Optical Coherence Tomography Angiography Characteristics of Polypoidal Lesions in Caucasians

Purpose The aim of the study is to analyze the swept source-optical coherence tomography angiography (SS-OCTA) characteristics of polypoidal lesions in Caucasian patients. Methods In this retrospective observational case series, 43 polypoidal lesions in 32 eyes of 32 patients were diagnosed using indocyanine green angiography (ICGA) and compared to SS-OCTA at a tertiary medical retina center (Clinic Landstraße, Vienna Healthcare Group, Austria) between June 2017 and March 2020. Vascularity was identified by color-coded B-scan SS-OCTA while morphology was described as revealed by en face SS-OCTA after alignment with ICGA-confirmed findings. Results In total, SS-OCTA detected all polypoidal lesions, as identified by ICGA. On B-scan SS-OCTA, circumscribed flow was detected in 33 (76.7%) polypoidal lesions and diffuse flow in 10 (23.3%) lesions. On en face SS-OCTA, polypoidal lesions appeared morphologically as 19 tangled vessel balls (44.2%), 6 tangled vessel balls next to dilated vessels (13.9%), 8 vascular dilatations (18.6%), and 8 ill-defined vascular networks (18.6%), leaving 2 lesions (4.6%) undetected. Circumscribed flow was significantly associated with tangled vessel balls (p = 0.005). Conclusion This study highlights the importance of a multimodal imaging approach, including SS-OCTA, for the evaluation of polypoidal lesions. Our findings suggest a morphological heterogeneity of vascular patterns in Caucasian patients with polypoidal lesions, as pictured by SS-OCTA.


Introduction
Polypoidal lesions were frst described as a distinct macular disorder characterized by nodular excrescences with interconnecting vascular channels more than 30 years ago [1]. Later on, two basic choroidal vascular changes were identifed with the help of indocyanine green angiography (ICGA): a branching neovascular network (BNVN) in the inner choroid and vascular dilations at the border of the vascular network [2]. Since then, ICGA has been considered the gold standard for the diagnostic evaluation of polypoidal lesions independent of their origin. More recently, optical coherence tomography (OCT) has become valuable as a dyeindependent diagnostic tool [3]. However, OCT is limited to structural fndings as it is unable to illustrate the vascular fow. Te era of OCT angiography (OCTA)-commencing with the spectral domain (SD) technology-enabled the visualization of diferent vascular patterns such as roundshaped or cluster-like structures attributed to polypoidal lesions in a noninvasive setting [4,5]. Te introduction of swept source OCT angiography (SS-OCTA) with a deeper penetration and lower sensitivity roll-of facilitated further insights into the morphology [6][7][8]. Recently described in an Asian population, B-scan SS-OCTA demonstrated a high diagnostic accuracy of specifc fow signals in polypoidal lesions inside or adjacent to pigment epithelial detachments (PED) [9]. In another study, polypoidal lesions appeared as tangled vascular structures in en face SS-OCTA [10]. In Caucasians, polypoidal lesions are most commonly associated with age-related macular degeneration (AMD) and the pachychoroid disease spectrum besides having an idiopathic appearance [11]. Te observed features were reported to be similar to those in Asians using multimodal imaging without OCTA [12].
In the light of the above, this study investigated the detection of ICGA-confrmed polypoidal lesions by combining B-scan SS-OCTA and en face SS-OCTA-based characteristics in a Caucasian patient cohort.

