Phenotypic Distinctions Between EYS- and USH2A-Associated Retinitis Pigmentosa in an Asian Population

Abstract

Purpose: This study compares clinical characteristics of retinitis pigmentosa (RP) associated with mutations in the EYS and USH2A genes in a Southeast Asian cohort.

Methods: Prospective single-center study of families with EYS- or USH2A-associated RP seen at the Singapore National Eye Centre. Comprehensive ophthalmic evaluations, multimodal imaging, genetic testing, and longitudinal follow-up identified clinically useful differentiating features between the two genotypes.

Results: A total of 300 families with RP were enrolled, with EYS- and USH2A-associated RP, accounting for 24.7% of all probands and 50.7% of solved or likely solved cases. USH2A cases were predominantly nonsyndromic RP (75%). EYS-associated RP was more severe in functional and structural outcomes, and patients were more myopic than USH2A (SE −3.31 vs. −0.69; P < 0.0001). EYS RP displayed peripapillary nasal sparing on autofluorescence imaging more frequently than USH2A (57.6% vs. 26.7%; P = 0.006), whereas USH2A cases more often had a parafoveal ring (73.3% vs. 30.3%; P = 0.0002). Multiple logistic regression identified diagnostic features with 83.2% accuracy in distinguishing between EYS and USH2A, validated in a second unrelated clinical cohort.

Conclusions: EYS- and USH2A-associated RP have overlapping clinical presentations but can often be distinguished based on a constellation of phenotypic features including disease onset and severity, refractive error, and fundus autofluorescence. These diagnostic features may support a more effective diagnostic strategy for these common forms of RP.

Translational Relevance: Distinct clinical features differentiating EYS- and USH2A-associated RP provide valuable diagnostic tools that may inform personalized management and facilitate targeted interventions in clinical practice.

Introduction

Pathogenic variants in the EYS (eyes shut homolog) and USH2A (Usher syndrome type 2A) genes are responsible for a substantial proportion of retinitis pigmentosa (RP) worldwide and are the most prevalent genetic causes of RP in East Asia.^1–8 The coding sequences of both genes are exceptionally large—EYS is 9.4 kb and USH2A is 15.6 kb—and both are essential for structural integrity and function of photoreceptors,^9,10 whereas USH2A is also important for inner ear cell function and is associated with both syndromic RP (Usher syndrome) and nonsyndromic autosomal recessive RP.¹¹ The large size of both genes explains at least in part the high prevalence of RP secondary to variants in these genes,¹² but presumed founder mutations such as EYS c.4957dupA (p.S1653fs) in Japanese¹³ and EYS c.6416G > A (p.C2139Y)⁴ and USH2A c.2802 T > G (p.C934W)¹⁴ in ethnic Chinese also influences the prevalence of RP caused by EYS and USH2A variants. 

Retinal phenotypes of EYS and USH2A-associated RP have been described previously.^2,5,15–18 Pathogenic variants in both genes cause generalized rod-cone degeneration, although atypical forms have been observed with sectoral and irregular distributions.^4,17 Symptom onset varies between studies and in the case of USH2A, between syndromic and nonsyndromic disease, but has been reported to occur between the second to fourth decades of life,^14,19–21 with widely variable median delays in presentation, from three to 24 years depending on the study.^{19,20,22–24} Cohort studies have demonstrated wide regional variability in presenting vision for both forms of RP, with median presenting best-corrected visual acuity (BCVA) ranging from logMAR 0.1 to 0.6.^14,20,22,23 USH2A-associated RP has most commonly been reported as a typical generalized RP pattern of retinal involvement, whereas EYS-associated RP is associated with atypical retinal distributions that may be genotype-specific.^4,17,18,25 There is also some evidence that patients with EYS is associated with myopia,^22,23 although this is not uncommon among patients with RP.²⁶ In general, however, there are no established phenotypic features that can reliably discriminate between these two prevalent forms of RP beyond the presence of congenital hearing loss in USH2A-associated Usher syndrome. This limitation is important in East Asia, where both EYS and USH2A are the most prevalent causes of RP, and standard diagnostic approaches such as exome sequencing may disclose a single genetic variant in either gene or multiple variants in both genes, confounding attempts to solve the genotype of an affected individual. 

