Novel OCT Angiography Features, von Hippel–Lindau Disease Association, and Genetic Characterization of Juxtapapillary Retinal Capillary Hemangiomas

Abstract

Purpose: To present new clinical features of juxtapapillary retinal capillary hemangiomas (JRCHs), assess the risk of von Hippel–Lindau (VHL) disease, and explore the genotype-phenotype correlations in patients with JRCH.

Methods: Fifty patients with JRCH were included. Multimodal retinal imaging including optical coherence tomography angiography (OCTA), visual acuity, presence of peripheral RCHs, affected lateralities, systemic evaluation for VHL disease, and underlying VHL variants were reviewed.

Results: Of 59 eyes, 48 had classic JRCHs, whereas 11 had atypical JRCHs (type B, if it broke through the inner limiting membrane: 3 eyes; type A, if not: 8 eyes). Compared with atypical type A, which was indolent, type B might warrant surgical interventions. Better final visual acuity (P < 0.0001), fewer peripheral RCHs (P = 0.02), and lower prevalence of large peripheral RCHs (>1.5 mm) (P = 0.027) were observed in eyes with atypical JRCHs than classic JRCHs. VHL was diagnosed clinically in 72% of patients, and 22 VHL variants were identified, including 5 novel variants. Patients with truncating variants had a higher prevalence of atypical JRCHs than those with single amino acid substitution/deletion variants (P = 0.009). Patients with bilateral VHL–JRCHs were more likely to have large peripheral RCHs (P = 0.02) and less likely to harbor β-domain single amino acid substitution/deletion variants (P = 0.066) than those with unilateral VHL–JRCHs.

Conclusions: Atypical JRCHs, with distinctive OCTA characteristics and favorable visual outcomes, are less complicated by peripheral RCHs and more relevant to truncating variant genotypes. JRCH monitoring should incorporate OCTA classification and genotype analysis.

Juxtapapillary retinal capillary hemangiomas (JRCHs) are vascular tumors arising within the peripapillary regions or on the optic nerve disc. Von Hippel–Lindau (VHL) disease is a rare, dominantly inherited syndrome that predisposes patients to multisystem benign and malignant tumors, including RCHs.¹ JRCHs are less commonly (approximately 15%) found in eyes affected by RCHs in patients with VHL disease.² In other cases, JRCHs occur sporadically. Typical JRCHs have been classified separately owing to their distinct appearance in natural history, difficulties in ablative treatments,³ and more severe visual impairments than peripheral RCHs.² Atypical juxtapapillary vascularized lesions have been reported rarely in patients with VHL disease. They are termed “vascularized glial veils” and “retinal vascular proliferation,”⁴ or, most recently, “flat RCHs.”⁵ However, given the atypical appearance of the juxtapapillary vascularized lesions, their prevalence is likely to be underestimated. 

Optical coherence tomography angiography (OCTA) is becoming an important part of multimodal imaging metrics for RCH monitoring, and it is regarded helpful in delineating atypical features of RCHs.⁵ However, the last overview of a large case series of JRCHs was published in 2000, when most of the clinical information was obtained through fundus photography and fluorescein angiography.⁶ The application of OCTA was largely limited to symptomatic JRCHs in case reports.^7,8 Therefore, the study aims to explore the OCTA features of different types of JRCHs, including asymptomatic occult JRCHs and/or atypical JRCHs. 

Previous large cross-sectional studies found no correlations between VHL genotype and various aspects of ocular phenotype, including the number of peripheral RCHs, laterality, and overall severity of angiomatosis.^9,10 Few conclusions could be made about the effects of genotype on JRCHs, except that the truncating type was associated with a lower prevalence of JRCHs.¹⁰ Nevertheless, the genotype–phenotype correlation in a longitudinal cohort of patients with JRCHs remains to be elucidated. 

