Clinical and Imaging Characteristics of Thin and Thick Idiopathic Epiretinal Membranes Assessed by High-Resolution OCT

Abstract

Purpose

The purpose of this study was to determine the prevalence and imaging characteristics of thin and thick idiopathic epiretinal membranes (ERMs) using high-resolution (HR) optical coherence tomography (OCT) and to identify clinical factors associated with ERM thickness phenotype.

Methods

All patients underwent comprehensive ophthalmologic examination and multimodal imaging, including HR-OCT. ERM morphology was evaluated on fovea-centered OCT scans and classified as thin or thick based on preretinal reflectivity patterns. Multivariable generalized estimating equations (GEEs) were used to identify factors independently associated with ERM type while accounting for inter-eye correlation.

Results

Fifty-four eyes included in this analysis and 28 (51.9%) were classified as thin ERM, whereas 26 (48.1%) were thick ERM. Patients with thick ERM were significantly older than those with thin ERM (81.6 ± 6.2 vs. 75.7 ± 10.8 years, P = 0.039). Pseudophakia was more prevalent in the thick ERM group (84.6% vs. 60.7%, P < 0.001), as was the presence of AMD (69.2% vs. 28.6%, P = 0.003). Mean central macular thickness (CMT) and best-corrected visual acuity (BCVA) did not differ significantly between thin and thick ERM groups (P > 0.05). In multivariable analysis, pseudophakia (odds ratio [OR] = 10.33, 95% confidence interval [CI] = 1.29–83.10, P = 0.028) and age-related macular degeneration (AMD; OR = 4.12, 95% CI = 1.01–16.87, P = 0.049) were independently associated with thick ERM.

Conclusions

HR-OCT enables in vivo differentiation of thin and thick idiopathic ERMs. Thick ERM is independently associated with pseudophakia and AMD but not with greater CMT or worse BCVA, suggesting that ERM thickness phenotype may reflect differences in factors other than vitreoretinal interface biology alone.

Idiopathic epiretinal membrane (ERM) is a fibrocellular proliferation that develops on the inner retinal surface typically at the macula and can produce retinal wrinkling, distortion of the foveal contour, and variable degrees of visual impairment.^1–3 Although ERM is common in older adults, its clinical phenotype is heterogeneous ranging from a delicate “cellophane” membrane with minimal traction to a thicker, highly contractile tissue associated with marked retinal distortion.

Within this framework, Sebag's concept of anomalous posterior vitreous detachment (PVD) and vitreoschisis provides a compelling explanation for phenotypic variability in ERM.^1,4 Vitreoschisis refers to splitting within the posterior vitreous cortex during attempted separation, leaving an outer lamella adherent to the retina while the remaining vitreous detaches anteriorly. This residual lamella may contain hyalocytes and collagen, potentially driving a more cellular and contractile preretinal proliferation. Sebag hypothesized that PVD may occur with or without hyaloid splitting, which could plausibly yield different ERM “types,” including relatively thin membranes versus thicker, more substantial preretinal tissue.^1,4–6 However, despite the biologic appeal of this hypothesis, in vivo characterization of ERM thickness phenotypes and their vitreoretinal correlates has been limited, and controversy persists particularly because standard clinical OCT has historically offered axial resolution on the order of approximately 5 to 7 µm constraining confident assessment of ultrathin preretinal tissue layers, subtle posterior hyaloid configurations, and microstructural vitreoretinal interface features.

However, the staging schemes primarily reflect secondary retinal remodeling and do not specifically characterize the structural organization of the preretinal tissue itself. The architecture of the ERM at the vitreoretinal interface may represent an additional feature of disease phenotyping that is not captured by current staging systems and could have practical implications for intraoperative visualization and membrane peeling strategies. High-resolution optical coherence tomography (HR-OCT) may therefore provide an opportunity to further refine ERM phenotyping by directly evaluating preretinal tissue architecture.

