eyes with Stargardt disease. In addition to analyzing the total choroidal thickness, the present study also aimed to assess the morphology and the individual vascular layers of the choroid in eyes with Stargardt disease using SD OCT.
Methods
Subjects
Twenty-eight patients with a clinical diagnosis of Stargardt disease (56 eyes) and 30 healthy subjects (30 eyes) who underwent high-definition 1-line raster scanning at New England Eye Center, Boston, between November 1, 2009 and September 30, 2012 were retrospectively identified. Retina specialists at the New England Eye Center made the diagnosis of Stargardt disease. The diagnosis was clinical and was based on an ophthalmic history and a complete ophthalmic evaluation including dilated fundus examination, fundus photography, OCT imaging, fundus autofluorescence, and, in selected cases, fluorescein angiography, visual fields, and electroretinogram. Clinical features such as pisciform (yellow) flecks, macular atrophy, dark choroid on fluorescein angiography, and increased fundus autofluorescence were used to diagnose patients. Genetic diagnosis was not available. The exclusion criteria for this study included patients with a myopic refractive error of 6 diopters or worse, concomitant posterior segment pathologies, and a history of intraocular surgery. Details of the best-corrected visual acuity and automated central retinal thickness measurements were obtained from the patients’ charts. The healthy subjects (control group) had a best-corrected visual acuity of 20/20 or better, underwent fundus examination and OCT imaging, and were found to have no chorioretinal pathology. Healthy subjects with myopic refractive error of 6.00 diopters or worse were excluded. This study was approved by the institutional review board (IRB) of Tufts Medical Center and is adherent to the tenets of Declaration of Helsinki. Informed consent was considered exempt by the IRB owing to the retrospective design of this study.
Spectral-Domain Optical Coherence Tomography Imaging
SD OCT imaging was performed using the Cirrus high-definition SD OCT (Carl Zeiss Meditec Inc, Dublin, California, USA) that employs a wavelength of ∼840 nm, performs imaging at a speed of 27 000 A-scans per second, and has an axial resolution of ∼5.3 μm in tissue. The scan pattern employed was the 6 mm 1-line raster that acquires 20 frames (B-scans) at the same retinal location that are then averaged together to increase the signal-to-noise ratio. Averaging of images obtained from the same retinal location allows for a better visualization of the choroid. Images were not inverted during acquisition to bring the choroid in closer proximity to the zero-delay because image inversion using the Cirrus software leads to a low-resolution pixilated image. The enhanced depth imaging (EDI) protocol, which theoretically improves the visualization of the choroid, could not be employed in the present study because it was not available on the Cirrus device at the time the scans were performed. All scans included in the study had an intensity of 6/10 or better. All images were taken horizontally as close to the center of the fovea as possible to ensure that the thinnest point of the macula was imaged in both groups. This was performed with the understanding that slight differences in positioning may cause discrepancies in the choroidal analysis. Trained operators performed all OCT imaging. Scans of both eyes in each patient with Stargardt disease were used for the purpose of analysis. For comparison, 1 scan from each healthy subject that had a clearly visible choroidoscleral interface was selected for the purpose of analysis.
