Aims To evaluate the variable depth tomographic features of choroidal vasculature in acute and chronic central serous chorioretinopathy (CSC) using swept source optical coherence tomography (SS-OCT) en face imaging.
Methods We retrospectively reviewed the en face SS-OCT images of 29 patients that presented with acute (12 eyes) or chronic (17 eyes) CSC. All of the patient eyes underwent 6×6 macular scans with SS-OCT (DRI OCT-1, Topcon, Tokyo, Japan), fluorescein angiography and indocyanine green angiography. The en face image was used to investigate the choroidal vasculature of each layer. Moreover, we determined that some parts corresponded to choriocapillaris and Sattler's layer attenuation, whereas choroidal vessel dilatation was associated with Haller's layer.
Results At Haller's layer level, choroidal vessel dilatation was observed in 11 of 12 acute CSC (91.7%) and 15 of 17 chronic CSC (88.2%). In acute CSC, choroidal vessel dilatation was divided into focal (9/11; 81.8%) and diffuse (2/11; 18.2%) patterns. The chronic CSC cases demonstrated different patterns of choroidal vessel dilatation: focal (5/15; 33.3%) and diffuse (10/15; 66.6%). Ten of the acute CSC eyes (83.3%) and 14 of the chronic CSC eyes (82.4%) were found to have obscured choriocapillaris and Sattler's layers on en face imaging.
Conclusions En face imaging of SS-OCT is useful when combined with angiography in CSC for evaluating choroidal vessel dilatation at Haller's layer and to identify obscured upper layers. We identified different choroidal vessel dilatation patterns between acute and chronic CSC. These findings might be useful for pathophysiological understanding of CSC.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Several studies have suggested that CSC originates in abnormal choroidal vessel; however, the pathophysiology of CSC is not fully understood.3–5 Functional and morphological features on indocyanine green angiography (ICGA) and optical coherence tomography (OCT) strongly suggest that the choroid is primarily involved. Choroidal hyperpermeability, vascular congestion and increased choroidal thickness (CT) were found to be possible in the pathophysiology of CSC.6 ,7
A number of studies have reported choroidal vascularity in CSC. Yang et al8 measured the largest subfoveal lumen in eyes with CSC using enhanced depth imaging OCT (EDI-OCT) line scanning, but the entire choroid was not evaluated. Kuroda et al9 reported that CSC CT was greater at the fovea and leakage point on high penetration OCT compared with that of a normal person, although there were some limitations to choroidal vessel evaluation.
OCT is a non-invasive tool that provides retinal and choroidal cross-sectional images and is essential for diagnosing and evaluating CSC.10–12 Spectral domain OCT (SD-OCT) has higher axial resolution and faster acquisition time compared with time domain OCT. Conventional SD-OCT involves a large amount of light scattering at the retinal pigment epithelium (RPE), resulting in poor visualisation beneath the RPE. In contrast, swept source OCT (SS-OCT) with a long-wavelength light source can visualise the microstructures beneath the RPE and choroid.13–17 Recent software developments can construct en face images of the posterior pole from SS-OCT data. The en face images facilitate screening of the entire lesion, which is difficult with a conventional cross-sectional B-scan view.18–23 Recently, the en face SS-OCT has been used to study various retinal disorders, and one group reported various en face SS-OCT images of chronic CSC.19 However, the report was limited to chronic CSC.
In this study, we observed acute and chronic CSC focusing on choroidal vascular topographic features in variable depths on en face SS-OCT images, and investigated the relationship between en face SS-OCT choroidal vascular patterns and chronicity of CSC.
This retrospective cross-sectional study design was approved by the Institutional Review Board of Hanyang University Medical Center, Seoul, South Korea. From May 2014 to September 2014, we retrospectively reviewed the medical records of consecutive patients with idiopathic CSC who were examined at Hanyang University Hospital. The review of the subjects' medical history in regard to hypertension, diabetes, smoking and use of mineralocorticoid antagonists was included. All patients underwent comprehensive ophthalmic examinations, including fluorescein angiography (FA) and ICGA. The FA and ICGA were obtained using an F-10 confocal scanning laser ophthalmoscope (cSLO) (Nidek, Gmagori, Japan), and the size of the scanning field was 40°. Systemic work-up in all patients also included SS-OCT with a B-scan and en face image.