Herein, we present a comprehensive review of Chinese patients with JRCHs, delineating the OCTA characteristics and longitudinal changes, especially for occult and atypical JRCHs. We determined the frequency of VHL disease in patients with JRCHs and investigated the genetic spectrum of JRCHs and its correlation with phenotype. 

Patients and Methods

This was a retrospective study of patients with JRCHs who visited the Eye and ENT Hospital of Fudan University between 2008 and 2024. The study was approved by the Institutional Review Board at the Eye and ENT Hospital of Fudan University, Shanghai, China and adhered to the Helsinki Declaration of 1975. Informed consent was obtained from the patients. JRCHs refers to an RCH on or within one disc diameter of the disc margin. Patients were further evaluated for family history and systemic involvement of VHL disease. The diagnosis of VHL disease was based on previously described clinical criteria.¹ Accordingly, patients with JRCHs were classified into VHL and non-VHL groups. Moreover, patients with VHL–JRCHs who did not have any JRCH in the contralateral eye until the most recent visit were categorized as the unilateral group. 

Ophthalmic Examinations

The patients underwent thorough ophthalmic examinations, including best-corrected visual acuity (BCVA), slit-lamp biomicroscopy, color fundus photography (Topcon TRC50LX; Topcon, Tokyo, Japan), ultra-wide field fundus photography (Optos 200Tx; Optos PLC, Dunfermline, UK), spectral-domain OCT (Spectralis OCT; Heidelberg Engineering, Heidelberg, Germany), fluorescein angiography (Topcon TRC50LX, Heidelberg Engineering, or Optos 200Tx), and swept-source OCTA (SS-OCTA) (PLEX Elite 9000; Carl Zeiss Meditec, Dublin, CA). The peripheral RCHs were evaluated for the number and size if available. A peripheral RCH of >1.5 mm (1 disc diameter) was regarded as a large RCH.¹¹ BCVA was converted into logMAR units for statistical analysis. For semiquantitative low vision scales, counting fingers, hand motion, light perception, and no light perception were specified as logMAR of 1.98, logMAR of 2.28, logMAR of 2.7, and logMAR of 3.0, respectively.^12,13 En face OCTA slabs were generated with automatic segmentation, and those deeper than superficial slabs were generated with automatic projection removal. 

Genetic Testing and Genotypic Categorizing

Targeted next-generation sequencing panel analysis, including VHL gene analysis (transcript reference: NM_000551), was performed. Whole-genome sequencing was performed if the panel analysis failed to detect potential disease-causing variants. In silico molecular genetic analysis was performed on all the available family members. After sequencing, the raw data was stored in FASTQ format and processed with cutadaptor software to remove MGI sequencing adapters and low-quality reads (<80 bp). The cleaned reads were then mapped to the UCSC hg19 human reference genome using Sentieon software with the BWA parameter. Duplicated reads were eliminated, and base corrections were made to enhance the accuracy of the final BAM file output. Single nucleotide polymorphisms and insertion–deletion variants were identified using Sentieon's driver parameter, and the data were converted to VCF format. Further annotation of variants was performed using ANNOVAR software, which also integrated information from databases like 1000 Genomes, ESP6500, dbSNP, EXAC, Inhouse (MyGenostics), HGMD, and predictions from SIFT, PolyPhen-2, MutationTaster, and GERP++. To select potential pathogenic mutations for downstream analysis, we applied four criteria: (1) mutation reads had to exceed five, with a mutation rate of at least 30%; (2) mutations were excluded if their frequency was over 5% in 1000 Genomes, ESP6500, or Inhouse databases; (3) mutations were dropped if present in the InNormal database (MyGenostics); and (4) synonymous mutations were removed unless listed in HGMD. The remaining mutations were considered potentially pathogenic for further analysis and were confirmed by Sanger sequencing. The pathogenicity of the VHL gene variants was evaluated according to the American College of Medical Genetics and Genomics standards. 