Recent advances of HR-OCT with approximately 3 µm axial resolution provide an opportunity to revisit these questions with improved anatomic fidelity and superior visualization of fine retinal details.^7,8 Because idiopathic ERM and age-related macular degeneration (AMD) are both strongly age-associated⁹ and frequently coexist, and because AMD primarily involves the outer retina rather than the vitreoretinal interface, eyes with AMD were not excluded provided that the ERM was idiopathic and that preretinal morphology could be reliably assessed on HR-OCT unlike many prior idiopathic ERM studies.^2,3 In this cohort of eyes with idiopathic ERM, we used HR-OCT to (1) determine the prevalence of thin versus thick ERM, (2) describe their imaging characteristics at the vitreoretinal interface and within the retina, and (3) identify clinical and structural predictors associated with ERM type. By leveraging improved axial resolution to refine phenotyping, this work aims to clarify whether thickness-based subtypes consistent with proposed differences in PVD mechanics and internal limiting membrane (ILM) microdefect-mediated proliferation can be reproducibly identified in vivo, and whether such subtypes are associated with distinct anatomic signatures that may inform prognosis and future mechanistic study.

Methods

Study Design and Participants

This cross-sectional study included consecutive patients with a diagnosis of ERM who underwent multimodal imaging at the Doheny – UCLA Eye Centers (Pasadena, CA, USA) between April 2025 and October 2025. The study adhered to the principles of the Declaration of Helsinki, and complied with the Health Insurance Portability and Accountability Act (HIPAA). Written informed consent for imaging and use of anonymized data for research purposes was obtained from all participants.

Inclusion criteria were: (1) diagnosis of idiopathic ERM confirmed by clinical examination and OCT; (2) availability of baseline volumetric OCT scans (20 × 20 degrees, 1024 × 97, with ART >9 centered on the fovea); and (3) uncomplicated cataract surgery from more than 1 year.^10,11 Eyes with AMD were not excluded, provided that the ERM was deemed to be otherwise idiopathic and that the vitreoretinal interface and preretinal tissue could be reliably visualized on the selected HR-OCT scan. Because ERM originates at the inner retinal surface, the presence of non-neovascular AMD below the retina was not deemed to interfere with the ability to phenotype the ERM. AMD was diagnosed according to Beckman criteria. Exclusion criteria were (1) secondary ERM (e.g., associated with retinal vascular disease, uveitis, trauma, retinal detachment, laser retinopexy, or prior retinal surgery), coexisting macular disease affecting retinal architecture like neovascular AMD defined by CONAN criteria,¹² and (2) OCT images with significant motion artifacts or segmentation errors precluding reliable analysis.

Study Population and Procedures

All subjects underwent a comprehensive evaluation including detailed medical history collection, best-corrected visual acuity (BCVA) assessment, intraocular pressure measurement, slit-lamp biomicroscopy, ultra-widefield fundus photography (Optos California; Optos, Dunfermline, UK), cSLO Multicolor, and HR-OCT and spectral-domain OCT (SD-OCT; Spectralis, Heidelberg, Heidelberg, Germany) with an axial resolution of approximately 3 µm. ERM definition and AMD definition was assessed by two masked graders with confirmation of a third senior investigator. ERM diagnosis was confirmed by a senior retina specialist according to fundus and OCT examination.

Image Grading

After initial image quality screening, grading was performed on a horizontal high-definition OCT B-scan passing through the foveal center, magnified to 200% to enhance visualization of preretinal and retinal microstructural features. The foveal center was defined as the point of maximal foveal depression. In eyes with loss of normal foveal contour due to ERM related traction, the foveal center was defined as the point of maximal inward displacement of the outer nuclear layer (ONL).¹³ When ectopic inner foveal layer (EIFL) was present, the scan passing through the center of the EIFL complex was selected and ERM staging was assessed accordingly.³ The region of analysis was defined according to the Early Treatment Diabetic Retinopathy Study (ETDRS) grid, centered on the fovea. For descriptive purposes, the B-scan was subdivided into the central 1-mm subfield (fovea), inner ring (parafovea), and outer ring (perifovea). ERM morphological assessment and thickness measurements were confined to the central 1-mm ETDRS subfield, corresponding to the area of greatest tractional relevance for foveal integrity. Central macular thickness (CMT) was calculated as previously described.² The vitreous status was assessed for PVD detection on OCT by reviewing the entire macular scan volume. Eyes with vitreomacular traction were excluded.

OCT Image Analysis and ERM Thickness Measurement

HR SD-OCT images were analyzed using the device's built-in caliper tool with 1:1 micron scaling. The central 1-mm ETDRS subfield was predefined as the region of interest for ERM thickness assessment. ERM thickness was measured on the horizontal fovea-centered B-scan at the location of maximum preretinal thickness within the central 1-mm subfield. Thickness was defined as the axial distance between the vitreous–ERM interface and the ERM–retina interface (internal limiting membrane), with calipers placed perpendicular to the retinal surface (Figs. 1, 2).