Study Population.
Tis retrospective observational case series included 32 eyes of 32 patients diagnosed with polypoidal lesions of diferent origins, independent of any prior treatment at our tertiary medical retina center (Medical Retina Unit, Department of Ophthalmology; Clinic Landstraße, Vienna Healthcare Group, Karl Landsteiner Institute for Retinal Research and Imaging, Austria) between June 2017 and March 2020. All patients underwent comprehensive ophthalmic examinations including dilated fundoscopy and multimodal imaging with SD-OCT, fuorescein angiography and ICGA (SPECTRALIS HRA-OCT Confocal Scanning Laser Ophthalmoscope and Angiography; Heidelberg Engineering, Heidelberg, Germany), SS-OCT, and SS-OCTA (DRI OCT Triton Plus; Topcon Corporation, Tokyo, Japan) as baseline standards of care on the day of admission. Te best corrected visual acuity (BCVA) was measured using the Early Treatment Diabetic Retinopathy Study (ETDRS) letter score (4 m) and converted to Snellen (Sn). Te SS-OCTA device works at a wavelength of 1050 nanometers under an acquisition rate of 100.000 Ascans per second with a motion contrast algorithm called OCTARA ™ [13]. It operates on 1 mW input power with a digital axial resolution of 2.6 µm and a transverse digital resolution range of 9.4 to 18.8 µm, depending on the selected cube. Two trained operators captured SS-OCTA B-scans and standardized 4.5 × 4.5 mm, 6 × 6 mm 2 , and 9 × 9 mm 2 en face SS-OCTA macular cubes with a scan resolution of 320 × 320 or 512 × 512 B-scans for each eye. Te study adhered to the tenets of the Declaration of Helsinki. All subjects provided informed consent to analyze their data retrospectively and, hence, to participate in the study at the frst presentation. Te Federal Hospitals Act §15a Abs. 3a states that approval from the Viennese ethics committee is not needed for this study design.

Imaging and Grading.
Polypoidal lesions and the BNVN were diagnosed based on the EVEREST criteria and termed as suggested by medical retina experts [3,14]. Two independent graders who were blinded to the angiographic fndings evaluated all acquired SS-OCTA images for fow detection in B-scans and the structural morphology in en face scans by using the integrated OCTA analysis software IMAGEnet 6 (Version 1.24.1.15742, Topcon Corporation, Tokyo, Japan). Te SS-OCTA grading approach was based on qualitative parameters for feasible reproducibility in clinical settings. In cases of grading disagreement, the images were independently reevaluated. If a grading disagreement existed, a senior clinical advisor was consulted. Te fnal grade was determined by the majority vote of all three retinal specialists. Te segmentation protocol defned the retinal pigment epithelium (RPE) as the inner line and Bruch's membrane (BM) as the outer segmentation boundary. In the case of segmentation artifacts, resegmentation was performed either semi-automatically by altering the predefned sections or by drawing the lines manually. First, SS-OCTA B-scans were investigated to identify fow inside various confgurated PEDs [15]. Color-coded fow signals were either found to be circumscribed with a focal increase of density existing in various shapes and forms within the PED (Figures 1 and 2) or difuse as scattered fow along the PED or as complete fow within the PED ( Figure 3). Circumscribed fow was defned by the presence of increased fow signals within a focal area inside the PED, while every other abnormal fow without precise localization was determined to be difuse.
Te corresponding en face SS-OCTA images were superimposed with the hypercyanescent spots in ICGA by an alignment of the inner retinal vasculature to confrm the topography of the respective structures. After locating the vascular complex as the lesion of interest correctly, a number of diferent en face SS-OCTA patterns were elaborated: First, the previously described tangled vessel ball type (Figure 1), a neovascularization with convoluted vessels in a ball-like formation. Second, a tangled vessel ball adjacent to dilated vessels inconsistent with the BNVN (Figure 2) was detected. Tird, single vascular dilatations were defned by the presence of larger vessel calibers exceeding the size of other pathologic vessel formations (Figures 3(a)-3(c)). Fourth, neovascular lesions without a tangled ball-shaped appearance or abnormal dilatations but unclear demarcation from the BNVN were termed as ill-defned vascular networks (Figures 3(d)-3(f )). More than 50% of the polypoidal lesion area had to be consistent with the respective morphological pattern for valid grading. On structural OCT B-scans, the PED height, confguration, and the localization of the polypoidal lesion inside the PED were analyzed as covariables. In addition, subfoveal choroidal thickness (SFCT) was measured as the greatest vertical distance between the BM and sclerochoroidal interface to diferentiate the pachychoroid.

