Purpose To compare the findings of optical coherence tomography angiography (OCTA) with indocyanine green angiography (ICGA) in polypoidal choroidal vasculopathy (PCV) that was divided into two types: polypoidal choroidal neovascularisation (CNV) and typical PCV (type 2 PCV).
Methods We studied a retrospective case series of 32 patients with treatment-naïve PCV (24 men, eight women; mean age 65.4 years). PCV was categorised into polypoidal CNV (type 1 PCV) and type 2 PCV based on ICGA findings. OCTA was performed using the RTVue XR Avanti. Macular cubes (3×3 or 6×6 mm) were acquired. To evaluate the locations of polyps and branched vessel networks (BVNs), we used B-mode scan.
Results OCTA clearly depicted only 17% of the type 1 PCV polyps and 46% of the type 2 PCV polyps which were detectable by ICGA. All type 1 PCV polyps detectable by OCTA were located just beneath the retinal pigment epithelium (RPE). On the other hand, type 2 PCV polyps were detected in various locations. All BVNs of type 1 PCV were located between the RPE and Bruch's membrane on OCTA images. However, the BVNs in type 2 PCV were located mainly under the RPE, though some were located in the choroid.
Conclusions Polyps of type 1 PCV were more difficult to detect with OCTA than those of type 2 PCV. Polyps of type 1 PCV were located just beneath the RPE. The BVNs of type 1 PCV were located between the RPE and Bruch's membrane.
Statistics from Altmetric.com
Polypoidal choroidal vasculopathy (PCV) is a subtype of neovascular age-related macular degeneration ((NV-AMD) and is characterised by a branching vascular network (BVN) with polypoidal lesions detectable by indocyanine green angiography (ICGA).1 ,2 There has been controversy regarding the origin and location of both the BVNs and the polypoidal lesions. Our group categorised PCV into two types: polypoidal choroidal neovascularisation (CNV) (type 1 PCV) and typical PCV (type 2 PCV). Type 1 PCV was shown to be located in the sub-retinal pigment epithelium (RPE) space, whereas type 2 PCV was shown to be located in the choroidal space.3 Coscas et al4 recently reported two forms of PCV, that is, idiopathic PCV and secondary polyps associated with NV-AMD. These reports may support the division of PCV into type 1 PCV and type 2 PCV. There are, however, reports showing all PCV lesions located in the sub-RPE space to have the features of CNV.5 ,6 The controversy is due to ICGA not having the capacity to visualise three dimensionally the structure of a polypoidal lesion. Optical coherence tomography angiography (OCTA), a new, non-invasive imaging technique recently became available for clinical use and might be the key to resolve the controversies faced in studying PCV. OCTA can acquire volumetric angiographic information without dye. The en-face images allow visualisation of individual vascular plexus, the inner retina, the outer retina and the choriocapillaris without being impacted by dye leakage.
The present study aimed to evaluate PCV types by employing OCTA and comparing the findings obtained with those of ICGA.
Subjects and methods
Thirty-two patients (24 men, eight women; mean age 65.44 years) were enrolled at Nihon University Hospital in Tokyo between June 2015 and October 2015. PCV was diagnosed based on the presence of polypoidal lesions on ICGA. Then, we categorised each PCV into either polypoidal CNV (type 1 PCV) or typical PCV (type 2 PCV), based on ICGA, as previously reported.3
The diagnosis and categorising were ascertained by two retinal specialists (RM and KT). If the interpretation differed between the two readers, another retinal specialist (AK) interpreted the results. Central macular thickness and central choroidal thickness (CCT) were measured by enhanced depth imaging (EDI)-optical coherence tomography (OCT) (SPECTRALIS HRA–OCT; Heidelberg Engineering, Heidelberg, Germany). EDI-OCT images were obtained using the Spectralis SD-OCT unit. A single 30° horizontal line scan (approximately 9 mm), centred on the fovea, taken in high-speed EDI mode (768 A-mode scans per B-mode scan) was selected for analysis. Using the eye-tracking automatic real-time mode in Spectralis, about 100 B-scans were averaged per image.7 OCTA was performed using the RTVue XR Avanti with Angio Vue (Optovue, Fremont, California, USA). This instrument has an A-scan rate of 70 000 scans per second, using a light source centred on 840 nm and a bandwidth of 50 nm. Each OCTA volume contains 304×304 A-scans with two consecutive B-scans captured at each fixed position before proceeding to the next sampling location. Split-spectrum amplitude-decorrelation angiography was used to extract the OCTA information. Each OCTA volume is acquired in 3 s and two orthogonal OCTA volumes were acquired in order to perform motion correction to minimise motion artefacts arising from microsaccades and fixation changes. Macular cubes (3×3 or 6×6 mm) were acquired. To check for retinal haemorrhage or pigment epithelial detachment (PED), colour fundus photography and OCT were performed. Before any treatment was administered, we performed ICGA, OCTA and then compared the images to detect any differences. To evaluate polyp and BVN locations, we used B-mode scan. First, we automatically set the layer to the choriocapillaris on OCTA. Then, we manually operated the moving of the scan line to detect polyps and BVNs. The locations were ascertained by two retinal specialists (RM and KT). If the interpretation differed between the two readers, another retinal specialist (AK) interpreted the results. Furthermore, polyp pulsation was evaluated by making an ICGA video. We then determined whether the polyps showing pulsation on ICGA were detectable on OCTA.