Idiopathic CSC was diagnosed based on the presence of serous detachment of the neurosensory retina involving the macula, as demonstrated by SS-OCT, leakage at the level of RPE on FA and hyperpermeability on ICGA. From the result of multimodal imaging, the patients with associated choroidal neovascularisation were excluded from the study. Patients with steroid-induced CSC, organ transplant-associated CSC or a history of treatment such as laser photocoagulation or photodynamic therapy were excluded. We also excluded severe media opacity such as cataract, which could degrade image quality to produce an unreadable en face SS-OCT image.
If the duration of subjective symptoms such as visual loss, metamorphopsia, chromatopsia or micropsia was less than 6 months, the CSC was classified as acute; if the symptoms had been present for 6 months or more, the disease was classified as chronic.10
SS-OCT system and en face imaging
We used DRI-OCT1 (Topcon, Tokyo, Japan) based on SS-OCT technology, with a scanning speed of 100 000 A-scans/s. The scan protocol was a 3D volumetric macular scan 6×6 mm square containing 512×128 A-scans. The centre wavelength of the probe beam was 1060 nm. This long-wavelength probe enables deep choroidal penetration. The axial resolution of this system is 8 µm in tissue. Automated built-in software was used to create a CT map of 6×6 mm from 128 images of each 3D data set. The consecutive en face images between RPE and chorioscleral interface were constructed using software codeveloped with Topcon.21
An intensity- and gradient-based, semiautomatic algorithm was used to segment the RPE to generate a reference surface for en face display. En face images of the choroid were extracted at varying depths from the RPE reference surface. This analysis involved point-to-point correlation between en face images and cross-sectional B-scan images. Atypical features that appeared on the en face images were characterised through systematic analysis of multimodal fundus imaging records, including FA and ICGA.19 To compare the two sets of images accurately, the en face images were resized to fit the scale of the fundus image (FA and ICGA) provided by cSLO (F-10).
We obtained representative en face images at different en face depths, including the choriocapillaris and choroidal layers with medium-sized vessels (so-called Sattler's layer) and large-sized vessels (so-called Haller's layer). En face SS-OCT images were retrospectively evaluated and compared with FA and ICGA findings by three of four investigators (WJL, JWL, BRL) in a blinded fashion. The presence and the patterns of choroidal vascular dilatation were evaluated using the method introduced in several previous studies.19 ,22 ,23 For standardised comparison, reference photographs were used (figure 1). Theses photographs depict Haller's layer. If there was disagreement among investigators, the majority of author's results were applied. We performed the examination of images in a fully masked fashion.
Statistical analyses were performed using PASW Statics V.18 (SPSS, Chicago, Illinois, USA). Pearson's χ2 test or Fisher's exact test was used to compare en face patterns between CSC groups.
A total of 29 eyes of 29 patients were included in this study; these included 12 eyes with acute CSC and 17 eyes with chronic CSC. The mean age of the patients in the study was 49.83±8.69 years (range, 28–65 years). Males represented 22 of the patients and females represented 7 patients. The mean symptom duration was 32.3±43.5 days in the patients with acute CSC and 271.8±140.7 days in the patients with chronic CSC. Medical history and detailed clinical profiles of the eyes studied are presented in online supplementary table. The mean CT at the fovea was 334.1±50.5 μm in the patients with acute CSC and 281.8±71.1 μm in the patients with chronic CSC (table 1). There was no significant difference between the CT of acute CSC and that of the chronic CSC (p=0.052), but the CT in acute CSC tended to be greater than that in chronic CSC.
Clinical profiles and choroidal vasculature patterns of the central serous chorioretinopathy patients.
Patterns of choroidal dilatation: focal and diffuse
Choroidal en face SS-OCT imaging in each layer revealed dilatation of choroidal vessels in two somewhat distinct patterns (figure 1). Some vascular branches that were found to have larger calibres than the surrounding choroidal vessels were characterised as ‘focal choroidal dilatation’. Relatively homogenous dilatation of vessels observed in the entire series of en face images was characterised as ‘diffuse choroidal dilatation’.19 ,23 All dilated choroidal vessels seen on the en face images coincided with those seen in the ICGA images.
At the Haller's layer level, choroidal vessel dilatation was observed in 11 of 12 acute CSC (91.7%) and 15 of 17 chronic CSC (88.2%) eyes. In acute CSC, choroidal vessel dilatation was divided into two patterns: focal (9/11; 81.8%) and diffuse (2/11; 18.2%) (figure 3). Chronic CSC cases demonstrated a different distribution of the two patterns: focal (5/15; 33.3%) and diffuse (10/15; 66.6%) (figures 4 and 5). The difference between the two CSC groups was statistically significant (p<0.001) (table 1). Image analysis showed a potential correlation between the location of choroidal vessel dilatation on en face imaging and the thickened area on CT map by OCT in both acute and chronic CSC.