Analysis of Choroidal Morphology
Two independent raters (M.A., D.F.) from the Boston Image Reading Center evaluated the choroid for the morphologic features using previously described methods. Raters were masked to the diagnosis for evaluation of the choroidal morphology on SD OCT images. Any disagreements between the raters were resolved by open adjudication to make a final decision on the presence or absence of a particular feature. The clarity of the choroidoscleral interface was examined throughout the 6 mm line scan owing to its importance in choroidal thickness measurements. The contour and shape of the choroidoscleral interface was labeled as being either (1) convex (or bowl shaped) or (2) “S” shaped (having an irregular or concave/convex/concave shape with more than 1 inflection point). Foveal center was defined as the thinnest point of the macula. In cases where macular atrophy caused difficulty in assessing the center of the fovea, the SD OCT image was superimposed on the corresponding SD OCT fundus projection and/or a fundus photograph and the center of the fovea was found by eyeballing important anatomic landmarks such as the density of the macular pigment and the location of the foveal center in relation to the optic. The site of the thickest point of the choroid was identified and its location was compared to the overlying retina. It was defined as being subfoveal when the thickest point of the choroid fell within a 200 μm region centered on and beneath the foveal center. Areas of focal thinning of the choroid were identified. Focal thinning was defined as a thinning of the choroid at any location within the 6 mm line scan that is 50% or more than that in healthy eyes at the corresponding location. Mean choroidal thickness measurements in healthy eyes reported previously were used as a reference to determine areas of focal thinning in eyes with Stargardt disease. This previously published study assesses choroidal thickness in normal eyes belonging to a wide age range using the same SD OCT system as that used in the present study and it computes choroidal thickness at 11 locations that is beneath the foveal center and at 500 μm intervals up to 2500 μm temporal and nasal to the foveal center encompassing the full 6 mm line scan. An attenuation of the large choroidal vessel layer was defined as an absence of large choroidal vessels measuring ≥100 μm within a 1000 μm region centered on and beneath the foveal center. Preservation of the large choroidal vessel layer was defined as observation of large choroidal vessels measuring ≥100 μm within a 1000 μm region centered on and beneath the foveal center. RPE atrophy was defined as an interruption of the hyperreflective band between the inner hyperreflective interdigitation zone of the photoreceptor layer and the outer choriocapillaris layer, while outer retinal atrophy was defined as more than 50% reduction in the thickness of this layer compared to that in healthy eyes.
Analysis of Choroidal Thickness
Two independent raters (M.A., D.F.) from the Boston Image Reading Center, masked to the subject diagnosis, measured the total choroidal thickness perpendicularly from the outer edge of the hyperreflective RPE to the inner aspect of sclera at the fovea, and at 500 μm intervals up to 1500 μm temporal and nasal to the fovea (7 locations), using the Cirrus linear measurement tool. The average measurements of the 2 observers were used for the purpose of analysis.
Analysis of Choroidal Vasculature Beneath the Fovea
The choroidal vasculature was analyzed using a previously described method in healthy and diseased eyes. First the total choroidal thickness was measured beneath the fovea. A large choroidal vessel measuring ≥100 μm, located close to the choroidoscleral interface, and in the closest proximity to the fovea was selected and a perpendicular line from the innermost point of the identified vessel was drawn, to intersect the total choroidal thickness measurement line. The subfoveal large choroidal vessel layer (Haller layer) thickness was measured perpendicularly from the inner aspect of the sclera to the intersection point on the total choroidal thickness measurement line. The thickness of the Haller layer was subtracted from total choroidal thickness to obtain the medium choroidal vessel layer (Sattler layer)/choriocapillaris layer thickness. The ratio of the large choroidal vessel layer thickness to the total choroidal thickness was calculated to determine the contribution of large choroidal vessel layer to the overall choroidal thickness in both groups. Two independent raters (M.A., D.F.) from the Boston Image Reading Center, masked to the diagnosis of the subjects, performed all the measurements using the Cirrus linear measurement tool. The average measurements of the 2 observers were used for the purpose of analysis.
Statistical Analysis
All data were expressed as mean ± standard deviation. An unpaired Student t test was used to determine the difference in the thickness and vasculature measurements of the choroid in eyes with Stargardt disease and healthy eyes. One randomly selected eye per Stargardt disease patient was used for this quantitative comparison. Pearson correlation and intraclass correlation coefficient (ICC) were used to determine the interobserver agreement for all the measurements. Pearson correlation was used to assess the association of subfoveal choroidal vasculature measurements with the automated central retinal thickness measurements and the best-corrected visual acuity, as well as the association of the RPE and outer retinal atrophy with the preservation or attenuation of the large choroidal vessel layer in eyes with Stargardt disease. A 95% confidence interval and a 5% level of significance were adopted; therefore, results with a P value less than or equal to .05 were considered significant. All statistics were performed using Graph Pad Prism 5.0 software for Macintosh (GraphPad Software, La Jolla, California, USA) and SPSS software for Windows (Version 19.0; SPSS, Chicago, Illinois, USA).