Informed consent was obtained from each study participant as per the protocol approved by the Institutional Review Board of Nihon University. This investigation was performed according to the guidelines of the Declaration of Helsinki.
Data are presented as means±SD. The Mann-Whitney U test and Fisher’s exact test were performed to determine the significance of differences between type 1 and type 2 PCVs. Statistical significance was set at p<0.05. SPSS (V.21; SPSS, Chicago, Illinois, USA) was used for these analyses.
The clinical features of participants and the results of the comparison of type 1 and type 2 PCVs are shown in table 1.
The mean CCT of type 2 PCV was significantly greater than that of type 1 PCV. The maximum BVN dimension of type 1 PCV was significantly larger than that of type 2 PCV. Furthermore, there were significant differences in polyp numbers between type 1 and type 2 PCVs. Though 4.7 polyps on average per eye were clearly detectable on ICGA, only 0.8 polyps (17%) on average were seen on OCTA of type 1 PCV. When examining type 2 PCV, 1.6 polyps were detectable on ICGA and 0.77 polyps (46%) were detectable on OCTA. Nine of 22 eyes with type 2 PCV had no detectable polyps on OCTA. All nine of these eyes (100%) with type 2 PCV, in which no polyps were seen on OCTA, had haemorrhagic PED (hPED), PED or subretinal haemorrhage (SRH). There were three type 1 PCV cases in which polyps were not detectable on OCTA, but only one (33%) had hPED (data not shown).
Representative cases are shown in figure 1 for type 1 PCV and in figure 2 for type 2 PCV. Type 1 PCV has a feeder vessel in the early phase on ICGA. Type 2 PCV has no evidence of a feeder vessel on ICGA.
There were three polyp locations which could be identified by B-mode scan, that is, the outer retina, just beneath the RPE, and the choroid. Examples of each are shown in figure 3. Polyp locations detectable on OCTA are presented in table 2.
The type 1 PCV polyp seen on OCTA was located entirely beneath the RPE. On the other hand, type 2 PCV polyp locations were variable. In 22 eyes with type 2 PCV, 36 polyps were detected on ICGA. However, 17 polyps (47%) were detected on OCTA in these 22 eyes. A total of 1 (6%), 3 (18%) and 13 (76%) eyes had polyps located in the outer retina, the choroid and just beneath RPE (figure 3), respectively. BVN locations are also shown in table 2. All BVNs in eyes with type 1 PCV were located between the RPE and Bruch's membrane on OCTA. However, the BVN in eyes with type 2 PCV were located mainly under the RPE, but there were also cases with choroidal locations. Examples of each are presented in figures 1 and 2.
Polyp pulsation was visualised by ICGA in 50% of eyes with type 2 PCV and 10% of those with type 1 PCV. Eight (75%) of these 12 pulsating polyps were detectable by OCTA.
First, we used both 3×3 and 6×6 mm for this study. As Spaide et al8 noted, 3×3 mm is definitely better for detecting lesions on OCTA. However, in our cases, especially those with type 1 PCV, it was difficult to evaluate entire lesions due to their larger dimensions. We consequently used 6×6 mm for this study.
In this study, we examined the OCTA findings of two different types of PCV, types 1 and 2. We previously reported the distinct genetic backgrounds of these two types of PCV.9 ,10 Coscas et al4 recently reported two forms of PCVs, that is, idiopathic PCV (IPCV) and secondary polyps associated with NV-AMD. Comparing with our subtype of PCV, their categorisation had same feature. Their ‘IPCV’ is similar to our type 2 PCV, in terms of mean BVN diameter which is smaller and mean CCT which is larger than the other subtype. Furthermore, they stated that BVN had different feature between IPCV and NV-AMD categorising by leakage of fluorescein angiography. As mentioned before, Inoue et al11 also divided PCV into two forms, PCV and polypoidal choroidal NV (PCNV). Their categorisation was similar with Coscas et al,4 that is, CCT of PCV is larger than PCNV and large neovascular complex in PCNV. Although the existence of feeder vessel was our key to separate PCV, they did not state it. However, the big difference between our and their categorisation is that the location of polyp and BVN (or CNV). Nevertheless, these reports may support the division of PCV into two forms.