Based on the predicted effects on VHL protein, exon deletions, nonsense variants, and frameshift variants were categorized as the truncating variant (TV) group, and missense (substitution) and in-frame (deletion) variants were categorized as the single amino acid substitution/deletion variant (SAASDV) group.¹⁴ The VHL protein structurally contains α- (involving codons 155–192) and β-domains (involving codons 63–154 and 193–204).¹⁵ The SAASDV group was further subdivided into the α- and β-domains according to the location of the affected codon. 

Statistical Analyses

Statistical analyses were performed with SPSS version 20 (IBM Corp., Armonk, NY). The Shapiro–Wilk test was used to assess normality. Continuous data were presented as mean ± standard deviation or median with interquartile range (IQR). Categorical data were presented as percentage. Comparisons of continuous variables such as age were carried out with a t test or one-way analysis of variance, followed by Bonferroni analysis if necessary, whereas the number of peripheral RCHs, BCVA (logMAR), and the follow-up duration in different groups were compared using Mann–Whitney or Kruskal–Wallis tests. Comparisons of categorical variables such as JRCH subtype and genetic distribution were carried out with χ² or Fisher's exact tests, followed by χ² partition with Bonferroni correction for multiple testing if necessary. Differences were considered statistically significant at a P value of less than 0.05, except for the evaluation of differences in ages and JRCH subtypes between any two groups (TV, SAASDV, and non-VHL groups). In this case, the P value was adjusted for multiple comparisons for a threshold of 0.017 (0.050 ÷ 3.000). 

Results

Overall Clinical Characteristics and Subtypes of JRCHs

Seventy eyes of 59 patients who visited our hospital had JRCHs. Fifty-nine eyes of 50 patients from unrelated pedigrees and with follow-up for at least 3 months were included in the study. The mean patient age at first JRCH diagnosis was 34.1 ± 13.4 years. Twenty-two patients were male, and 28 patients were female. Nine patients had bilateral JRCHs. The median follow-up period was 39.5 months (IQR, 17.8–82 months). 

Based on the appearance in the fundus photograph and SS-OCTA scan, and in accordance with the characteristics described by Takahashi et al.,⁵ JRCHs were further categorized into classic and atypical types, which were identified in 48 and 11 eyes, respectively. Classic JRCHs refer to juxtapapillary nodular mass developing in any layer of the retina. Importantly, using multimodal imaging techniques especially SS-OCTA scan, occult classic JRCHs especially the exophytic/sessile type, could be easily identified (Figs. 1A–E) and longitudinally evaluated noninvasively (Supplementary Figs. S1A–S1H). Although some exophytic JRCHs were nearly invisible in the fundus photograph and en face OCTA image of the whole retina, their borders were best delineated at the avascular slab (Figs. 1C, S1C, S1G). 

Figure 1.

 

Identification of an occult classic JRCH with OCT angiography (OCTA). (A) Fundus photograph and (B) en face OCTA image of the total retina showing apparently normal optic disc and adjacent retina (case 1). (C) Avascular slab identifying a nodular JRCH (white arrow) on the superior side. (D) Corresponding B scan taken along the level of the yellow dashed line in (C) demonstrating blood flow in the exophytic JRCH (yellow arrow). (E) Fluorescein angiography image showing hyperfluoresence of the lesion.

By comparison, atypical JRCHs, which manifested as aberrant vascular proliferation, were in the superficial layer of the optic disc or adjacent retina. Among them, 73% were mainly located in the retinal nerve fiber layer, without breaking through the inner limiting membrane (ILM) (atypical type A). Type A was indolent (Figs. 2A–F) or slowly progressive (Figs. 2G–L). The rest fibrovascular lesions were classified as atypical type B. Type B broke through the ILM, adhered to and grew along the posterior hyaloid membrane (Figs. 3A–F), and warranted surgical intervention in some cases (Figs. 3G–L). 

Figure 2.