This prospective study aims to compare the clinical characteristics, disease progression, and genetic profiles of patients with EYS- and USH2A-associated RP. Conducted at tertiary eye centers in Singapore and Japan, this study enrolled patients diagnosed with RP and confirmed mutations in either the EYS or USH2A genes. Through comprehensive ophthalmic evaluations, genetic testing, and longitudinal follow-up, this research seeks to elucidate the differential impacts of these genetic mutations on retinal degeneration. The outcomes of this study aimed to provide insights into the natural history of EYS- and USH2A-associated RP, potentially informing personalized management strategies and future therapeutic developments. 

Methods

Study Subjects

This was a prospective study of nonsyndromic RP patients with EYS- or USH2A-associated RP that were sequentially enrolled at the Singapore National Eye Centre, Singapore, during the period from March 2021 to December 2022, and at the Department of Ophthalmology at the Nagoya University Hospital. Patients with a clinical diagnosis of retinitis pigmentosa (RP), or rod-cone dystrophy, were included for further analysis. Diagnoses were made by fellowship-trained retinal specialists based on clinical history, physical examination, ophthalmic imaging, psychophysical testing, and genetic testing. The study was approved by the SingHealth and Nagoya University Hospital Institutional Review Boards and carried out in accordance with the ethical standards of the 1964 Declaration of Helsinki and its later amendments. Informed consent was obtained from all participants. 

Clinical Evaluation

All subjects underwent clinical evaluation, including a targeted clinical history, assessment of BCVA by logMAR testing, intraocular pressure measurement, and fundus examination by slit lamp biomicroscopy. Age at symptom onset (“onset”) was defined as the age at which a patient reported the onset of symptoms that were in the opinion of the managing clinician thought to be related to underlying retinitis pigmentosa (Table), whereas age at presentation (“presentation”) was the age at which a patient was initially seen in clinic. Color fundus photography (Triton DRI; Topcon Healthcare, Tokyo, Japan), ultrawide fundus autofluorescence imaging (Optos, Marlborough, MA, USA), macular optical coherence tomography (OCT; Spectralis; Heidelberg Engineering, Heidelberg, Germany), Ishihara color vision testing, and Goldmann kinetic perimetry (Haag-Streit, Koeniz, Switzerland) was performed on all subjects. The extent of visual field was measured as the extent of the V4e isopter (field diameter in degrees) at the horizontal midline as previously described.⁴ Foveal-centered horizontal line scans from macular OCT (30°) raster images were used for measurement of the horizontal diameter of ellipsoid bands, using the built-in Spectralis Eye Explorer software caliper tool (Heidelberg Engineering), as described in detail elsewhere.²⁷ A value of zero was assigned if no discrete ellipsoid band was discernable. Horizontal and vertical visual field extent was measured as the number of degrees spanned by the V4e stimulus at the horizontal and vertical midlines on the Goldmann kinetic perimetry grid. Electroretinography was performed as an adjunctive test according to international society for clinical electrophysiology of vision (ISCEV) methodology were retained responses were anticipated based on clinical examination.²⁸ Where indicated, the spherical equivalent (SE) was calculated in patients for whom phakic refraction data were available and for whom no history of refractive surgery was present at the time of refraction. Selected phenotypic features were scored in a second cohort from Nagoya, Japan, using clinical and retinal imaging data acquired the same approaches to that used in the primary study cohort. All patients completed psychophysical and structural testing, and a subset (n = 20 of 35 for EYS, n = 18 of 36 for USH2A) also completed electrophysiological testing. 

Table.

