Remodeling of macular vortex veins in pachychoroid neovasculopathy

Superior and inferior macular vortex veins are divided by a horizontal watershed passing through the macula. We evaluated macular vortex vein remodeling in eyes with pachychoroid neovasculopathy (PNV). Thirty eyes of 30 patients with treatment-naïve PNV and 30 normal eyes of 30 age-, gender-, and refraction-matched subjects were studied. We assessed the features of macular vortex veins employing en face optical coherence tomography (OCT) and determined central choroidal thickness (CCT) using B-mode OCT. Of the 30 normal eyes, a horizontal watershed was identified in 24 eyes (80%), while venous anastomosis between the superior and inferior vortex veins was observed in 6 eyes (20%). Mean CCT was 233 μm. Of the 30 eyes with PNV, vortex veins were dilated in all 30 eyes with PNV. In 27 of the 30 PNV eyes (90%), the horizontal watershed had disappeared, and collateral veins had instead developed via anastomosis between the superior and inferior vortex veins, making this finding significantly more frequent than in normal eyes (P < 0.001). Mean CCT was 357 μm, significantly thicker than that of normal eyes (P < 0.001). Remodeling of choroidal drainage routes by venous anastomosis between superior and inferior vortex veins was common in eyes with PNV. This observation suggests longstanding congestion of the choroidal veins.


Results
The demographic and clinical characteristics of our PNV patients and normal subjects are presented in Table 1. The normal subjects included 20 men (66.7%) and 10 women (33.3%). The age of the normal subjects was 63.1 ± 9.2 years (mean ± standard deviation). All eyes were phakic with a refraction of −0.47 ± 2.17 diopters. The PNV patients in this study included 26 men (86.7%) and four women (13.3%). Their age was 64.1 ± 11.6 years. Twenty-six eyes (86.7%) were phakic and four eyes (13.3%) had pseudophakia. The refraction of the 26 phakic eyes was −0.21 ± 1.79 diopters. Best corrected visual acuity was 0.23 ± 0.28 logMAR units.
The en face OCT showed dilated vortex veins (outer choroidal vessels) in all 30 PNV eyes. Fifteen eyes (50%) showed a symmetrical distribution of the superior and inferior vortex veins. The superior and inferior vortex veins were both dilated in the posterior fundus. The other 15 eyes (50%) showed an asymmetrical vortex vein distribution, 8 eyes (26.7%) with superior dominant and 7 eyes (23.3%) with inferior dominant findings. Dominant vortex veins showed somewhat greater dilatation. Twenty-seven eyes (90%) showed anastomosis between the superior and inferior vortex veins at the horizontal watershed (Figs 2-4), making this finding significantly more frequent than in normal eyes (P < 0.001). In 6 of the 30 PNV eyes (20%), the anastomosis was observed mainly in the peripapillary area (Fig. 3). The vortex vein anastomosis showed sinusoid like dilatation in 2 eyes (6.7%) (Fig. 4). CCT was 357 ± 100 µm, significantly thicker than that of normal eyes (P < 0.001). ICGA confirmed perfusion in the collateral vessels which had developed from anastomosis. OCT angiography demonstrated network vessels of CNV between the detached retinal pigment epithelium (RPE) and Bruch's membrane, which consistently arose at the site of dilated vortex veins in all 30 eyes with PNV (Figs 2-4).

Discussion
We conducted a retrospective investigation of 30 eyes with treatment-naïve PNV using multimodal imaging. En face OCT revealed an anastomosis between the superior and inferior vortex veins in 27 eyes (90%) with PNV, from which the horizontal watershed zone had disappeared (Figs 2-4). Sub-RPE CNV had consistently arisen at the site of dilated vortex veins in all 30 PNV eyes. In 24 of the 30 normal eyes (80%), as no anastomosis was present, the horizontal watershed zone was identified between the superior and inferior vortex veins (Fig. 1).
Collateral formation due to venous anastomosis at the watershed appears to be a common response to choroidal vortex vein congestion. Takahashi and Kishi evaluated choroidal drainage routes by applying wide-angle  ICGA in eyes with occluded vortex veins after scleral buckling surgery for retinal detachment 15 . In 10 of the 12 eyes that underwent angiography 3 months or more postoperatively, the collateral veins that developed connected the sector harboring the occluded vortex veins to that with intact vortex veins via venovenous anastomoses 15 . The collateral veins which compensate for the occlusion of vortex veins also formed as a consequence of radiation retinochoroidopathy 14  A recent study focusing on the choroidal features of peripapillary PCV found that eyes with this type of PCV had thinner subfoveal choroids than those with macular PCV 17 . However, at the foci of extrafoveal disease, choroidal thickness, Haller's layer thickness, and its ratio to total choroidal thickness were all relatively high, findings www.nature.com/scientificreports www.nature.com/scientificreports/ consistent with those of macular PCV 17 . Peripapillary pachychoroid syndrome is a newly proposed variant of pachychoroid disease spectrum, which demonstrates thicker nasal versus temporal macular choroidal layers 18 . Peripapillary choroidal thickening is reportedly associated with nasal macular intraretinal and/or subretinal fluid accumulation and also with disk edema in some cases 18 . In our present study, 6 PNV cases (20%) showed anastomosis between the superior and inferior vortex veins mainly in the peripapillary area. Peripapillary anastomosis between the superior and inferior vortex veins might develop in peripapillary PCV as well as in peripapillary pachychoroid syndrome. The peripapillary watershed zone appears to be another preferential site, besides the macula, of anastomosis development between the superior and inferior vortex veins.
Lee and colleagues described attenuation and thinning of the choriocapillaris and Sattler vessels overlying dilated outer choroidal vessels (pachyvessels), based on their investigation evaluating B-mode OCT features in eyes with PCV 4 . They suggested the following mechanism to underlie the development of CNV in eyes with PCV: Attenuation of the choriocapillaris produces a relatively ischemic environment at the RPE-Bruch membrane complex level, triggering expressions of various angiogenic factors 4 . We previously reported a choriocapillaris filling delay in CSC in the area of dilated vortex veins. In the current study, dilated vortex veins were consistently observed beneath the CNV in eyes with PNV. Therefore, outer choroidal vessel dilatation associated with chronic choriocapillaris ischemia might lead to the development of CNV in eyes with PNV. www.nature.com/scientificreports www.nature.com/scientificreports/ The limitations of our study include its retrospective nature, the single-center design, and the relatively small number of patients. OCT focused on the posterior pole of the fundus demonstrates only the posterior portion of the choroidal circulation. Measurements of CCT were performed manually. The current subjects were all Japanese, such that the results may not be generalizable to a larger PNV population, including other racial groups.
In conclusion, remodeling of choroidal drainage routes via venous anastomosis between the superior and inferior vortex veins commonly develops in eyes with PNV, suggesting longstanding congestion of the choroidal veins.