 

Atypical JRCHs without breaking through the ILM (atypical type A). (A) Fundus photograph showing a nearly invisible JRCH at the nasal disc border, which is hyperfluorescent on fluorescein angiography (B, case 3). (C) En face retinal OCTA identifying a tiny vascular focus (red arrow). (D) Along the yellow dashed line in (C), the B-scan verifying blood signal of the atypical JRCH (white arrow), in the superficial layer of the retina. Thirty-two months later, the atypical JRCH remains stable on the en face retinal OCTA image (E, red arrow) and corresponding B scan (F, white arrow). (G–I) Reddish lacy vascularity around bifurcation of the upper retinal vein is shown in fundus photograph (G), en face OCTA scan (H), and B scan (I, white arrows; OCT as shown in the inset, case 4). (J–L) The lesion is demonstrated to grow with increasing blood signal and shadowing effects (L, white asterisk) after 42 months of follow-up.

Figure 3.

 

Atypical JRCHs breaking through the ILM (atypical type B). (A) Fundus photograph showing a faint red lesion at the temporal disc border (white arrow, case 5). Along the yellow dashed line in en face retinal OCTA image (B), the atypical JRCH breaking through the ILM (C, white arrow) and attaching to the posterior hyaloid membrane (C, white asterisk). (D) The OCTA B-scan shows weak blood flow inside. (E, F) Twenty-one months later, the lesion grows along the posterior hyaloid membrane (E), exerting tractional effects on the retina (E, yellow arrow) with enhancement of blood flow signal (F). However, the BCVA remains 20/20. (G–I) En face retinal OCTA scans at baseline with BCVA of 20/30 (G, the red arrow indicates the lesion), thirteen months later with BCVA of 20/50 (H), and 4 months after the surgery with BCVA of 20/25 (I, case 6). (J–L) Corresponding B scans demonstrating the fibrovascular lesion adhering on the posterior hyaloid membrane (J, white asterisk) with lamellar macular hole; the formation of a full-thickness hole (K) and macular hole healing after surgical removal of the lesion (L).

Treatment options specific to JRCHs used in our study are summarized in Table 1. Eyes with classic JRCHs were treated with various options, including observation (22 eyes), pars plana vitrectomy (5 eyes), photodynamic therapy (10 eyes), oral propranolol (9 eyes), intravitreal anti-VEGF injection (3 eyes), and external beam radiotherapy (1 eye). Stabilization or improvement of visual acuity of the treated eyes was observed in 91%, 80%, 90%, 67%, 67%, and 0% for observation, pars plana vitrectomy, photodynamic therapy, oral propranolol, intravitreal anti-VEGF, and external beam radiotherapy, respectively. However, for eyes with atypical JRCHs, the management was mostly observation, except for one eye that underwent pars plana vitrectomy, and visual acuity was stable or improved in all eyes at the most recent visit. At the end of the follow-up, the median BCVAs of eyes with classic JRCHs and those with atypical JRCHs were 1.3 logMAR (IQR, 0.1–2.0 logMAR) and 0 logMAR (IQR, 0–0 logMAR), respectively. The difference was statistically significant (P < 0.0001) (Supplementary Fig. S2). 

Table 1.

 

Treatment Modalities Specific to JRCHs and Visual Outcomes

 

Association With VHL Disease, Genetic Profiles, and Genotype–Phenotype Correlation

Overall, VHL disease was present in 72% of patients with JRCHs, based on the clinical diagnostic criteria for VHL, and they were further included for VHL mutation analysis. As shown in Figure 4A, of the 36 patients diagnosed clinically with VHL disease, 97% were genetically positive for VHL mutations. A total of 22 germline VHL variants were detected, among which c.341-13_350del, c.453-454insA, c.484_490delinsCT, c.585_586del, and c.329A>C were novel (Table 2, Fig. 4B). 

Figure 4.