Figure 1.

High-resolution OCT of epiretinal membrane morphology and vitreoretinal interface. (A) High-resolution spectral-domain optical coherence tomography (OCT) horizontal B-scan through the foveal center shows an epiretinal membrane overlying the internal limiting membrane with associated inner retinal surface undulation and foveal contour distortion. The preretinal tissue shows variable reflectivity and focal thickening, consistent with heterogeneous ERM architecture. Scale bar = 200 µm. (B) Corresponding multicolor fundus image shows retinal surface wrinkling and vessel distortion in the macular region reflecting tangential traction exerted by the epiretinal membrane. (C) Magnified OCT view of the vitreoretinal interface shows a preretinal layered tissue with focal elevation, separation from the retinal surface and complete posterior vitreous detachment. Arrowheads indicate stratified preretinal tissue consistent with thick ERM and focal points of vitreoretinal adhesion. The underlying inner retinal layers show surface irregularity and inward displacement.

Figure 2.

High-resolution OCT phenotyping of thin and thick epiretinal membrane patterns. (A–C) Representative example of a thick epiretinal membrane (ERM) phenotype. A En face infrared image with Early Treatment Diabetic Retinopathy Study (ETDRS) grid centered on the fovea, illustrating the central 1-mm subfield used for ERM assessment. B High-resolution horizontal OCT B-scan through the foveal center shows a continuous hyper-reflective preretinal band overlying the internal limiting membrane, associated with mild inner retinal surface undulation and limited foveal contour distortion. C Magnified view of the vitreoretinal interface shows a layer preretinal hyperreflective line (arrowhead) with internal stratification, consistent with a thick ERM. (D–F) Representative example of a thin ERM phenotype. D En face infrared image with ETDRS overlay centered on the fovea. E High-resolution OCT B-scan shows prominent preretinal tissue associated with marked inner retinal surface irregularity and inward displacement of the inner retinal layers. F Magnified view highlights a thin, heterogeneous preretinal complex with increased thickness and reflectivity, consistent with a thin ERM phenotype. Scale bars = 200 µm.

ERM phenotype was classified as thin or thick based on preretinal tissue morphology and supported by a quantitative thickness criterion similarly to past strategies adopted for outer retina assessment (Fig. 3).^14,15 Thin ERM was defined as a single, continuous hyper-reflective preretinal band without internal stratification, whereas thick ERM was characterized by multilamellar or stratified preretinal tissue with heterogeneous reflectivity. Specifically, thin ERM was defined as a single continuous hyper-reflective line overlying the internal limiting membrane without visible stratification on the fovea-centered scan. Thick ERM was defined by the presence of two or more distinguishable layers above the retinal surface, with a multilayered configuration visible across at least two consecutive B-scans within the central ETDRS subfield. To minimize misclassification related to OCT system resolution and manual boundary placement, ERM phenotype classification was primarily based on morphologic features observed on OCT. For membranes classified as thick ERM, a minimum thickness of approximately 15 µm at the point of maximal preretinal tissue thickness within the central ETDRS subfield was used as a supporting criterion. This cutoff exceeds approximately four to five times the nominal axial resolution (approximately 3–4 µm) reported for HR SD-OCT, a margin commonly used in optical imaging to ensure reliable separation of tissue boundaries despite speckle noise, point-spread-function broadening, and partial-volume effects.¹⁶ In addition, transverse (lateral) resolution in SD-OCT systems is typically on the order of 15 µm, which can further affect boundary definition for ultrathin preretinal structures, supporting the use of a conservative thickness threshold well above the resolution limit.⁷ The thickness of the ERM was measured at the point of maximal preretinal tissue thickness on the fovea-centered scan using manual calipers. Two masked graders (authors A.Q. and J.H.) independently performed measurements, and discrepancies were adjudicated by a senior grader (author S.S.).

Figure 3.

High-resolution OCT of thick epiretinal membrane (ERM). (A) High-resolution OCT B-scan through the foveal center showing a thick ERM overlying the internal limiting membrane (ILM). The inset shows the area enlarged in panel B. (B) Magnified view of the boxed region in panel A showing multilayered preretinal tissue above the ILM. Arrowheads indicate distinct hyper-reflective layers within the ERM, demonstrating internal stratification of the membrane. (C) High-resolution OCT B-scan from a second eye with thick ERM, again showing a hyper-reflective preretinal membrane adherent to the inner retinal surface. The boxed region indicates the area enlarged in panel D. (D) Magnified view of the boxed region in panel C. Lines outline the boundaries of the preretinal tissue, illustrating the increased thickness of the ERM.