Results
The demographic and clinical characteristics of the patients with Stargardt disease and healthy subjects are shown in Table 1 . The clinical phenotype of the Stargardt patients and thus the stage of disease were variable, from a few pisciform (yellow) flecks visible in the posterior pole without macular atrophy in some eyes to an exaggerated macular atrophy in others. There was no significant difference in the mean age and mean myopic refractive error between the 2 groups ( P = .32 and P = .83, respectively). Of the 28 patients (56 eyes), 1 eye of 1 patient had a history of membrane peeling for epiretinal membrane, 1 eye of 1 patient had a history of intraocular surgery for traumatic retinal detachment, and 1 eye of 1 patient had a myopic refractive error of more than 6 diopters. These eyes were excluded. Of the remaining 53 eyes of 28 patients, 12 eyes were excluded from the choroidal analysis owing to the lack of precise visualization of the choroidoscleral interface on SD OCT that did not allow the choroidal analysis performed in this study. Hence, 41 eyes of 28 patients were used for the purpose of choroidal analysis.
| Healthy Subjects | Stargardt Disease Patients | |
|---|---|---|
| Number of patients (eyes) | 30 (30 eyes) | 28 (53 eyes) |
| Mean age, y (range) | 49 (22–79) | 46 (15–79) |
| Mean myopic refractive error (range) | 1.25 (1.00–1.60) | 1.65 (1.15–2.25) |
| Sex | Male: 15 (50%) Female: 15 (50%) | Male: 12 (43%) Female: 16 (57%) |
| Duration of diagnosis, y (range) | NA | 10.2 (3–42) |
| Mean best-corrected visual acuity (range) | 20/20 | 20/200 (20/20 – light perception) |
Analysis of Choroidal Morphology
The morphologic features of the choroid and representative images with illustration of some of the morphologic features in eyes with Stargardt disease in comparison to healthy eyes are shown in Table 2 and Figure 1 , respectively. There was no association of the presence or absence of any choroidal morphologic feature with the best-corrected visual acuity and the duration of diagnosis except the association of best-corrected visual acuity with attenuation of the large choroidal vessel layer (see below).
| Choroidal Morphologic Feature | % of Healthy Eyes With Feature (N = 30) | % of Eyes With Stargardt Disease With Feature (N = 41) | P Value |
|---|---|---|---|
| 1. “Bowl”-shaped contour of the choroidoscleral interface (1 point of inflection) | 100% | 36% | <.0001 a |
| 2. Border of the choroid and sclera clearly identifiable throughout the 6 mm line scan | 78% | 68% | .12 |
| 3. Thickest point of choroid beneath the fovea | 97% | 22% | <.0001 a |
| 4. Focal thinning of the choroid | 0.0% | 51% | <.0001 a |
| 5. Attenuation of the large choroidal vessel layer b | 0.0% | 20% | .003 a |
b No large choroidal vessels ≥100 μm in diameter within a 1000 μm region centered on and beneath the foveal center.
Analysis of Choroidal Thickness
Mean choroidal thickness at all locations combined was significantly lower in eyes with Stargardt disease, when compared to healthy eyes ( P = .0007, Figure 2 ). The subfoveal choroidal thickness was significantly reduced in eyes with Stargardt disease, when compared to healthy eyes (272.8 ± 32.8 μm vs 225.4 ± 69.9 μm; P = .03, Figure 2 ). Mean choroidal thickness in healthy eyes showed a pattern of thinnest choroid nasally, thickest in the subfoveal region and thinning again temporally ( Figure 2 ), as has previously been reported. In eyes with Stargardt disease, however, the maximal choroidal thickness was not subfoveal in 32 of 41 eyes (79%) and focal thinning was observed in 21 of 41 eyes (51%) ( Table 2 , Figure 1 ). Focal thinning was exaggerated in the nasal region in 12 of 41 eyes with Stargardt disease (30%) ( Table 2 , Figure 1 ). This may explain the significant differences in the choroidal thickness measurements in the nasal locations, but not the temporal locations, in eyes with Stargardt disease compared to healthy eyes ( Figure 2 ). The choroidal thickness measurements in healthy eyes and those in eyes with Stargardt disease had a strong interobserver correlation (r = 0.98; P < .0001 and r = 0.93; P < .0001, respectively). The ICCs for choroidal thickness measurements between 2 observers were 0.96 (0.92–0.98) and 0.90 (0.88–0.94) for healthy eyes and eyes with Stargardt disease, respectively ( P < .0001).