 

Baseline Characteristics of the EYS- and USH2A-Associated Nonsyndromic RP Cohorts

 

Genetic Analyses

DNA was obtained from peripheral blood leukocytes according to standard procedures. Whole exome sequencing (WES) was performed by Macrogen (Geumcheon-gu, Seoul, South Korea) and Novogene (Singapore), and bioinformatics analysis of sequence data was performed at the SingHealth Duke-NUS Institute of Precision Medicine (PRISM), Singapore. When necessary, direct sequencing was performed with the BigDye Terminator Cycle Sequencing kit (Applied Biosystems, Foster City, CA, USA) with 10 ng of template DNA in each reaction and using a PCR program of 25 cycles of denaturation at 97°C for 30 seconds, annealing at 50°C for 15 seconds, and extension at 60°C for four minutes. Samples were analyzed in a 3130 Genetic Analyzer (Applied Biosystems). Genetic variant calls were made by comparison with SG10K,²⁹ a Singapore population-based genome reference database, with additional reference to the gnomAD genomic database³⁰ and ClinVar^31,32 for variants absent from SG10K. Variant phase was established by direct sequencing of the relevant variants in one or more unaffected first-degree relatives. The gnomAD genomic database²⁰ was queried to identify variants that were classified as pathogenic or likely pathogenic according to annotations from the ClinVar variant database. Determination of variant pathogenicity was based on criteria from the American College of Medical Genetics and Genomics. A subset of patients underwent panel-based clinical exome sequencing, performed by Molecular Vision Laboratories (Hillsboro, OR, USA), and these patients were also enrolled into the study and their clinical and genetic data analyzed with the research testing cohort. Genetic variant data are provided in Supplementary Tables S1, S2, and S3. 

Statistical Analyses

Statistical analyses were performed using Prism 10.2 (GraphPad Software, Boston, MA, USA). Comparisons between clinical subgroups were performed using Mann-Whitney (pairwise) and Kruskal-Wallis tests (>2 group analyses). LogMAR visual acuity values of 1.9, 2.3, 2.7, and 3.0 were used to represent counting fingers, hand motions, light perception, and no light perception, respectively, for the purposes of statistical comparisons. Kaplan Meier survival analyses and Mantel Cox log-rank tests were performed using the specified endpoints with ages at which patients reach each endpoint being imputed for analysis. Regression analyses used to correlate spherical equivalent with age at onset were performed using t tests for slope, with significance set at P < 0.05. Simple linear regression was used to compare changes in BCVA, goldmann visual field (GVF), and OCT outcomes between groups over time (age and years since symptom onset). Regression lines were compared using analysis of covariance (ANCOVA). Multiple logistic regression was performed following conversion of clinical features into categorical variables, whereby continuous variables, including age at symptom onset, spherical equivalent in diopters, and presenting logMAR BCVA, were binned according to the median value for each variable after pooling data from EYS and USH2A cases. 

Results

We included 300 unrelated individuals with RP in the initial cohort (Fig. 1), for whom 48.7% had a solved or likely solved disease-causing genotype. Collectively, EYS and USH2A accounted for 24.7% of all probands with RP and 50.7% of probands with solved or likely solved genotypes. Baseline characteristics of patients with nonsyndromic EYS- and USH2A-associated RP are described in the Table. Individuals with biallelic EYS- and USH2A-associated RP comprised 10.7% (n = 32 families; 35 affected individuals) and 14.3% (n = 43 families; 48 affected individuals) of families enrolled and genotyped during the study period, respectively (Fig. 1). Among EYS probands, 62.5% (n = 20) contained the common variant EYS c.6416G > A (p.C2139Y), whereas the common USH2A variant c.2802T > G (p.C934W) was present in 14.3 (n = 13) of USH2A probands. Usher syndrome was present in 12 (25%) individuals in the USH2A cohort, and these individuals were excluded from subsequent analysis. There were significantly more females in the USH2A cohort compared to the EYS cohort (62.5% vs. 37.1%; P = 0.007). Symptom onset occurred far earlier in patients with EYS compared to USH2A (median onset 25.0 and 35.0; P = 0.947; Fig. 2A), although limited sample sizes meant that this did not reach statistical significance. 

Figure 1.

