Optical Coherence Tomography–Assisted Enhanced Depth Imaging of Central Serous Chorioretinopathy

Abstract

Purpose.: To describe characteristics of central serous chorioretinopathy (CSC) imaged by optical coherence tomography–assisted enhanced depth imaging (EDI-OCT).

Methods.: The prospective observational case series study consisted of patients with acute or chronic CSC. All subjects underwent fundus fluorescein angiography, indocyanine green angiography (ICGA), and EDI-OCT.

Results.: The study included 68 eyes (68 patients) with 35 eyes showing signs of acute CSC. Mean subfoveal choroidal thickness, 478 ± 114 μm, was larger than the normative value from the Beijing Eye Study 2011 (254 ± 107 μm) on the same ethnic group. In the hyperfluorescent ICGA areas, EDI-OCT revealed a thinning of the inner choroidal layers and enlargement of the underlying hyporeflective lumina in all eyes. The diameter of the hyporeflective lumina (mean: 330 ± 103 μm) was significantly (P < 0.001) associated with subfoveal choroidal thickness (correlation coefficient [r]: 0.68). An RPE detachment was detected in 36 (53%) eyes. A double-layer sign defined as an undulated RPE layer and intact underlying Bruch's membrane (seen in 51 [75%] eyes) was significantly (P = 0.025) more often in the chronic CSC group (29/33; 87%) than in the acute CSC group (22/35; 63%). Prevalence of an RPE microrip (8/68 [12%] eyes) did not differ (P = 0.14) between the chronic CSC group (6/33; 18%) and the acute CSC group (2/35; 5.7%).

Conclusions.: CSC is characterized by a thinned inner choroidal layer and enlarged underlying hyporeflective choroidal lumina in all eyes, in addition to a dome-shaped RPE elevation, a double-layer sign of the RPE/Bruch's membrane complex, and RPE microrips in some eyes. EDI-OCT may be helpful in the diagnosis of CSC.

Introduction

Central serous chorioretinopathy (CSC) is characterized by a serous macular detachment in predominantly middle-aged men. Studies applying indocyanine angiography (ICGA) or recent investigations using enhanced depth imaging optical coherence tomography (EDI-OCT) have shown a vascular choroidal hypermeability in eyes affected by the disease and in clinically unaffected contralateral eyes. ^1–3 These findings on choroidal abnormalities prompted the hypothesis that hemodynamic changes in the choroid may cause an extravasation of excessive fluid leaking into the space beneath the retinal pigment epithelium (RPE). Mechanical forces might then lead to a breakdown of the RPE, resulting in a serous retinal detachment in the macular region. ⁴ In that model, primary choroidal vascular changes and secondary RPE lesions played important roles in the presumed pathophysiology of CSC. Minor morphologic changes of the choroid and the RPE could only poorly be visualized by conventional OCT, since often the wavelength of the light source used to image the retina was not long enough to penetrate into the choroid. This was due to wavelength-dependent light scattering and signal loss that occur in the image path, decreased sensitivity away from zero-delay, and the various inherited limitations with Fourier transformation. ^5–8 Conventional OCT could thus document mostly a serous retinal detachment and an RPE detachment in eyes with CSC. ⁶  

EDI-OCT introduced by Spaide and colleagues provides an in vivo cross-sectional information of the choroid, including choroidal thickness, large choroidal vessel, and RPE in addition to fine changes of retina. ^6,9–11 Using the EDI-OCT technology, recent studies have shown an abnormal thickness of the subfoveal choroid and dilated choroidal vessels in eyes with CSC. ^{1–3,6,9–15} Based on these investigations, it was the purpose of our study to examine in which percentage of a relatively large group of patients with CSC these changes occur and which additional abnormalities of the choroid and RPE may be found. 

Methods

The prospective clinical observational study included 68 eyes of 68 Chinese patients with CSC. The study was approved by the Institutional Review Board and Ethics Committee of the Capital University of Medical Science, and informed consent was obtained from all patients. The patients were examined between October 2011 and May 2012. For all patients, CSC was diagnosed by fluorescein angiography showing a focal leak at the level of the RPE, and by ICGA showing a choroidal hyperfluorescence. A subretinal detachment of the retinal photoreceptors from the RPE layer was demonstrated by spectral-domain OCT (Heidelberg Spectralis; Heidelberg Engineering Co., Heidelberg, Germany). Exclusion criteria were any history of eye diseases or any kind of ocular surgery or photocoagulation, current pregnancy, history of taking corticosteroids and sildenafil or related drugs, systemic diseases such as endogenous hypercortisolism, and arterial hypertension. Arterial hypertension was an exclusion criterion to avoid any potentially confounding effect of elevated blood pressure on the vascular system of the choroid. Polypoidal choroidal vasculopathy, masquerading as CSC with signs of a branching vascular choroidal network or polypoidal lesions in ICGA images, was specifically looked for and excluded. 