Statistical Analysis

Continuous variables are reported as mean ± standard deviation, whereas categorical variables are presented as counts and percentages. Comparisons between eyes with thin and thick ERM were performed to evaluate differences in demographic, clinical, and imaging characteristics. Variables were assessed for normality distribution using Shapiro-Wilk test. Group comparisons for continuous variables were performed using independent-samples t-tests or Mann–Whitney U tests, whereas categorical variables were compared using chi-square or Fisher's exact tests as appropriate. To identify factors independently associated with ERM type while accounting for within-patient correlation in individuals contributing both eyes to the analysis, a multivariable analysis was performed using generalized estimating equations (GEEs) with a logistic link function, with eye and patient's ID as random effects. ERM type (thin versus thick) was specified as the binary dependent variable. Age, CMT, presence of AMD, and lens status were included as covariates. Results from the multivariable model are presented as odds ratios (ORs) with corresponding 95% confidence intervals (CIs). Intergrader reliability for ERM category was assessed using Cohen's kappa for phenotype classification. All statistical tests were two-sided, and a P value less than 0.05 was considered statistically significant. All analyses were performed using R software version 4.1.0 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Cohort Characteristics and Comparisons

A total of 82 eyes from 70 patients were initially screened. Fifty-four eyes from 42 patients with idiopathic ERM were included in the final analysis after application of inclusion and exclusion criteria, and 12 patients contributed with both eyes. Twenty-eight eyes were classified as thin ERM, and 26 eyes were classified as thick ERM. Baseline demographic and clinical characteristics of the two groups are summarized in Table 1. Complete PVD was observed in 50 of 54 eyes (92.6%). With regard to ERM staging, 27 eyes (50.0%) were classified as stage 1 ERM, 21 eyes (38.9%) as stage 2, and 6 eyes (11.1%) as stage 3. No eyes met the criteria for stage 4 ERM. Patients with thick ERM were significantly older than those with thin ERM, with a mean age of 81.62 ± 6.24 years compared with 75.68 ± 10.77 years, respectively (mean difference 5.94 years, 95% CI = 1.15–10.73, P = 0.039). Pseudophakia was more prevalent in the thick ERM group, present in 22 eyes (84.6%), compared with 17 eyes (60.7%) in the thin ERM group (P = 0.0006). Conversely, phakic eyes were more frequently observed in the thin ERM group. The mean thickness in the thick ERM group was 19 ± 2.5 µm.

Table 1.

Comparison Between Thin and Thick ERM Groups

Variables	Thin ERM	Thick ERM	P Value
Patients, n	24	22	—
Eyes, n	28	26	—
Lens status			0.0006
Pseudophakic	17 (60.7%)	22 (84.6%)
Phakic	11 (39.3%)	4 (15.4%)
AMD			0.002
Yes	8 (28.6%)	18 (69.2%)
No	20 (71.4%)	8 (30.8%)
AMD stage			—
Intermediate	4 (50%)	17 (94.4%)
Late AMD	4 (50%)	1 (5.6%)
Age, y	75.68 ± 10.77	81.62 ± 6.24	0.039
CMT, µm	338.68 ± 99.94	302.27 ± 67.85	0.197
BCVA, logMAR	0.22 ± 0.21	0.37 ± 0.36	0.139

AMD, age-related macular degeneration; BCVA, best corrected visual acuity; CMT, central macular thickness.

The P values in bold represent statistical significance.

AMD was observed in 18 eyes (69.2%) in the thick ERM group compared with 8 eyes (28.6%) in the thin ERM group (P = 0.002). Mean CMT was 338.68 ± 99.94 µm in the thin ERM group and 302.27 ± 67.85 µm in the thick ERM group (mean difference −36.41 µm, 95% CI = −82.87 to 10.05, P = 0.197). Mean BCVA was 0.22 ± 0.21 logMAR in eyes with thin ERM and 0.37 ± 0.36 logMAR in eyes with thick ERM (P = 0.139). Distributions of BCVA, CMT, and age by ERM type are illustrated in Figure 2³. Cohen's kappa was 0.89.

