Aim To observe the choroidal microstructure in polypoidal choroidal vasculopathy (PCV) using high-penetration optical coherence tomography (HP-OCT) with a long-wavelength light source that visualises tissue beneath the retinal pigment epithelium (RPE) and deep choroid, and to compare the findings with those of indocyanine green angiography (ICGA).
Methods In this retrospective, non-invasive, observational case series, 19 eyes (18 patients) with PCV were observed using HP-OCT (swept source, 100 000 A-scans/s, 1060 nm wavelength) and ICGA. The HP-OCT scan protocol was a 3×3-mm or 6×6-mm square containing 256×256 or 512×128 A-scans. The choroidal thickness (CT) was measured using HP-OCT.
Results ICGA showed 43 polypoidal lesions in 14 eyes and a vascular network in 17 eyes. HP-OCT showed 41 of the 43 polypoidal lesions visualised by ICGA as RPE rings with inner reflectivity and 15 eyes with a vascular network. Six eyes with RPE rings with inner reflectivity on HP-OCT were not visualised on ICGA images. The choroidal vascular network was dilated in 14 (33%) of 43 polypoidal lesions and 22 (47%) of 47 polypoidal lesions on ICGA and HP-OCT images, respectively. The mean CT at the fovea was 250 μm. The CT at the dilated choroidal vessels beneath the polypoidal lesions was significantly (p = 0.0095) thicker than that of the undilated choroidal vessels beneath the polypoidal lesions.
Conclusions HP-OCT can visualise choroidal vascular abnormalities in eyes with PCV and should be useful for understanding the pathogenesis of these abnormalities.
Statistics from Altmetric.com
Polypoidal choroidal vasculopathy (PCV) is characterised by polypoidal dilatations and branching choroidal vascular networks observed on indocyanine green angiography (ICGA).1 ,2 Several investigators have suggested that PCV originates in an abnormality of the inner choroidal vessels1 ,3; however, the pathogenesis of PCV is not fully understood. Optical coherence tomography (OCT) is a non-invasive tool that provides retinal cross-sectional images and is essential for diagnosing various macular diseases,4–8 In eyes with PCV, OCT has shown protrusions of the retinal pigment epithelium (RPE) with moderate inner reflectivity and two highly reflective lines (‘double layer sign’). Ueno et al9 observed eyes with PCV before and after intravitreal ranibizumab (Lucentis, Genentech, Inc., South San Francisco, California, USA) injections using ICGA and spectral domain (SD)-OCT, which has higher axial resolution and faster acquisition time compared with time domain (TD)-OCT, and reported that some polypoidal lesions were visualised by SD-OCT but not by ICGA. Conventional SD-OCT with a 840 nm wavelength light source has a great deal of light scattering at the RPE, resulting in poor visualisation beneath the RPE. In contrast, high-penetration (HP)-OCT with a long-wavelength light source (1060 nm) can visualise the microstructures beneath the RPE and deep choroid. Custom-built software, a recent innovation, enables construction of en-face images of the posterior pole from the HP-OCT data. The en-face images facilitate screening of the entire lesion, which is difficult with conventional longitudinal sections. In the current study, we observed the PCV lesions, especially polypoidal lesions and vascular networks, and choroidal dilatation connected to the polypoidal lesions on the en-face HP-OCT images.
Patients and methods
We retrospectively reviewed the records of patients with PCV who were examined in the Department of Ophthalmology, Osaka University Medical School. All eyes underwent slit-lamp biomicroscopy, fluorescein angiography (FA) and ICGA at the initial visit. FA and ICGA were obtained using a confocal laser scanning system (HRA-2, Heidelberg Engineering, Heidelberg, Germany). The diagnosis of PCV was based on the findings of polyp-like choroidal vascular dilatation with or without a choroidal vascular network on ICGA images at the initial visit. The diagnosis was made independently by the authors who were retina specialists (KS, FG, MS, CH). The research adhered to the tenets of the Declaration of Helsinki, and all participants provided written informed consent. A trained technician performed all HP-OCT examinations through dilated pupils. The best-corrected visual acuity was measured and FA and ICGA images were obtained at the same visit. Patients were excluded who had severe cataracts, poor fixation, and posterior abnormalities other than PCV that prevented us from obtaining good-quality images. The institutional review board of Osaka University Hospital approved this retrospective study.