Inner choroidal layer obscuration
We found attenuation of the inner choroidal layer (choriocapillaris and Sattler's layer) on a cross-sectional B-scan, which might have been caused secondarily by the dilatation of outer choroidal vessels (Haller's layer). Also, the dilatation of choroidal vessels was visible at inner en face depths than expected, which suggests that the inner choroidal attenuation is associated with outer choroidal vessel dilatation. We defined these findings as inner choroidal layer obscuration (figure 2).
Ten of the acute CSC eyes (83.3%) and 14 of the chronic CSC eyes (82.4%) revealed obscured choriocapillaris and Sattler's layer on the en face image. The presence of inner choroidal obscuration showed no significant difference between patients with acute and chronic CSC (p=0.51).
This study showed different choroidal vessel dilatation patterns on en face SS-OCT images between acute and chronic CSC. The majority of acute CSC cases showed focal choroidal vascular dilatation on en face images that were correlated to hyperpermeable lesions on ICGA (figure 3). In comparison, chronic CSC cases demonstrated different choroidal vessel dilatation patterns that had a general diffuse distribution (figure 5). This is a novel observation, and this discrimination might support the hypothesis that the origin of the disease is the choroidal vessels level. In the acute phase, choroidal dilatation might start focally in the origin of the disease focus. As the disease course becomes chronic, generalised vascular changes occur, which could be caused by the remodelling of choroidal vasculature or exudative changes of choroid.
Recently, one group reported choroidal dilatation of patients with CSC using en face images. However, the OCT used in that study was EDI-OCT not SS-OCT, and discrimination of focal or diffuse patterns of choroidal dilatation and analysis according to disease chronicity were not performed. In addition, this study reported that choroidal dilatation was found in all of the enrolled patients with CSC irrespective of active or quiescent disease, whereas the minority of patients from our study was not found to show choroidal dilatation.24
Another recent report analysed the en face SS-OCT images of chronic CSC and reported that a diffuse choroidal dilatation pattern in Haller's layer was more frequently observed than a focal pattern.19 This result was consistent with ours, although there were some differences. First, we enrolled both patients with acute and chronic CSC and compared their en face image patterns. Second, the previous study reported dilatation of choroidal vessels found in all choroidal layers, including choriocapillaris and Sattler's layer. However, in the present study, we observed the vessel dilatation to be originated from Haller's layer. The inner choroidal layer including the choriocapillaris and Sattler's layer was attenuated or compressed by the dilated outer choroidal layer, and these findings were observed on cross-sectional B-scans. The previous study reported that the dilated choroidal vessels visible on the en face image section of choriocapillaris or Sattler's layer to be the dilated vessels from the inner choroidal layer itself.19 From our study, however, we noted that the dilated vessels visible on inner choroidal en face images seemed to be from Haller's layer. Because of the inner choroidal obscuration, the dilated vessels derived from outer layer were observed in the inner en face section (figure 2).
The concept of ‘inner choroidal obscuration’ in CSC has been introduced recently, and studies have reported that the hyporeflective reflex extends from the outer choroidal layer and obscures parts of the inner layer corresponding to choriocapillaris and Sattler's layer in patients with CSC using SS-OCT, and this might be associated with exudative changes in the choroid caused by CSC.25 Another group reported thinning of the inner choroidal layers and enlargement of the underlying hyporeflective lumina in the hyperfluorescent ICGA area of patients with CSC using EDI-OCT.26 We noted similar findings using SS-OCT cross-sectional B-scan.
Recently, Spaide and Ryan27 described the loculation of fluid in the posterior choroid in eyes with CSC. They introduced the images from EDI-OCT, SS-OCT and en face OCT visualising the fluid behind the large-sized choroidal vessels. As the imaging modality is evolving, further evaluation imaging of deeper portion of choroid would be anticipated.
The precise mechanisms of inner choroidal obscuration are unclear, and there is no evidence of this in pathological studies from patients with CSC whose eyes have been examined at autopsy. However, we hypothesised that inner obscurations are the result of the dilated outer choroid compressing the inner choroid. As mentioned above, the en face images at the choriocapillaris and Sattler's layer are the direct consequence of inner choroidal obscuration. These findings might demonstrate how to differentiate between normally dilated choroidal vessels such as the vortex vein and pathologically dilated choroidal vessels like those in CSC. Inner choroidal obscuration was not found in the vortex vein, and normal inner choroidal architecture was visible with focally enlarged Haller's choroidal vessels. However, in patients with CSC, dilated Haller's choroidal vessels were visible and combined with inner choroidal obscuration.