Statistics.
Te sensitivity of each investigated SS-OCTA imaging modality was calculated in comparison to ICGA. Due to the small numbers, the diferentiation between scattered fow along the PED and complete fow within the PED was set in contrast to circumscribed fow on B-scan SS-OCTA. Te relationship between B-scan SS-OCTA and en face SS-OCTA was evaluated using the chi-square test. Age, sex, laterality, underlying disease, BCVA, SFCT, PED height, PED confguration, lesion localization inside the PED, presence of hemorrhage, and previous treatment with intravitreal antivascular endothelial growth factor (anti-VEGF) were investigated as potential infuencing factors using generalized mixed-efect regression analyses for the binary dependent variable B-scan SS-OCTA (comparing the probability for difuse fow vs. circumscribed fow) as well as for the binary dependent variable en face SS-OCTA (comparing the probability for tangled vessel balls vs. other results). All models were accounted for with random factors for patients. Statistical signifcance was set at p < 0.05. Intergrader variability was calculated using Cohen's kappa coefcient (K) and corresponding 95% confdence intervals (CI). All analyses were performed using R,

Results
In total, 43 polypoidal lesions were identifed by the ICGA and were hence eligible for enrollment. Te descriptive data and patient information summary are displayed in Tables 1  and 2.

Discussion
In this study, the diagnostic value of SS-OCTA in polypoidal lesions regarding functional characteristics of fow signals and morphology was investigated and set in contrast to the    Journal of Ophthalmology  invasive imaging modality ICGA. All lesions seen in ICGA were identifed using color-coded B-scan SS-OCTA. Fujita and colleagues primarily focused on the analysis of B-scan SS-OCTA and presented a detection rate of 94.7% with typical fow characteristics in Asians [9]. Tey described polypoidal lesions in B-scans as complete or incomplete round/ring-like or fow signals adjacent to a PED notch. While the authors emphasized on a detailed description of the distributed fow, this study attached less importance to the morphology of fow patterns in B-scan SS-OCTA.
Instead, fow density was described as an expression of functionality, either circumscribed by a focal area of high fow or difuse without exact localization within the PED. Interestingly, a statistical tendency for male preponderance and better BCVA was related to a difuse distribution, suggesting a heterogenic variability in the respective characteristics. Circumscribed fow in B-scan SS-OCTA was signifcantly related to the appearance of tangled vessels balls in en face SS-OCTA (Figure 4 and Table 3). In fact, all tangled    Journal of Ophthalmology vessel balls observed in this study were visualized as circumscribed fow in B-scan SS-OCTA, which was attributed to the densely packed multiple vascular branches in tangled vessel ball formations. In total, en face SS-OCTA was inferior to color-coded B-scan SS-OCTA as it was not able to visualize the morphology of all ICGA-confrmed spots. Tangled vascular structures with or without dilatations besides ill-defned vascular structures were observed. Tese fndings partly coincide with the observation of Bo and colleagues who frst reported that polypoidal lesions in ICGA correspond to tangled vessel balls using en face SS-OCTA [10]. In their Asian study population, all confrmed polypoidal lesions were pictured as tangled vascular structures by en-face SS-OCTA. Te authors concluded that polypoidal lesions are a distinct form of neovascularization rather than an aneurysmal structure. On the other hand, the BNVN was considered a variant of MNV type 1 and the term "aneurysmal" was found to be more appropriate than "polypoidal" to describe the vascular dilatations observed by multimodal imaging [16]. While retinal experts widely agree upon the vascular-some even refer to as arterial -origin rather than solid feshy polyps, consensus has not been reached whether polypoidal lesions are simple aneurysms or more complex vascular structures [17]. A Japanese study group detected tangled vascular structures in 58 of 72 (80.5%) polypoidal lesions using en-face SS-OCTA as a secondary outcome measure [9]. Nevertheless, they reported difculties in determining detailed structural fndings in the remaining polypoidal lesions, which were attributed to image resolution limits or low quality. In our cohort of Caucasian patients, a variety of morphological appearances besides tangled vessels balls were expressed such as vascular dilatations in connection with tangled vessels, mere fndings, and even ill-defned vascularity without a specifc pattern. Tis complexity suggests a more heterogenic structure of polypoidal lesions and could very well be the manifestation of the same disease in a diferent stage-treated or treatment-naive p.e.