HP-OCT (Topcon Corporation, Tokyo, Japan) is based on swept-source OCT technology, with a scanning speed of 100 000 A-scans/s. The scan protocol was a 3×3-mm or 6×6-mm square containing 256×256 or 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 μ in tissue. The consecutive en-face images between the RPE and chorioscleral interface were constructed using the software co-developed with Topcon and were compared with the ICGA findings obtained on the same day as the HP-OCT images.
Choroidal thickness measurements
The choroidal thickness (CT) was measured manually at the fovea and the dilated and undilated choroidal vessels beneath the protruding segment of the RPE, using the scale included in the software. We identified dilated or undilated choroidal vessels beneath the protruding segment of the RPE by moving the en-face images back and forth, and the CT was measured on B-scan images in which the protruding segment of the RPE was seen (figure 1). The CT was defined as the distance from the RPE line to the hyperreflective line behind the large vessel layers of the choroid, presumed to be the chorioscleral interface. The CT measurements represented the average of all measurements performed by two authors (KS, FG).
Statistical analysis was performed using JMP software V.10.0 (SAS Inc., Cary, North Carolina, USA). The comparisons performed to identify significant differences in the CT were analysed using the t test. p<0.05 was considered significant.
Nineteen eyes of 18 patients (11 men, 7 women; mean age 74.0±7.1 years) were included. Six eyes were treatment naïve and 13 eyes had been treated with intravitreal injections of an anti-vascular endothelial growth factor (VEGF) drug and/or photodynamic therapy (PDT). In 5 (26%) of 19 eyes, all of which had been treated previously, ICGA did not show polyp-like choroidal vascular dilatation when ICGA and HP-OCT were performed simultaneously; however, the previous ICGA images showed that polypoidal lesions were present in all eyes (tables 1 and 2). ICGA showed 43 polypoidal lesions as clusters (17/43; 40%), masses (14/43; 33%), and looped vessels (12/43; 28%).
The en-face HP-OCT images showed 47 sites in 17 eyes with an RPE ring with inner reflectivity in the anterior slices obtained from the RPE layer that corresponded to the polypoidal lesions on ICGA images (figure 2). En-face HP-OCT images showed six RPE rings with inner reflectivity that were not seen on ICGA images (figure 2). The inner reflectivity within the RPE rings was divided into two patterns: mosaic (34/47; 72%) and homogenous (13/47; 28%). In 35 polypoidal lesions, the RPE rings with inner reflectivity were the same size as the polypoidal lesions on ICGA images. In eight polypoidal lesions, the RPE rings with inner reflectivity were larger than the polypoidal lesions on the ICGA images and the same size as the hypofluorescence around the polypoidal lesions on the ICGA images. In 8 (17%) of 47 sites, the hyporeflective dots within the RPE rings continued into the choriocapillary layer (figure 3).
ICGA clearly showed the choroidal vascular networks in 15 eyes; in 2 eyes the networks were blurry. In contrast, the en-face HP-OCT images showed the vascular networks in 15 eyes as a hyperreflective mesh-like configuration (figure 4). This type of configuration was seen clearly in 13 eyes, and in 10 eyes (67%), hyporeflective tortuous lines corresponding to choroidal vessels consisting of networks were clearly visualised within the hyperreflective mesh-like configuration (figure 4).
Dilated choroidal vessels beneath the polypoidal lesions were seen in 14 (33%) of 43 polypoidal lesions on the ICGA images (figure 1). A dilated choroidal vessel beneath the RPE ring with inner reflectivity (presumably polypoidal lesions) was seen in 22 (47%) of 47 RPE rings on the HP-OCT images. All dilated choroidal vessels seen on the en-face HP-OCT images coincided with those seen in the ICGA images.
The mean CT at the fovea was 250 μm (range 133–441 μ). The mean CT at the dilated choroidal vessels beneath the RPE ring at the fovea was 256.8±109.0 μm (range 124–534 μm), and the mean CT at the undilated choroidal vessels was 185.6±44.0 μm (range 96–265 μm) (table 3). The CT at the dilated choroidal vessels beneath the RPE ring was significantly (p = 0.0095) greater than that of the undilated choroidal vascular networks (figure 1).