Generalized Estimating Equations Model

After adjustment for age and CMT, both lens status and the presence of AMD were still independently associated with ERM type. Eyes with AMD had significantly higher odds of having a thick ERM compared with eyes without AMD (OR = 4.12, 95% CI = 1.01–16.87, P = 0.049). Pseudophakic eyes were also significantly more likely to exhibit a thick ERM compared with phakic eyes (OR = 10.33, 95% CI = 1.29–83.10, P = 0.028). Age and CMT were not independently associated with ERM type in the multivariable model. Age showed an OR of 0.99 (95% CI = 0.89–1.10, P = 0.813), and CMT showed an OR of 0.99 (95% CI = 0.99–1.00, P = 0.215; Table 2).

Table 2.

GEE to Model ERM Type, Accounting for Within-Patient Correlation (Eye-Level Data)

Term	Estimate	OR (95% CI)	P Value
Intercept	0.517	1.68 (0.00–1911.78)	0.885
Age	−0.012	0.99 (0.89–1.10)	0.813
CMT	−0.006	0.99 (0.99–1.00)	0.215
AMD	1.416	4.12 (1.01–16.87)	0.049
Lens status	2.335	10.33 (1.29–83.10)	0.028

The P values in bold represent statistical significance.

Discussion

In this study, HR-OCT enabled in vivo differentiation of thin and thick idiopathic ERM based on distinct preretinal morphologic features. Using this approach, we found that thick ERM was significantly associated with pseudophakia and the presence of AMD, whereas CMT and BCVA were not independently related to ERM type.

Within this framework eyes with AMD were included provided that the ERM was idiopathic and that the preretinal membrane and inner retinal architecture could be reliably assessed on the scan of choice. As an ERM originates at the inner retinal surface, the presence of AMD does not inherently preclude meaningful evaluation of preretinal tissue morphology when imaging quality is sufficient, although several studies have demonstrated the potential influence of incomplete PVD on the evolution of AMD.^17–20 Accumulating evidence, however, suggests that AMD may be associated with secondary changes in Müller cell function, microglial activity, and ILM integrity, which could plausibly influence ERM development and maturation.^21–23 Moreover, vitreoretinal disorders are gaining importance in patients with AMD because it seems that they are associated with higher risk of atrophy progression and macular neovascularization.^18,19,24 Including eyes with AMD, therefore permits exploration of whether a degenerative retinal environment is associated with specific ERM thickness phenotypes, rather than assuming AMD to be solely a confounding condition to be excluded a priori. Finally, exclusion of AMD eyes could introduce selection bias by preferentially removing older patients, and potentially obscuring biologically relevant relationships among aging, vitreous dynamics, and ERM morphology.

Previous OCT-based classification systems have largely focused on secondary retinal changes rather than primary preretinal morphology.^3,13,25,26 Studies by Govetto and colleagues proposed ERM staging schemes based on the presence of EIFL, retinal thickening, and foveal architecture disruption.³ Although these classifications have been valuable in correlating retinal distortion with visual outcomes, they do not explicitly address the structural properties of the epiretinal tissue itself, which could potentially impact the understanding of tractional forces and have surgical implications. Earlier OCT studies may have failed to detect these differences due to axial resolution limitations of approximately 5 to 7 µm, which constrain reliable discrimination of ultrathin preretinal layers. Investigational HR SD-OCT achieves axial resolution on the order of approximately 3 µm and lateral resolution near approximately 10 to 15 µm allowing a detailed vitreoretinal interface and outer retina characterization.^7,16

Sebag et al. introduced the concept of anomalous PVD and vitreoschisis, proposing that incomplete separation of the posterior vitreous cortex may leave a residual cortical layer adherent to the macula.^1,4,6 This remnant vitreous cortex rich in hyalocytes and collagen may serve as a scaffold for more exuberant cellular proliferation and contractile membrane formation.⁴ Within this framework, ERMs arising in the setting of vitreoschisis could plausibly appear thicker and more multilayered than those forming after a cleaner vitreoretinal separation. Although this hypothesis has been influential, in vivo confirmation of thickness-based ERM phenotypes has been limited by the axial resolution of conventional OCT systems.

Recognizing thick ERM as a potentially more biologically complex entity, it is noteworthy that neither BCVA nor CMT differed significantly between groups. This finding is consistent with prior studies demonstrating imperfect correlation between OCT-derived thickness measures and visual function in ERM.^13,27,28 Visual acuity reflects only a limited aspect of macular function and may not capture subtle distortions, metamorphopsia, or contrast sensitivity deficits similarly to findings in eyes with inner retinal dimpling.²⁹ Similarly, CMT is influenced by multiple factors, including retinal stretching and segmentation variability, and may not directly reflect preretinal traction dynamics. The dissociation between ERM morphology and these conventional outcome measures underscores the value of refined phenotyping using HR-OCT and suggests that ERM thickness could potentially provide a complementary dimension of information.