All eyes underwent spectral-domain EDI-OCT (Heidelberg Spectralis HRA + OCT; Heidelberg Engineering Co.). For all eyes, six radial scans centered on the fovea were obtained in a 5° × 30° field in a first step (Fig. 1). In a second step, rectangular scans encompassing the ICGA derived hyperfluorescent area with a minimum of seven lines and an interline distance of 240 μm were obtained vertically and horizontally. To reduce the speckle noise, a proprietary software algorithm was used to align images (TruTrack Active Eye Tracking; Heidelberg Engineering Co.) and an average of 100 frames was produced for each scan. The diameter of hyporeflective lumen in the choroid (large choroidal vessels as shown in indocyanine green angiography) was assessed perpendicular to Bruch's membrane in the region of the widest lumen diameter. Taking the hyporeflective lumen as surrogate for the large choroidal vessels, the largest one from the series of horizontal sections, vertical sections, or radial sections was taken and included in the statistical analysis (Figs. 1, 2). For eyes with more than one hyperfluorescent areas in ICGA, the data of the largest area were taken for statistical analysis. The subfoveal choroidal thickness was measured as the distance between the outer portion of the hyperreflective line corresponding with the RPE to the inner surface of the sclera. All OCT findings and measurements were performed by two experienced examiners (LY, WW) independently of each other. If the measurements by the two examiners differed by more than 15%, the examiners reperformed the measurement together. Subfoveal choroidal thickness measurements primarily differed between both examiners by >15% for two of the 68 patients, and choroidal vessel diameter measurements primarily differed for 26 of 68 patients. The reproducibility of the technique was assessed in a previous investigation on 3233 subjects. The mean difference in subfoveal choroidal thickness measurements obtained by two independent examiners of the same team as the actual study was 3.14 ± 13.1 μm. ¹⁶ As qualitative markers, the presence or absence of a microrip of the RPE, ¹⁷ of an RPE detachment, and of a double-layer sign were assessed. The double-layer sign consisted of an undulated RPE line and a hyperreflective straight line representing Bruch's membrane. ¹⁸  

Figure 1. 

 

Digital subtraction ICGA with simultaneous enhanced depth imaging assisted optical coherence tomogram of an eye with chronic central serous chorioretinopathy. Six radial scans were centered on the fovea. In the horizontal scan through the fovea (top image) was the diameter of the hyporeflective lumen 229 μm. The oblique scan (bottom image) shows a larger diameter of the hyporeflective lumen of 407 μm. The largest measurement (i.e., 407 μm) was used for the statistical analysis. Note: hyperreflective material (black arrow) beneath the retinal pigment epithelium in the top figure.

Figure 2. 

 

Digital subtraction ICGAs with simultaneous enhanced depth imaging assisted optical coherence tomograms of a 46-year-old man with acute central serous chorioretinopathy. In the left photos, the ICGA was taken at 2:15 minutes after injection of indocyanine green. (A) The vertical black long line shows a relative thick inner choroidal layer in the region surrounding the hyperfluorescent area. The diameter of hyporeflective choroidal lumina (large choroidal vessels) was measured in the meridian of its widest diameter (green line in the top right photograph). (B) The fifth EDI-OCT section (from the bottom) showed a marked thinning of the inner choroidal layer in the region of the large underlying hyporeflective lumen with a diameter of 384 μm (green line).

Statistical analysis was performed using a commercially available statistical software package (SPSS for Windows, version 20.0; IBM-SPSS, Chicago, IL). We performed a bivariate correlation analysis of associations between the subfoveal choroidal thickness and lumen diameter corresponding to the hyperfluorescence area. P values were two-sided and were considered statistically significant when the values were <0.05. 