ICGA is essential for diagnosing PCV. ICGA shows the polypoidal structures that correspond to the reddish-orange lesions visible on biomicroscopy and a branching vascular network in typical cases of PCV.1 ,2 Recently, OCT has been recognised as useful for diagnosing PCV.4–8 Iijima et al4 and Otsuji et al5 reported observing steep dome-like elevations of the RPE with moderate inner reflectivity, presumably consisting of polypoidal lesions as a hallmark of PCV. Using TD-OCT with an axial resolution of 10 μ, Sato et al6 reported the presence of double reflective layers at the lesions in the branching vascular networks, presumably composed of the RPE and Bruch's membrane (double layer sign). Tsujikawa et al7 observed polypoidal lesions at the margin of pigment epithelial detachments (PEDs) as a notch in the PED. Using SD-OCT with 3 μ axial resolution, Ojima et al8 reported that a highly reflective line, presumably Bruch's membrane, was beneath the PEDs in eyes with PCV.
In the current study, we observed PCV using HP-OCT with a longer wavelength of 1060 nm and 10 μ axial resolution. We believe that HP-OCT is useful for observing eyes with PCV, since PCV lesions, such as polyps and vascular networks, exist beneath the RPE. The software co-developed with Topcon enables construction of en-face HP-OCT images that provide transversal slice images at arbitrary depths. Two reports on HP-OCT imaging in eyes with PCV have been published.10 ,11 Yasuno et al10 observed a case of PCV and noted that HP-OCT showed inner reflectivity within the PED and the hyperreflective area on the en-face image corresponding to the polyps on ICGA. Hong et al11 reported that the newly developed high-penetration Doppler optical coherence angiography (HP-OCA) visualised the feeder vessels and provided three-dimensional images of the lesions, including polyps and the vascular network. The authors also reported that the en-face OCA images were similar to the ICGA images.11 As these reports described, HP-OCT is useful for observing eyes with PCV.
Previously, Kameda et al12 and Saito et al13 observed the RPE rings corresponding to the polypoidal lesions in 84.2% and 96% of cases, respectively, and choroidal vascular networks as elevations of RPE or a geographic area in 52.6% and 68% of cases, respectively, on en-face OCT images of PCV obtained by OCT ophthalmoscopy. In the current study, the en-face HP-OCT images showed that 95% of the polypoidal lesions on ICGA were RPE rings with inner reflectivity. All RPE rings accompanied internal mosaic (72%) or homogeneous (28%) signals, that is, internal signals that are rarely seen on en-face OCT-ophthalmoscopy images. The superior ability of en-face HP-OCT to observe polypoidal lesions likely is due to its higher axial resolution and superior ability to visualise the deep layers. In addition, the hyporeflective dots in the RPE rings extended into the choriocapillary layer in some cases. We presumed that the hyporeflective dots are the sites at which polyps from the vascular networks enter Bruch's membrane.
En-face HP-OCT but not ICGA visualised some polypoidal lesions. The precise mechanism of the discrepancy between ICGA and the en-face HP-OCT images is unclear, but one possibility is that some polypoidal lesions, especially after treatment, lack internal blood flow but maintain their configurations. Ueno et al9 reported that the discrepancy between the polypoidal lesions on ICGA and OCT occurred after ranibizumab treatment. The current study included 13 (68%) of 19 cases that underwent anti-VEGF therapy and/or PDT; therefore, it might be reasonable that only HP-OCT detected some polypoidal lesions that shrank but persisted after treatment.
Choroidal vascular networks were visualised on ICGA images in 15 (79%) eyes and on en-face HP-OCT images as a tortuous line within a hyperreflective mesh-like configuration in 15 (79%) eyes. Because the height of the RPE elevation (PED) is lower at the choroidal vascular networks than that of the polypoidal lesions, it was difficult to differentiate the vascular networks from irregular RPE lines due to drusen or a serous PED. However, en-face HP-OCT can provide two-dimensional retinal views of the extent of the lesions, so that it is possible to differentiate the vascular networks from other pathologies such as drusen. The benefit of en-face imaging is non-invasive screening of the entire lesion.