Methods
We performed this study, in compliance with the Declaration of Helsinki guidelines, after obtaining approval from the Institutional Review Board of Gunma University Hospital. Informed consent was obtained from all individual participants included in the present study.
The DRI OCT-1 Triton and PLEX Elite 9000 devices incorporate a tunable laser with a 1050 nm central wavelength and acquire 100,000 A-scans/second. The DRI OCT-1 Triton has an axial resolution of 8 μm and a lateral resolution of 20 μm, while the PLEX Elite 9000 has an axial resolution of 6.5 μm and a lateral resolution of 20 μm. We obtained B-mode images of the horizontal and vertical line scans (12 mm) through the fovea employing the DRI OCT-1 Triton. Next, cube data were obtained with a raster scan protocol of 1024 (horizontal) × 1024 (vertical) B-scans, which covered the 12 × 12 mm area centered on the fovea by the PLEX Elite 9000. En face images were obtained from the vitreous to the choroidoscleral border with coronal slices from a 3-dimensional dataset included in the inner software. Then, we performed OCT angiography volume scanning, i.e. 300 × 300 pixels in the 3 × 3 mm area demonstrated by the PLEX Elite 9000. OCT angiography is based on optical microangiography algorithm.
All normal eyes underwent color fundus photography and SS-OCT to obtain B-mode and en face structural images.
Definition of pachychoroid neovasculopathy. Herein, we diagnosed PNV if all of the following criteria were met. (1) CNV was diagnosed or suspected by FA and ICGA in the affected eye. (2) A shallow irregular RPE detachment at the site of CNV was observed on OCT B-mode images, while network vessels of CNV were detected between the detached RPE and Bruch's membrane on OCT angiography 19 . (3) Clinical and anatomical features of the pachychoroid phenotype were present, i.e. pathologically dilated outer choroidal vessels on en face OCT images and regional choroidal vascular hyperpermeability on ICGA images. CCT was not included among the criteria for the pachychoroid phenotype because CCT is impacted both by age and refractive errors 20 . Furthermore, eyes with normal CCT can exhibit extrafoveal choroidal thickening at sites of CNV 4 . The absence of drusen was also excluded from among the criteria applied for diagnosing the pachychoroid phenotype because pachychoroid-associated drusen, so-called pachydrusen, can be seen in pachychoroid spectrum diseases 21,22 . CNV with polypoidal lesions, so-called polypoidal choroidal vasculopathy, was excluded from the PNV category. image analysis. We evaluated the remodeling of macular vortex veins; the presence of anastomosis between superior and inferior vortex veins in eyes with PNV was demonstrated using en face OCT and ICGA. The presence of dilated vortex veins, choroidal vascular hyperpermeability, and anatomical and functional anastomoses between the superior and inferior vortex veins were judged by two experienced retinal specialists (H. Matsumoto and S. Kishi), working independently of each other. CCT was measured on B-mode images using the computer-based caliper measurement tool included in the OCT system. CCT was defined as the distance between Bruch's membrane and the margin of the choroid and sclera under the fovea.
Statistical analysis. For statistical analyses, the Mann-Whitney U test was used to compare unpaired values of age, refraction, and CCT. The chi-squared independence test was used to determine differences in gender, vortex vein symmetry, and the incidence of vortex vein anastomosis. The data analyses were performed using Excel 2016 (Microsoft, Redmond, WA, USA) with add-in software Statcel4 23 . A P < 0.05 was considered to indicate a statistically significant difference.