 

Composition of RP genotypes in the Singapore RP study cohort. A total of 300 unrelated RP probands were included. Cases with biallelic EYS or USH2A RP, biallelic EYS or USH2A RP wherein one or both alleles included the EYS C2139Y or USH2A C934W variant or were only a single EYS or USH2A variant was identified, are shown in exploded slices. The remaining RP cases without EYS or USH2A, both solved/others and unsolved, are also shown.

Figure 2.

 

Baseline clinical features of RP cases with EYS or USH2A. (A) Age at symptom onset and presentation; (B) BCVA; and (C) horizontal visual field diameter (°) on the V4e isopter on Goldmann kinetic perimetry are sown as scatter plots. Median values are shown as horizontal bars. EYS and USH2A groups were compared with Mann Whitney tests. Ns, not significant.

Both groups initially presented at a similar age—median age at presentation for EYS was 38.5 and USH2A was 39.5 (P = 0.864; Fig. 2A). The median delay between onset of symptoms and presentation to the clinic was 9.49 years (interquartile range [IQR] 0.35–24.8 years) for EYS and 11.5 years (IQR 3.1–23.7 years) for USH2A, which was comparable between the groups (Mann-Whitney, P = 0.445). Presenting visual acuity and the horizontal diameter of the visual field (measured in degrees on GVF, V4e isopter) was also better in patients with USH2A compared to EYS, but this was not statistically significant (Figs. 2B, 2C). We further examined the two cohorts to compare the natural history of disease progression between the two cohorts and identify clinically useful phenotypic features that could be used to distinguish between the two genotypes. Survival curve analysis for BCVA of logMAR 0.5 (Snellen 20/60) and logMAR 1.0 (Snellen 20/200) revealed significant separation of the EYS and USH2A groups, with EYS median survival lower than USH2A in both cases (Figs. 3A, 3B). Median survival for logMAR 0.5 was 63.5 years for EYS and 72.5 years for USH2A (p = 0.010), and for logMAR 1.0 median survival was 70.1 and 79.3 for EYS and nonsyndromic USH2A, respectively (P = 0.002). Other structural and functional outcomes, including loss of visual fields, macular ellipsoid band loss, and ffERG extinction, were similar between EYS and USH2A, although there was a clear trend toward EYS being more rapidly progressive than USH2A (Fig. 3). Only a subset of cases (n = 20 for EYS and n = 18 for USH2A) underwent ffERG testing based on discretion of the managing clinician, so a selection bias for this aspect of the comparison cannot be excluded. Despite this, it appeared quite clear from BCVA, visual fields, OCT, and ERG metrics that EYS was the more severe and rapidly progressive disease. 

Figure 3.

 

Survival curves for EYS- and USH2A-associated RP clinical features. Visual acuity cutoffs of logMAR 0.5 (A) and 1.0 (B), visual field cutoffs of 60° (C) and 20° (D), ellipsoid band width of 15° (E) and 5° (F), and ffERG extinction for DA 0.01 (G), DA 10.0 (H), and LA 3.0 (I) are shown for both groups. Survival curves were compared using log-rank (Mantel Cox) tests for significance, with P values shown as insets.

Regression analysis of visual fields, visual acuity, and EZ band width as a function of patient age and years since symptom onset showed similar rates of functional and structural decline in EYS and USH2A, with years since symptom onset demonstrating better correlation over time than patient age (Fig. 4). Unlike the survival curve analyses in Figure 3, where EYS was clearly more severe than USH2A, the regression approach did not clearly separate two diseases. 

Figure 4.

 

Comparisons in outcomes for EYS- and USH2A-associated RP, based in patient age and time elapsed since symptom onset. Presenting BCVA (A and B), horizontal visual field diameter (C and D), and horizontal EZ band length (E and F) are plotted against patient age or years elapsed since symptom onset. Linear regression (lin. reg.) was used to compare the two genotypes, with ANCOVA analysis comparing the regression lines. Significance of ANCOVA is shown as P values inset.