Several investigators have reported that PCV originates in an abnormality of the inner choroidal vessels, and histopathologic studies have shown dilated venules and arterioles under Bruch's membrane or fibrovascular tissue within Bruch's membrane.1 ,3 ,14 ,15 Using ICGA, Sasahara et al16 reported choroidal vascular hyperfluorescence in 9.8% of eyes with PCV and that all eyes with choroidal vascular hyperpermeability had dilated choroidal vascular lesions and 33% of them had a history of central serous chorioretinopathy (CSC). Koizumi et al17 reported that 34% of eyes with PCV had choroidal vascular hyperpermeability. Several investigators have reported an association between PCV and CSC and identified choroidal vascular hyperpermeability in late-phase ICGA images.18 ,19 In the current study, we found dilated choroidal vessels beneath the polypoidal lesions in 14 (33%) of 43 polypoidal lesions and 22 (47%) of 47 polypoidal lesions on ICGA and HP-OCT images, respectively. ICGA could not visualise some dilated choroidal vessels due to the overlapping of the abnormal vessels of the PCV lesions. The consecutive en-face HP-OCT scans detected the dilated choroidal vessels more easily compared with ICGA in 8 (42%) of 19 eyes.
Using enhanced-depth (EDI)-OCT, several investigators have reported that the CT in eyes with PCV was significantly greater than that in control eyes and those with age-related macular degeneration, and the subfoveal CT in eyes with PCV with vascular hyperpermeability was significantly thicker than in eyes without it.20–22 The mean subfoveal CT in the current study was 250±81.0 μ, which was lower than the values reported previously. Several studies have found that the CT changes after treatment, that is, notably the CT decreases after PDT.21–23 In addition, the CT decreased in the area of treatment and in other choroidal regions after PDT.24 In the current cases, 13 eyes received anti-VEGF therapy with or without PDT; therefore, we presumed that these therapies might affect the CT. However, the CT at the dilated choroidal vessels beneath the RPE rings was significantly greater than that at the undilated choroidal vessels. Using EDI-OCT, Yang et al25 measured the largest diameter of the choroidal hyporeflective lumina in the macular area as a surrogate for choroidal vessels and reported that the mean largest diameter of the choroidal vessels in the macular area was significantly larger than that of the control group and the CT was correlated significantly with the largest diameter of the choroidal vessels. The mechanism of the association of the choroidal vascular dilatation, that is, the thickened choroid, and PCV remains unclear. Several investigators have observed the choroidal circulation in PCV using ICGA or a laser speckle phenomenon; however, they did not mention the deep choroidal layers.3 ,26 Further investigations should be performed to monitor the circulation in PCV.
The current study had several limitations. One was that this report was a retrospective case series with a limited number of patients and a heterogeneous group of treated and treatment-naïve eyes. In addition, the presence or absence of treatments might affect the CT as we mentioned previously. Another limitation was that the axial resolution of the HP-OCT images is 8 μ and the lateral resolution is 20 μ, whereas, the choriocapillary layer thickness is about 10 μ. Therefore, en-face OCT visualises medium-sized and larger choroidal vessels, but assessment of the choriocapillaris may be challenging. We found that the hyporeflective dots within the RPE rings continued into the choriocapillary layer in 8 (17%) of 47 sites. The hyporeflective dots were seen as a continuous area of low reflectivity from Bruch's layer to the choroid. Therefore, we believe these hyporeflective dots were not an artefact but the site at which polyps from the vascular networks enter Bruch's membrane.
In summary, we observed the abnormalities of PCV and measured the CT at the fovea and dilated choroidal vessels using HP-OCT. This technology provides clear images of the polypoidal lesions, vascular networks, and choroidal vascular dilatations beneath the polypoidal lesions. The advantage of ICGA is that it provides useful information about leakage from the early to late phases, and the current accumulated knowledge about ICGA findings is greater than that of HP-OCT. The ability of en-face HP-OCT to detect abnormalities is almost the same as that of ICGA, although it is necessary to clarify the precise mechanism of the discrepancy between ICGA and HP-OCT. HP-OCT is advantageous in that it allows non-invasive screening of the entire lesion and choroid. In addition, the en-face HP-OCT images showed that some polypoidal lesions have a thicken choroid presumably due to the dilated choroidal vessels. En-face HP-OCT should be useful to clarify the morphology of PCV.
Competing interests None.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The authors have full access to all the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis as well as the decision to submit for publication.
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.