-or a distinct subtype. One hypothesis is based on the transformation of neovascular patterns following repeated intravitreal antiVEGF injections. Terefore, prior antiVEGF treatment was analyzed as an infuencing factor without signifcant fndings. Other reasons including higher luminal pressure, focal vessel wall weakness, and genetic variability could be responsible for the difering appearances. Te question remains as to whether dilated vessels are possibly aneurysmal neovascularizations as polypoidal lesions typically but not solely arise secondary to the pachychoroid disease spectrum. Te presence of hemorrhage had no impact on the appearance of polypoidal lesions. It is well known that polypoidal lesions pose a risk of extended subretinal hemorrhage as frequently seen in our study cohort (Figure 1(a)). ICGA, as a validated method, has a limited capability to detect polypoidal lesions underneath blood due to its attributes in the blood circulation. Te noninvasive SS-OCTA technique highlights blood fow as void signals below the RPE in B-scans but also as vascular structures in en face scans. Te herein detected relation between circumscribed high fow signals and tangled vessel balls could be useful to narrow the topographic localization of polypoidal lesions and might be preferential in the presence of subretinal hemorrhage, in the case of a strategical therapeutic change, for example. A prior anti-VEGF treatment was also investigated as a cofactor for the occurrence of morphologic patterns, but it showed neither a signifcant infuence nor a tendency at the time of observation. Te transformation of a polypoidal lesion into a MNV type 1 after previous antiVEGF treatment was not investigated in this study [18]. Structural OCT demonstrated a high diagnostic accuracy by applying a combination of 3 OCT-based criteria: sub-RPE ring-like lesions, en face OCT complex RPE elevation, and a sharp-peaked PED [3]. In our study, the identifed polypoidal lesions were set in relation to the PED height and PED formation as pictured by OCT B-scans. All polypoidal lesions were elucidated below the RPE in B-scan OCTA. Tey were either located within a sharp-peaked PED (Figure 3(c)), below the top of a PED (Figures 2(e), 2(f ), and 2(h)) or adjacent to a PED notch (Figures 1(d), 2(d), and 2(g)). However, the univariate analysis of coexisting OCT B-scan features showed no signifcance to either of the OCTA fndings. Other OCT-based criteria such as the subfoveal choroidal thickness of the afected and the contralateral eye were also investigated without signifcant relation. Attributes such as age or origin were separately investigated as cofactors without statistically signifcant results (Tables 2  and 3). Tis does not automatically imply that the diferent characteristics are independent of the etiologic nature of the underlying disease. It simply states that the herein applied covariables had no infuence on the appearance of polypoidal lesions in SS-OCTA.
Te limitations of this study are attributed to its retrospective nature. Te majority of polypoidal lesions were found to be solitary while only 6 eyes showed multiple lesions. Arguably, the cluster-like appearances in the same eye were likely to demonstrate similar SS-OCTA characteristics as described by Bo when in fact they showed a heterogenic picture as displayed in Figure 2 [10]. Nonetheless, no conclusion could be drawn as only two eyes were afected by this multiplicity of polypoidal lesions. Te consensual descriptions of morphological appearances were partially extracted from previous studies, which could lead to a grader's bias. Another limitation is the description of qualitative fow parameters in B-scans, namely, focal or difuse fow. Tis lack of diversity was partially driven by the low image resolution and the number of investigated eyes. However, good initial grader agreement underlines the predefned morphological expressions of fow signals and vascular patterns. To the best of our knowledge, this is the frst study investigating the feasibility of SS-OCTA in Caucasian patients diagnosed with polypoidal lesions by combining fow signals in B-scan OCTA and vascular patterns in en face SS-OCTA.
In conclusion, this study elucidates the value of SS-OCTA and contributes to the noninvasive ICGAindependent diagnostic characteristics of polypoidal lesions. Our fndings suggest a more diverse appearance of vascular patterns as a spectrum of the disease in Caucasians.

Data Availability
DAB and SAS had full access to all the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis. Te datasets used and/or analyzed to support the fndings of this study are available from the corresponding author on reasonable request.

Ethical Approval
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Te Federal Hospitals Act §15a Abs. 3a states that a vote from the Viennese ethics committee is not obligatory for this study design.

Consent
Informed consent was obtained from all participants. Publication is approved by all authors as well as the responsible authorities. Te statistical analysis was carried out by Alexandra Graf, an independent professional without afliations to our department or the institute.

Disclosure
Te abstract of the paper was presented at the annual meetings of the Austrian ophthalmological society in May 2022, Villach, Austria and at the 22nd EURETINA meeting in September 2022, Hamburg, Germany.