Results

The prospective study included 68 eyes from 68 patients (57 males), with a mean age of 44.5 ± 8.0 years (range, 30–60 years). Of the 68 eyes, 35 eyes showed signs of acute CSC, with one or several leakage points at the level of RPE and duration of symptoms of less than 6 months, and 33 eyes were considered to have chronic CSC showing a mottling hyperfluorescence by fluorescein angiography and duration of symptoms of more than 6 months. All eyes demonstrated a serous macular detachment. The fellow eyes of all patients showed no serous retinal detachment. One eye (1.5%) had three hyperfluorescent areas and two eyes (3%) had two hyperfluorescent areas; all other eyes showed one hyperfluorescent area on ICGA. The hyperfluorescent area was located subfoveally in seven eyes (10%); in the other eyes, the hyperfluorescent areas were located in the parafoveal region or outside of the foveal region. 

Mean subfoveal choroidal thickness was 478 ± 114 μm (range, 232–695 μm). This value was larger than the mean subfoveal choroidal thickness of 253.8 ± 107.4 μm in the population of the Beijing Eye Study 2011 on the same ethnic group, in particular if a decrease in subfoveal choroidal thickness by 4 μm per year of life and by 15 μm per diopter of myopia is taken into account. ¹⁹  

In the hyperfluorescent ICGA areas, EDI-OCT revealed a relative thinning of the inner choroidal layer, including the small and medium large vessels, as indicated by a lower hyperreflectivity beneath Bruch's membrane in the hyperfluorescent region compared with the surrounding nonhyperfluorescent regions (Fig. 3). Beneath the thinned inner choroidal layer, hyporeflective lumina were identified in all eyes (Figs. 1–5). The internal reflectivity within these lumina appeared to be similar to that of subretinal fluid. The diameter of the hyporeflective lumina ranged from 104 to 549 μm, with a mean value of 330 ± 103 μm. This value was larger than the corresponding value for normal eyes (140 ± 40 μm), which had been examined in a recent study. ²⁰ The largest diameter of the choroidal lumina (vessels) in the hyperfluorescent areas was significantly (P < 0.001) associated with subfoveal choroidal thickness (r: 0.68). 

Figure 3. 

 

Digital subtraction ICGA with simultaneous enhanced depth imaging assisted optical coherence tomogram of an eye with chronic central serous chorioretinopathy. The region with dilated large choroidal vessels was demarcated on the oblique radial scan on the angiogram (two short green lines on the top image) and on the scan itself (long green lines in the bottom image). We measured the vertical distance between the retinal pigment epithelium and inner border of the dilated choroidal vessel and defined it as “inner choroidal layer.” This inner choroidal layer was thinner in the region with large (dilated) choroidal vessels (short black line) than in the neighboring region without dilated choroidal vessels (long black line).

A detachment of the RPE characterized by a dome-shaped RPE elevation with an intact underlying Bruch's membrane was detected in 36 (53%) eyes. The internal reflectivity within the RPE detachment was hyporeflective in 35 (97%) eyes (Fig. 4), and in one (3%) eye, it appeared to be hyperreflective (Fig. 1). A double-layer sign defined as an RPE layer with an undulated and not dome-shaped contour and intact underlying Bruch's membrane was seen in 51 (75%) eyes. It was detected significantly (P = 0.025; odds ratio [OR]: 4.28; 95% confidence interval: 1.23, 15.0) more often in the chronic CSC group (29/33 or 87%) than in the acute CSC group (22/35 or 63%). In the eyes with a double-layer sign, the space between the undulated RPE line and the straight Bruch's membrane line appeared hyporeflective in 41 (80%) eyes. In the remaining five (10%) eyes with a double-layer sign, the space was hyperreflective, and in the remaining five (10%) eyes, the reflectivity in the space between the undulated RPE line and Bruch's membrane line could not be determined on the images. In five of the 51 eyes with a double-layer sign, the double-layer line became continuous with a dome-shaped RPE detachment (Fig. 4). The double-layer sign and the dome-shaped RPE detachment were located within the hyperfluorescent area on the ICGA images. 

Figure 4. 

 

(A) Double-layer sign of the RPE/Bruch's membrane complex in CSC on digital subtraction ICGA and simultaneous enhanced depth imaging assisted optical coherence tomogram of a 58-year-old man with acute CSC. Corresponding to the hyperfluorescent area on ICGA, the simultaneous EDI-OCT showed an undulated RPE contour line (white arrow) and intact underlying Bruch's membrane with the material between the two layers showing a medium reflectivity. (B) Dome-shaped RPE detachment continuous with the double-layer sign in the same patient as shown in Figure 2. The first EDI-OCT section (from the bottom) showed an RPE detachment, which was continuous with the double-layer sign (short white arrow). The space beneath the RPE detachment was hyporeflective. Long white arrow: subfoveal choroidal thickness.