 

Overview of patients with JRCHs and follow-up information. (A) The pie chart illustrates the distribution of 50 patients with JRCHs according to the association with VHL disease and genetic results. (B) The spectrum of VHL variants along the length of the VHL gene among patients with VHL–JRCHs. The subtypes of JRCHs and affected lateralities are also marked for a certain VHL variant.

Table 2.

 

Detected Germline Variants of VHL Gene in Patients With VHL-related JRCHs

 

Patients with detected VHL variants were subdivided into the TV and SAASDV groups. Clinical characteristics of different groups are presented in Table 3. The average age of JRCH diagnosis for non-VHL patients (45.2 years) was significantly older than that of patients with TV (28.8 years) and SAASDV (31 years) (P = 0.0006; non-VHL vs. TV, P = 0.001; non-VHL vs. SAASDV, P = 0.003; TV vs. SAASDV, P = 0.85). Nine eyes (43%), 2 eyes (9%), and no eye (0%) were affected by atypical JRCHs in the TV, SAASDV, and non-VHL groups, respectively (P = 0.002; TV vs. SAASDV, P = 0.009; TV vs. non-VHL, P = 0.005; non-VHL vs. SAASDV, P = 0.52). 

Table 3.

 

Clinical Characteristics of Patients With JRCH According to Association With VHL Disease and Underlying Genetic Information (n = 49)

 

No difference was noted among the TV, SAASDV, and non-VHL groups in sex, baseline BCVA, frequency of JRCH diagnosed symptomatically, follow-up period, or final BCVA. In addition, no significant difference between the TV and SAASDV groups was found in family history of VHL disease or VHL-related systemic manifestations including neural hemangioblastomas, renal/pancreatic cysts or renal cell carcinoma, and neuroendocrine tumors (Table 3). 

By definition, patients in the non-VHL group had only one JRCH, as patients with JRCH and any peripheral RCH fulfilled the clinical criteria for VHL. Thus, we investigated the relationship between peripheral RCH and JRCH in the VHL group. Eyes in the SAASDV group tended to be more commonly affected by peripheral RCH than those in the TV group, however, the difference was not statistically significant (95% vs. 74%; P = 0.08). A significantly greater proportion of eyes in the SAASDV group was complicated by at least one large peripheral RCH (>1.5 mm) than those in the TV group (45% vs. 16%; P = 0.042) (Table 3). Furthermore, compared with VHL eyes with classic JRCHs, those with atypical JRCHs were similarly affected by any peripheral RCH (72.7% vs. 87.1%; P = 0.53), whereas they had a lower prevalence of large (>1.5 mm) peripheral RCHs (0% vs. 41.9%; P = 0.027). The median numbers of peripheral RCHs were 2 (IQR, 1–7) and 1 (IQR, 0–2) for VHL eyes with classic and atypical types, respectively (P = 0.02) (Supplementary Fig. S3). 

VHL-Related JRCHs With Bilateral and Unilateral Involvement

Among 36 patients with VHL-related JRCHs, 30 patients were further categorized into bilateral (9 patients) and unilateral JRCH groups (21 patients), depending on the state of bilateral eyes at the most recent visit. Six patients were excluded owing to enucleation or phthisis of the contralateral eyes. 

Clinical and genetic characteristics of patients with VHL-related bilateral JRCHs and those with unilateral JRCHs are shown in Table 4. Comparisons between them revealed no significant differences in sex ratio, age at diagnosis, positive family history, prevalence of atypical JRCHs, and follow-up period (Table 4). Although the proportion of eyes complicated with peripheral RCHs was similar between the two groups, patients with bilateral JRCHs were significantly more likely to present with large peripheral RCHs (>1.5 mm) in the initially affected eye than those with unilateral JRCHs until the final visit (86% vs. 30%; P = 0.02). Regarding genetic spectrum, patients with bilateral JRCHs had a lower prevalence of SAASDVs residing in the β-domain than those with unilateral JRCHs, however, the difference was not statistically significant (0% vs. 33%; P = 0.066). 

Table 4.