We next examined additional baseline features that differed between the EYS and USH2A groups. Patients with EYS were significantly more myopic than patients with USH2A (SE −3.38 vs. −0.68; P < 0.0001; Supplementary Fig. S1). We hypothesized that the earlier onset of disease among EYS cases, with attendant loss of photoreceptors and potential disruption of normal emmetropization, may have contributed to higher myopia in EYS cases. There was a modest but significant correlation between age at symptom onset and myopia for USH2A patients, with earlier onset of symptoms being associated with higher myopia (r² = 0.1; P = 0.037; Supplementary Fig. S1). However, there was no correlation between symptom onset and myopia for EYS (r² = 0.009; P = 0.607; Supplementary Fig. S1). We also excluded a possible gender effect on myopia by separately comparing males and females with EYS and USH2A (Supplementary Fig. S2), but the significant finding of higher myopia in patient with EYS-associated RP remained. 

Fundus autofluorescence (AF) patterns are useful for the diagnosis and characterization of inherited retinal diseases, and EYS-associated RP is known to have atypical patterns of AF.^4,17,18 We thus reviewed the ultra-widefield AF patterns of the EYS and USH2A cohorts to identify distinguishing features. Two features, namely sparing of the peripapillary retina and the presence of a parafoveal hyperautofluorescent ring (the so-called Robson-Holder ring) were identified as potentially discriminating features between the two genotypes. Peripapillary nasal sparing was observed in 57.6% of EYS cases, compared to 26.7% of USH2A cases (P = 0.006; Fig. 5). Conversely, a parafoveal ring was observed in 30.3% of EYS cases but occurred in 73.3% of USH2A cases (P = 0.0002; Fig. 5). Given the progressive nature of RP and changes in AF patterns over time in a particular eye, we compared the ages of patients with nasal sparing and parafoveal ring to patients without these signs (Fig. 5). No significant difference in age was observed between individuals with and without the AF features, in both the EYS and USH2A cohorts. Logistic regression analysis demonstrated that an USH2A genotype was associated with a parafoveal ring (z value 3.443; P = 0.001) independent of patient age and ellipsoid band width (Supplementary Table S4), whereas an EYS genotype (z value 2.802; P = 0.005) and the ellipsoid band width (z value 3.352; P = 0.001) were independently associated with the presence nasal sparing (Supplementary Table S5). 

Figure 5.

 

Autofluorescence imaging distinguishes between EYS and USH2A RP. (A) Peripapillary nasal sparing was more prevalent among individuals with EYS (P = 0.0019) whereas a parafoveal ring was more prevalent among individuals with USH2A (P = 0.0002). (B) Illustrative cases for both features are shown, along with the EYS and USH2A genotypes identified for each case.

We pooled our findings for each of the potentially discriminating features for EYS and USH2A to determine whether a combination of phenotypic features would be clinically useful for distinguishing between the two genotypes. Cases with USH2A-associated Usher syndrome were excluded from this analysis. Multiple logistic regression identified five useful diagnostic features, namely age at symptom onset (<28 years), spherical equivalent (≤−2D), presenting BCVA (>0.2 logMAR), absent parafoveal ring on AF, and presence of nasal peripapillary sparing on AF. When combined these features yielded a diagnostic accuracy of 83.2% for distinguishing between EYS- and USH2A-associated RP, with a negative predictive power of 77.8% and positive predictive power of 80.0% (Fig. 6). The weightings of each clinical feature are described in Supplementary Table S6. 

Figure 6.

 

Multiple logistic regression analysis incorporating age at symptom onset, presenting BCVA, spherical equivalent, peripapillary nasal sparing, and perifoveal ring or fundus AF, were used to distinguish between EYS and USH2A (A). Negative predictive power (NPP) and positive predictive power (PPP) are shown. (B) The predicted probabilities for EYS and USH2A cohorts are shown as violin plots, where median and quartiles are shown as horizontal bars.