Figure 5. 

 

Digital subtraction ICGA with simultaneous enhanced depth imaging assisted optical coherence tomogram of a 54-year-old man with chronic central serous chorioretinopathy. Arrowhead: microrip of the RPE.

A microrip of the RPE or fissure defined as a breach in the RPE contour was detected in eight (12%) eyes. Its prevalence did not differ significantly (P = 0.14) between the chronic CSC group (6/33 or 18%) and the acute CSC group (2/35 or 5.7%). All RPE microrips were located within the areas of hyperfluorescence on the ICGA images (Fig. 5) 

Discussion

The results revealed that all eyes with CSC showed, in addition to a thickened subfoveal choroid, a thinning of the inner choroidal layer and an enlargement of the underlying hyporeflective lumina in the ICGA hyperfluorescent regions. The diameter of the hyporeflective lumina was significantly (P < 0.001) associated with subfoveal choroidal thickness. A double-layer sign of the RPE/Bruch's membrane complex (seen in 51 [75%] eyes) was significantly (P = 0.025) more common in the chronic CSC group than that in the acute CSC group. The prevalence of an RPE microrip (8/68; 12 eyes) did not differ significantly between the chronic CSC group and the acute CSC group. A dome-shaped RPE detachment was present in 36 (53%) eyes. 

The results of our study agree with previous investigations on an increased choroidal thickness in eyes with CSC and dilated choroidal vessels in the ICGA hyperfluorescent area. ^{1–3,6,12–15} The measurements in the previous studies (e.g., by Imamura et al. ³ : 505 ± 124 μm) yielded results similar to those in our study, 478 ± 114 μm (range, 232–695 μm). The finding of dilated choroidal vessels confirmed previous studies that applied indocyanine green angiography and reported on congested and dilated choroidal vessels close to the fovea in eyes with CSC. ^1,2 The diameter of the hyporeflective choroidal lumina with a mean value of 330 ± 103 μm agreed with another recent study on patients with CSC, in which the diameter of the largest choroidal vessels was 305 ± 101 μm. ²⁰ Besides the vessel dilatation as cause for the choroidal thickening, an interstitial choroidal edema may have played a role. Hyperpermeability of the choriocapillaris has been considered to be responsible for the focal choroidal hyperfluorescence observed on ICGA images in patients with CSC. ² The reasons have remained unclear as to why the choroidal vessels showed a dilatation with secondary choroidal thickening in patients with CSC. Since choroidal vessels are innervated by the autonomic system and since ganglion cells have been found in the posterior choroid, the dilatation of the choroidal vessels may have been regulated by innervation. ^21,22 Additionally, substances such as endothelin-1 or inflammatory cytokines may participate in the pathogenesis of the choroidal swelling in eyes with CSC, since endothelin-1 has been reported to be involved in the autoregulation of the choroidal vasculature. ^23,24  

An RPE detachment is a well-known feature of CSC. The prevalence of an RPE detachment of 53% (36/68) partially differs from the findings obtained in other investigations. Using a time-domain OCT and six radial scans, Montero and Ruiz-Monero ²⁵ detected an RPE detachment in 8% of eyes with acute CSC. Applying en face OCT, van Velthoven and colleagues ²⁶ found an RPE detachment in 52% of eyes with active CSC and in 100% of eyes with inactive CSC. Mitarai and associates ²⁷ reported on the prevalence of 63% for an RPE detachment in eyes with acute CSC. Reasons for the discrepancies between the studies may be the difference in the composition of the study population (acute CSC versus chronic CSC), the relatively small size of the study samples, and differences in the methodology such as the type of OCT device and the type of scanning pattern used. In addition to an already described dome-shaped RPE detachment, we found in our study a so-called double-layer sign characterized by two highly reflective layers: an undulated RPE line and a hyperreflective straight line representing Bruch's membrane. A similar finding has been reported for eyes with polypoidal choroidal vasculopathy. ¹⁸ Our study showed that the double-layer sign was not pathognomonic for polypoidal choroidal vasculopathy, but that it could also be found in CSC. In five eyes with a double-layer sign, the latter became continuous with a dome-shaped RPE detachment in the ICGA hyperfluorescent region. It may lead to the assumption that both changes were the result of choroidal hypermeability, and that the double-layer sign was a special form of an RPE detachment. Examining eyes with polypoidal choroidal vasculopathy, Ojima and colleagues ²⁸ noted a hyperreflective material between the two layers of the double-layer sign, and considered it as a sign of polypoidal lesions and branching vascular networks. In our study, however, the space between the RPE line and Bruch's membrane line in the region of the double-layer sign appeared to be hyporeflective in the majority (80%) of the eyes. One may postulate that the difference in the reflectivity between eyes with PCV and eyes with CSC may be due to differences between the two diseases in the hyperpermeability of choroidal vessels. 