The choroidal vascular dilatation locations confirmed with en face SS-OCT were correlated with the thickened choroidal area in the CT map (figures 3⇓–5). This result was partially consistent with a previous report that showed a high correlation between angiography pattern and thickened area of the CT map.28 However, our image showed choroidal vessel dilatation that corresponded to the CT map, without using an invasive imaging modality.
En face SS-OCT images are advantageous for evaluating choroidal vasculature. First, with a faster scan rate and longer wavelength than conventional SD-OCT, SS-OCT can provide higher resolution images, especially those of choroidal vascular changes beneath the RPE. Second, the software developed by Topcon enables en face SS-OCT images to be constructed and provides transversal slice images at arbitrary depths. This can provide more realistic visualisations of choroid at each level, and in vivo dye-free choroidal vascular imaging. Third, en face imaging provides an overview of the whole choroid and precise information (such as location, description, and size) that is not available in standard cross-sectional B-scans. En face imaging is also non-invasive and can be applied to screen the entire lesion.
However, an en face image is not an actual image of direct visualisation of choroid, but is a reconstructed image. Without the guidance of other imaging modalities such as ICGA, it would be limited to determine the presence of choroidal vessel dilatation by using only en face SS-OCT. In addition, image production can be time consuming. Although en face SS-OCT cannot replace ICGA, it can be used as an adjunctive imaging modality for CSC choroidal vascular evaluation.
Recently, OCT angiography using software called ‘split-spectrum amplitude-decorrelation angiography’ allows for the visualisation of both the retinal and choroidal vasculature without dye injection, and several studies investigated the choroidal vascular changes in the eyes with CSC.29–31 Unfortunately, the patients did not undergo OCT angiography evaluation in this study. Further studies comparing or co-analysing en face SS-OCT with OCT angiography would be desirable.
The current study had several limitations. One was the retrospective case series nature with a limited number of patients. Another limitation was that the axial resolution of the SS-OCT is 8 μm and the lateral resolution is 20 μm, whereas the choriocapillaris layer thickness is about 10 μm. Therefore, although en face OCT visualised medium-sized and large-sized choroidal vessels, the assessment of the choriocapillaris could be challenging.21 The CSC stage was determined by the onset of subjective symptoms; therefore, there might be controversy over the exact duration of symptoms.10 Also, we especially focused on the imaging analysis itself after simply dichotomising the CSC as acute and chronic. The concept of recurrence or sequelae was not taken into consideration for the sake of simplicity. This aspect probably resulted in the unbalanced sample in our study. Furthermore, there might be controversy in the differentiation of focal and diffuse choroidal dilatation patterns due to the lack of objective criteria.19 ,22 ,23 To minimise the subjectivity bias, three of the authors analysed the en face SS-OCT images in a masked manner and the majority of author's results were applied. It would have been ideal to have matched eyes as a reference point or have a large reference group. Furthermore, measuring the actual vessel diameters and comparing them qualitatively would yield an analysable data.23 In further studies, it would be desirable to have more delicate measurement methods of the vessel diameters in en face imaging and an improved study design. More objective grading scheme should be needed for future study.
To our knowledge, this is the first report about the distinct patterns of choroidal vascular changes in eyes with CSC according to chronicity based on en face SS-OCT. We observed that focal choroidal vascular dilatation was common in acute CSC, whereas diffuse choroidal vascular dilatation was common in chronic CSC. Inner choroid obscuration was observed in both acute and chronic CSC. En face imaging using SS-OCT can be a reproducible, non-invasive and effective tool to analyse choroidal changes in acute and chronic CSC. In vivo documentation of the choroidal vasculature at variable depths using en face SS-OCT in patients with CSC will help understand the pathophysiology of CSC and can contribute to the diagnosis and treatment of CSC.
Correction notice This article has been corrected since it was published Online First. A second affiliation has been added for author Won June Lee.
Contributors WJL and BRL: concept and design; WJL, BRL, JWL and SHP: analysis and interpretation; WJL and JWL: writing the article; WJL, BRL, JWL and SHP: critical revision of the article; WJL, JWL and SHP: data collection; WJL, BRL, JWL and SHP: provision of materials, patients or resources; WJL and JWL: statistical expertise; and WJL, JWL and SHP: literature search.
Competing interests BRL is a consultant for Nidek, Gamagori, Japan.
Ethics approval Hanyang University Medical Center.
Provenance and peer review Not commissioned; externally peer reviewed.