Finally, we tested the clinical utility of the model with a second cohort of patients previously enrolled at the University of Nagoya in Japan. For this test we scored each of the distinguishing features for 10 patients with EYS and nine patients with USH2A. In total, eight of 10 EYS and eight of nine USH2A cases were corrected predicted. This amounted to an accuracy of 84.2%, sensitivity (recall) of 80%, specificity of 88.9%, precision of 88.9%, and an F1 score (harmonic mean of precision and recall) of 84.2% for prediction of an EYS or USH2A genotype. 

Discussion

RP associated with variants in EYS and USH2A is by far the most common cause of RP in East Asia, and this was particularly evident in the current study, where EYS and USH2A were collectively responsible for more than half of RP cases with an identifiable disease-causing genotype. Distinguishing between these two genetic forms of RP is clinically useful both in Singapore and other East Asian nations, given their high prevalence and the implications for diagnosis, genetic counseling, and potential therapies. Our findings demonstrate clear differences in the phenotypic characteristics of these two forms of RP, contributing knowledge to the understanding of their natural history and clinical manifestations. We found that the age of symptom onset was generally earlier for EYS-associated RP, with our findings broadly aligning with previous studies reporting similar age ranges for symptom onset in East Asian populations. Japanese patients with EYS mutations presented with symptoms in their mid-20s,³ compared to age 27 for Chinese patients with USH2A mutations.² Both findings are consistent with our cohort's median onset age of 25 years for EYS and 28 years for USH2A, respectively. Although both groups exhibited comparable baseline visual acuity and horizontal visual field extents, survival curve analysis demonstrated that EYS was the more severe of the two genotypes across a range of functional and structural outcomes. However, our study also identified specific phenotypic markers that distinguish the two groups. Patients with EYS-associated RP were more myopic, and distinct fundus AF patterns differed between the groups, findings that have not been consistently emphasized in prior studies. 

AF patterns are potentially useful for diagnosing and characterizing inherited retinal diseases.^33–35 Our observation that EYS-associated RP commonly exhibits peripapillary nasal sparing corroborates earlier work^16,17 documenting unusual AF features among these individuals. Recent in vitro studies of EYS suggest that the EYS protein may play a role in photoprotection.⁹ This potentially explains the atypical patterns of retinal degeneration in a subset of patients with EYS RP, wherein loss of EYS function results in retinal phototoxicity and degeneration. 

The high prevalence of EYS c.6416G > A (p.C2139Y) and USH2A c.2802T > G (p.C934W) variants in our cohort is consistent with reports of presumed founder mutations in ethnic Chinese populations. Lin and colleagues¹⁴ recently reported on the high prevalence of USH2A c.2802T > G in Taiwanese patients with RP, while we previously demonstrated prevalence of EYS c.6416G > A among ethnic Chinese in Singapore.⁴ The precise geographic origins of these two genetic variants remain indeterminate, although both appear to be common among ethnic Chinese, with the EYS c.6416G > A variant in particular being prevalent among individuals of Southern Chinese descent. Interestingly, a recent analysis by the Human Genome Diversity Project identified EYS c.6416G > A exclusively in the She Chinese ancestral group that arose from Southern China.³⁰ This contrasts with USH2A c.2802T > G, which was present among ancestral groups from Central and Northern China, Mongolia, the Sakha region of Russia, and to a lesser extent Japanese.³⁰ This may explain the higher prevalence of EYS c.6416G > A and relatively lower prevalence of USH2A c.2802T > G among Singaporean Chinese, who are predominantly of Southern Chinese descent. The potentially large number of individuals affected by RP involving these variants makes both extremely attractive as gene therapy targets.⁴ Indeed, Pongpaksupasin and colleagues³⁶ recently developed a retinal stem cell line with EYS c.6416G > A, which may prove a useful model for testing targeted therapies. 