A tear or rip of the RPE in CSC has been described in CSC eyes of patients after a steroid exposure or of younger patients showing subretinal fibrin and a rapidly increasing fluid accumulation. ²⁹ The patients included in our study had no history of steroid intake and were predominantly middle-aged. These differences may explain why the prevalence of RPE microrips was lower in our study population than that in the patients of previous investigations. An increased choroidal hydrostatic pressure due to the hypermeability of the choroidal vessels has been considered to be responsible for the RPE detachment. ^1,2,30 This mechanism may lead to a breach or microrip of RPE by which the pressure beneath the RPE and above the RPE equalizes. It has remained unclear whether such an RPE microrip is a necessary condition for the fluid access from the sub-RPE space into the subretinal space in eyes with CSC. In that case, a prevalence of 100% instead of 12% as in our study should be expected. It has remained elusive whether limitations in the axial and lateral resolution of available OCT images and the scanning pattern may have been reasons for missing some microrips. The spectral domain OCT (Spectralis; Heidelberg Engineering Co.) has roughly a lateral resolution of 14 μm and an axial resolution of 7 μm. In contrast to a typical large RPE rip characterized by a torn RPE and a bared choroid, an RPE microrip is a small defect in the RPE layer, which may close spontaneously. ³¹  

The present study in association with the investigations mentioned above showed that OCT characteristics of CSC were an increased subfoveal choroidal thickness with dilated hyporeflective lumina (vessels) beneath a thinned inner choroidal layer in the ICGA hyperfluorescent area, RPE detachments with hyporeflectivity beneath the detached RPE, and RPE microrips. These features may be useful to differentiate CSC from masquerading diseases affecting the macular region, such as Vogt-Koyanagi-Harada's syndrome, uveal effusion syndrome, polypoidal choroidal vasculopathy, posterior scleritis, pregnancy-induced choroidopathy, choroidal hemangioma, lymphoblastic leukemia, or other metastatic choroidal disease. ^32,33  

The internal reflectivity within the dilated choroidal lumina partially appeared to be similar to that of subretinal fluid. Since this was a study based on imaging techniques, one could not infer that the similar appearance of the fluid in the hyporeflective lumina and in the subretinal fluid could be taken as a hint that the biochemical compositions of the fluids were similar. 

Potential limitations of our study should be mentioned. First, the enhanced depth imaging mode of the OCT technique is an indirect method to image and measure the RPE, Bruch's membrane, and the choroid. It has not yet been unequivocally proven which histologic structures correspond to which line in the EDI-OCT images. Second, due to limitations in the spatial resolution of the EDI-OCT and the scanning pattern, some RPE microrips may have been overlooked, leading to an underestimation of the prevalence of this abnormality. Third, since it is an imaging study, nothing could be said about the biochemical composition of the material beneath the RPE and in the dilated choroidal lumina. 

In conclusion, acute and chronic CSCs are characterized by a thinning of the inner choroidal layer with an enlargement of the underlying large hyporeflective choroidal lumina found in all eyes, in addition to a dome-shaped RPE elevation, a double-layer sign of the RPE/Bruch's membrane complex, and RPE microrips in some eyes. EDI-OCT may be a valuable help in the diagnosis of CSC. 

Acknowledgments

Supported by Beijing Municipal Excellent Talent Foundation and Training Plan of High-Level-Health Talent of Health System Grant 2009‐3‐32, Beijing, China; Beijing Natural Science Foundation Grants 7092021 and 7112031; and National Natural Science Foundation of China Grant 81041018. 

Disclosure: L. Yang, None; J.B. Jonas, None; W. Wei, None