In our present study, there were significant differences between the polyps of type 1 and type 2 PCVs, as seen on OCTA. There are two possible reasons for these differences. The first would be a segmentation error occurring just beneath the RPE. Polyps are often located just under the RPE and affected eyes also tend to have PED. These PEDs result in a steep slope making it difficult to place the sliced line at a relevant depth. Also, hPED, PED or SRH would be the factors for the differences. In our study, 100% (nine cases) with type 2 PCV, in which no polyps were seen on OCTA, had hPED, PED or SRH. However, only one case (33%) with type 1 PCV had PED undetectable polyp on OCTA. Type 2 PCV might be difficult to detect polyps when there were hPED, PED or SRH.
Inoue et al11 reported the characteristics of PCV, divided into a form similar with our type 2 PCV and polypoidal CNV similar with our type 1 PCV. They noted that polyps were visualised on OCTA in only 43% of the eyes in their study. Kim et al12 also reported that only 50% of polyps were hyper-reflective on OCTA. Srour et al13 recently reported that 25% of polypoidal lesions showed increased flow around lesion structures. Although they did not mention which type was more easily detected, we suspect that type 2 would have been more readily detectable. Inoue et al also noted that it was difficult to detect lesions in the presence of haemorrhage and PED.
Polyp locations are a source of controversy. We were not able to detect polyps in the z-axis direction on ICGA. However, using OCTA, locations could be visualised if we carefully considered the projection artefact. Recently, Miura et al,14 using Doppler OCT, reported 87% of polypoidal lesions to be located in the PED, 6.7% in the choroid and 6.7% in both the sub-RPE space and the choroid. We clearly showed polypoidal lesions of type 1 PCV to be located in the sub-RPE space, while those of type 2 PCV showed greater diversity, with 76% being located in the sub-RPE space, 18% in the choroid and 6% in the outer retina. Yuzawa noted that in some cases, a break in the RPE might cause the polypoidal lesion to be located in the outer retina space.15 Nakashizuka et al16 showed type 2 PCV specimens to have features of arteriosclerotic change. They noted that in type 2 PCV, polypoidal lesions are located under Bruch's membrane. These results may support the PCV origins suggested in our prior report. We found type 2 PCV to have the characteristic features of choroidal vascular abnormalities.
In our dataset, all BVNs of type 1 PCV were located between the RPE and Bruch’s membrane. Though BVNs of type 2 PCV were mainly located under the RPE, some were located in the choroid. Inoue et al11 also reported BVNs and polyps to be limited to sites between the RPE and Bruch’s membrane. However, Kim et al17 reported the colocalisation between ICGA and OCT of PCV. They stated that the large vessels of the BVN were located in the inner choroid in some cases. These reports support our results.
Polyp pulsation can be seen on ICGA videos and is a feature of type 2 PCV.3 Polyps showing pulsation were much more easily detected that those that did not. Pulsation of polyps indicates high flow, as illustrated by OCTA. This phenomenon might explain the high rate of polyp detection on OCTA. Further study is needed, as conclusions cannot be drawn from our small number of cases.
Pachychoroid neovasculopathy was recently reported.18 Pachychoroid neovasculopathy was described as type 1 CNV overlying a localised area of choroidal thickening and dilated choroidal vessels. In the report of Nakashizuka et al,16 they mentioned that the hyalinisation of the choroidal vessels in type 2 PCV would lead to the exudation of fibrin and blood plasma. The exudation would finally take to thick choroid. As choroidal thickness of type 2 PCV was more often thicker, we would say type 2 PCV was ‘Pachychoroid’, we speculate that dilated choroidal vessels may push up the polypoidal lesion such that the polyp is ultimately located under the RPE. Further study is needed to test our hypothesis.
We identified three major limitations of our study. The first was our small sample size. The second was the low detection of polypoidal lesion in OCTA. However, there were significant differences between subtypes with the detection of polyp even in these low rates. We will increase this number to show the fact in future studies. The third was the need for manual scanning. In order to detect the lesions, we needed to use manual scanning. However, due to its manual nature, our results may not be easily replicated. Further study is clearly needed. In conclusion, we found the polyps of type 1 PCV to be located just beneath the RPE, while those of type 2 PCV were detected in various locations. The BVNs of type 1 were located between the RPE and Bruch’s membrane. These findings support the existence of two types of PCV.
The authors thank all patients who participated in this study.
Funding This work was funded in part by the Division of Companion Diagnostics, Department of Pathology and Microbiology, Nihon University School of Medicine.
Contributors All persons who meet authorship criteria are listed as authors, and all authors certify that they have participated sufficiently in the work to take public responsibility for the content, including participation in the concept, design, analysis, writing or revision of the manuscript. Furthermore, each author certifies that this material or similar material has not been and will not be submitted to or published in any other publication before its appearance in the British Journal of Ophthalmology. Contributors—design and conduct of the study: KT, RM and MY. Collection, management and analysis: KT, RM, AK, YW and MY. Interpretation of the data: KT, RM and AK. Preparation and review of the manuscript: KT, RM, AK and HN.
Competing interests None declared.
Ethics approval Ethic committee of Nihon University Hospital, Tokyo, Japan.
Provenance and peer review Not commissioned; externally peer reviewed.
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.