Our findings have several important implications for clinical practice and research. The identification of phenotypic markers such as myopia and specific AF patterns provides valuable diagnostic tools for distinguishing between EYS- and USH2A-associated RP. The association between EYS and myopia, independent of age at symptom onset, is particularly notable. Myopia in RP ostensibly occurs during childhood, with severe or early onset RP, such as RPGR-associated RP, being known to associate with myopia.²⁶ However, we were unable to demonstrate a clear association between symptom onset and myopia with EYS-associated RP. One possible explanation for this is that the EYS gene may play a role in the structural integrity of the photoreceptors, influencing refractive development. Our findings suggest that EYS-associated RP is clearly associated with myopia, while USH2A cases did not appear to be significantly myopic. This might be due to differences in the underlying pathophysiology of photoreceptor degeneration between the two genes. 

The diagnostic utility of the clinical phenotypes we identified is another important consideration. The five features identified here yielded good diagnostic accuracy and are clinically facile. Unlike black-box deep learning approaches, the phenotypic features described here are easily applied using widely available tools. These phenotypic features may be particularly useful in unsolved cases with incomplete genotypes where cases with strong suspicion based on clinical and imaging features may be prioritized for genome or long-read sequencing to solve incomplete EYS or USH2A genotypes. 

We acknowledge several important limitations of the current work. The cohort size, while representative, may limit the generalizability of the findings to other regional populations. We did demonstrate clinical utility of several phenotypic features for EYS and USH2A RP in a Japanese cohort distinct from the primary study cohort from Singapore, but it remains to be seen if these features will carry over to, for example, European and U.S.-based cohorts. We would also like to emphasize that our use of fundus autofluorescence, while potentially useful as one of several phenotypic features to might distinguish between EYS and USH2A, is limited by the transient nature of autofluorescent fundus features observed in RP. Nasal sparing and the parafoveal ring seen on autofluorescence are expected to exist only during a specific period for any given patient,³⁷ so their use may not be relevant in very early or very late-stage disease. Both signs are also by no means genotype-specific and can appear across a wide range of RP genotypes.³⁸ Larger multi-center studies are needed to confirm our results and extend them to additional populations. Moreover, larger cohorts would enable better delineation of the clinical differences in EYS- and USH2A-associated RP. An obvious trend towards milder disease was observed in individuals with USH2A-associated RP, and it is likely that this difference would become significant with increased cohort sizes. It is also clear after reviewing the genotypes of our cohort that our results likely exhibit regional or ethnic bias. The large proportion of recurrent variants in the EYS and USH2A cohorts would exert a significant impact on any comparison made between the groups and potentially limit the generalizability to other geographic regions or ethnic groups with differing genetic structures. 

Exploring innovative gene therapy approaches that can overcome the challenges posed by the large size of EYS and USH2A genes will be crucial. Current gene therapy strategies are limited by the packaging capacity of adeno-associated viral vectors, commonly used for retinal gene therapy. Dual-vector strategies, which split the gene into two separate adeno-associated viral vectors, or the use of larger viral vectors such as lentivirus, could potentially address this issue. Additionally, gene editing technologies such as CRISPR/Cas9 offer promising alternatives for correcting mutations in large genes.³⁹ This study reports on the first comparative study of the two most common causes of RP in East Asia. The identification of distinctive phenotypic markers and progression patterns for EYS- and USH2A-associated RP offers practical diagnostic tools and highlights areas for future research and therapeutic development. 

Acknowledgments

Supported by research grants from the SingHealth Foundation (grant number R1748/71/2020), PRISM cohort study by EYE ACP-PRISM Precision Medicine Initiative Fund 05/FY2020/EX/06-A41, 05/FY2022/EX(SLP)/69-A131(b), 05/FY2023/EX(SLP_FY22)/162-A217, and 05/FY2023/EX(SLP)/208-A263. 

Disclosure: E.Y.H. Yeo, None; T. Kominami, None; T.-E. Tan, None; L. Babu, None; K.G.S. Ong, None; W. Tan, None; Y.M. Bylstra, None; K. Jain, None; R.W.C. Tang, None; S.Z. Farooqui, None; S.P.R. Kam, None; C.-M. Chan, None; R.S. Mathur, None; S.S. Jamuar, None; W.K. Lim, None; K. Nishiguchi, None; B.J. Fenner, None