 

Comparison of Clinical and Genetic Features of Patients With VHL-Related JRCHs According to Laterality (n = 30)

 

Discussion

In the current comprehensive retrospective study, we delineated the novel characteristics, unique natural history, and visual outcome of atypical JRCHs in the era of SS-OCTA, which were further classified according to whether the JRCH broke through the ILM. JRCH patients with VHL–TVs had a higher likelihood of developing atypical JRCHs. Compared with patients with unilateral JRCHs, those with bilateral JRCHs were significantly more likely to harbor large peripheral RCHs in the initially affected eye and were less likely to harbor VHL–SAASDVs in the β-domain, albeit without statistical significance. 

The value of SS-OCTA in screening silent JRCHs and aiding in the early diagnosis of occult classic JRCHs was verified in the study. Early detection of asymptomatic JRCHs is the prerequisite for proper management, such as adjusting follow-up plans. Although the en face OCTA slabs of the whole retina may omit the JRCHs located mainly in the middle/outer layer, the en face avascular slab provides the most sensitive way to detect exophytic JRCHs. 

Moreover, the depiction of the clinical course of atypical JRCHs based on the relationship with the ILM using SS-OCTA was another important finding. The previous atypical JRCHs were found to be associated with variable fibrotic components basically on fundus photograph, and their natural histories were uncertain.^4,16 The current study is the first to propose breaking through the ILM as an imaging marker for atypical JRCH activity. Those lesions breaking through the ILM (atypical type B) attach to the posterior hyaloid membrane, contain more fibrotic components, grow faster with greater traction on the retina, and are more likely to require surgical interventions than those limited to the ILM (atypical type A). Moreover, even the atypical type A JRCHs showed varying degrees of progression on the OCTA scan. The progressive changes support that they are hemangiomas rather than developmental anomalies, which should be quiescent. However, Takahashi et al.⁵ observed that flat RCHs, including two in the peripapillary region, showed no sign of progression in size and were free of exudative or tractional changes. The study did not include the atypical type B JRCHs, whose changes tended to be more remarkable. Additionally, the follow-up periods were relatively short. Compared with eyes with classic VHL–JRCHs, those with atypical VHL–JRCHs were complicated with fewer peripheral RCHs and less commonly affected by large peripheral RCHs, which are more difficult to destroy.^3,11 The relative sparing of the peripheral retina also plays a role in the generally favorable visual outcome. 

Genetic characteristics of JRCHs, especially the atypical type, remain long uninvestigated. Herein, for the first time, we demonstrated that VHL–TV correlated with a significantly higher prevalence of atypical JRCHs than VHL–SAASDV. Interestingly, VHL–null and VHL–type 2B (representing human R167Q) mutants have different effects on vascular development in murine retinas.¹⁷ Type 2B mutant endothelial cells exhibit a significantly higher transcriptional level of VEGF-A and Notch pathway genes than full VHL deletion. Furthermore, type 2B mutant vasculature displays more severe morphological defects. Atypical JRCHs, in our study and previous reports,^4,5 encompassed abundant flow similar to classic nodular JRCHs, but displayed milder biological behaviors. We speculate that TVs, with less perturbation of signaling cascade downstream of VHL protein in the retina, may give rise to atypical JRCHs, which represents a forme fruste. Given that the truncating type was reported to be linked to a lower predominance of JRCH,¹⁰ postulating that the omission of atypical JRCHs without SS-OCTA might contribute to the previous result is intriguing. Overall, our findings suggest that patients with VHL–TVs should consider the development of atypical JRCHs and potential surgical interventions. Although the biofunctional discrepancy across different VHL variants provides the rationale for the predilection of atypical JRCHs in certain genetic backgrounds, the newly uncovered genotype–phenotype correlations warrant further studies into the underlying molecular mechanism. 

In our study, 72% of patients with JRCHs were estimated to be associated with VHL disease clinically. Consistent with previous findings, VHL patients presented at a considerably younger age than non-VHL patients, and 17% of VHL patients without a family history were regarded to carry a de novo mutation.^6,18,19 The estimate of VHL disease approximates the data (84%) from a national study in which all the participants with nonspecified RCHs were tested genetically.²⁰ However, in our patients meeting the clinical criteria for VHL disease, the detection rate of next-generation sequencing in VHL mutations was 97%. A germline VHL variant can be identified with standard genetic testing in more than 95% of patients clinically diagnosed with VHL disease.²¹ The negative genetic results may be due to mosaicism with low variant allele fractions,^22,23 chromosomal translocation,²⁴ or a cryptic variant.²⁵ Thus, we believe a thorough VHL workup is indispensable for patients with a solitary JRCH, even if they are genetically negative using current methods. 

First described in 1966, bilateral optic disc involvement is an extremely rare phenotype in patients with VHL disease.²⁶ In our present study, bilateral JRCHs were present in 30% of patients with VHL disease, which was higher than the 17% estimate in a previous report.⁶ The difference should be attributed largely to the high sensitivity of SS-OCTA in detecting JRCHs. Furthermore, we found that the presence of large peripheral RCH in the initial eye affected by JRCH was associated with developing a contralateral JRCH. Because bilateral JRCHs pose a potential risk of bilateral central vision loss, patients with the aforementioned characteristics in the initially affected eye warrant particular attention to the contralateral eye. We acknowledge that the follow-up duration of the bilateral group was longer than that of the unilateral group, and the likelihood of bilateral involvement may increase with time. However, differences regarding age at diagnosis and follow-up duration were statistically insignificant. This finding suggests that the development of bilateral JRCHs may depend on genetic basis rather than time solely. Accordingly, we observed that VHL patients with β-domain SAASDVs were less likely to develop bilateral JRCHs (P = 0.066). This finding implies patients harboring α-domain SAASDVs and TVs may need closer monitoring in case of bilateral involvement. Although our interpretation is limited statistically by the small sample size, our results suggest personalized screening of bilateral eyes on a clinical and genetic basis. 

This study has limitations. First, it is a retrospective, single-center study. Although the number of patients was relatively large for JRCH, several patients were lost to follow-up. Second, owing to phthisical changes in the contralateral eye without past information on the juxtapapillary region, some VHL patients were unavailable for bilateral evaluation, which made statistical analysis challenging. Finally, the results with statistical significance were exploratory, and the corresponding hypotheses should be tested in further confirmatory studies. 

In conclusion, we updated the clinical spectrum of JRCH by investigating the occult and atypical types with SS-OCTA and elaborating on the characteristics of bilateral VHL–JRCHs as an extremely rare phenotype. SS-OCTA is valuable in the identification and longitudinal evaluation of JRCHs. In particular, atypical type B JRCHs are more aggressive in natural history, which necessitate surgical interventions more frequently than atypical type A. Patients with large peripheral RCHs in the initially affected eye are more susceptible to developing JRCHs bilaterally. The study also expanded the genetic profile of VHL–JRCHs by identifying five novel VHL variants and showed that VHL–TV was dramatically associated with atypical JRCHs. 

Acknowledgments

The authors thank the statistician Jie Hong for her guidance on the statistical analysis. 

Supported by the Shanghai Hospital Development Center Foundation (SHDC12023116), the Lumitin Vision to Brightness Research Funding for the Young and middle-aged Ophthalmologists (BCF-KH-YK-20221123-11), the National Natural Science Foundation of China (No.82101149, 82301213), the Shanghai “Rising Stars of Medical Talents” Youth Development Program. 

Disclosure: X. Zhuang, None; F. Gao, None; Y. Xuan, None; Z. Sun, None; X. Ye, None; X. Huang, None; R. Jiang, None; J. Wu, None; M. Wang, None; Q. Chang, None; G. Xu, None; W. Liu, None