The OR for pseudophakia was 10.33 indicating that pseudophakic eyes had substantially higher odds of exhibiting a thick ERM compared with phakic eyes, after accounting for age, CMT, and inter-eye correlation. The observed association between thick ERM and pseudophakia may support a mechanistic link between lens status, vitreous alterations, and ERM maturation.^6,30–33 Cataract surgery is known to accelerate vitreous liquefaction and PVD potentially increasing the likelihood of anomalous PVD with residual vitreous cortex adherence to the macula.^34,35 Such residual cortical vitreous may provide a scaffold for cellular proliferation and contribute to the formation of a more substantial, multilayered epiretinal tissue.^35,36 Although causality cannot be inferred from the present data, the strong association observed after multivariable adjustment seems to be consistent with this biologic framework.

The higher prevalence of thick ERM in eyes with AMD represents another notable finding. Importantly, AMD remained independently associated with thick ERM after controlling for age, suggesting that this relationship may not be attributable to age alone. Histologic and imaging studies have demonstrated Müller cell activation, microglial migration, and ILM alterations in AMD.^21,22 Although AMD has been excluded from many idiopathic ERM studies, emerging evidence suggests that degenerative changes in the inner retina, ILM, and Müller cell function may influence preretinal proliferation.^17,18,22,37 These changes may predispose the retinal surface to more extensive glial proliferation once an initiating event, such as PVD, occurs. These findings support the concept that a thick ERM represents a phenotype more likely to arise in a proinflammatory or degenerative retinal environment, with further potential impact on inner retina in addition to the impact on the outer retina by the degenerative process itself.

In contrast, age and CMT were not independently associated with ERM type in the multivariable model. The OR for age was close to unity (OR = 0.99) indicating that once AMD and lens status were accounted for, chronological age alone did not meaningfully influence the likelihood of thick versus thin ERM. Similarly, CMT showed an OR near 1.00, suggesting that ERM thickness phenotype is not simply a reflection of greater retinal thickening or traction severity. Collectively, these results support the hypothesis that thin and thick ERMs represent biologically distinct entities that may arise from different vitreoretinal interface conditions, potentially related to variations in PVD dynamics, residual vitreous cortex adherence, and ILM microdefects.

An additional consideration is the potential dynamic nature of ERM morphology over time. Although the present study focuses on cross-sectional phenotyping, ERMs are known to evolve, and spontaneous ERM separation or partial “auto-peeling” has been reported.³⁸ It is therefore conceivable that some thin ERMs observed in this study could represent a later stage following partial separation or remodeling of previously thicker membranes rather than an earlier stage of formation. Conversely, thin and thick ERMs may also reflect biologically distinct entities arising from different vitreoretinal interface conditions, such as variations in PVD dynamics or residual vitreous cortex adherence. Because the present study is cross-sectional, it cannot determine the temporal relationship between these phenotypes. Longitudinal HR-OCT studies will be required to clarify whether thin and thick ERMs represent sequential stages of disease evolution or distinct morphologic pathways.

The lack of a significant association between ERM phenotype and conventional outcome measures such as BCVA is consistent with prior literature highlighting the complex relationship between structural OCT findings and visual function in idiopathic ERM. This functional–structural dissociation has led to increasing interest in complementary outcome measures that better capture visual distortion experienced by patients. Quantitative assessment of metamorphopsia using tools such as M-CHARTS or other distortion-based metrics may provide a more sensitive functional correlate of macular traction and retinal surface irregularity. Future studies integrating HR-OCT characterization of epiretinal tissue architecture with metamorphopsia metrics may therefore help clarify the relationship between ERM phenotype and patient-reported visual symptoms.

This study has several limitations. The relatively small sample size and cross-sectional design limit causal inference and may reduce statistical power for detecting more subtle associations. Although HR-OCT improved visualization of the vitreoretinal interface, direct confirmation of vitreoschisis or residual vitreous cortex was not possible. In addition, functional outcomes beyond standard visual acuity were not assessed. Larger prospective studies incorporating longitudinal imaging and functional measures will be necessary to further clarify the clinical implications of ERM thickness phenotypes. A further limitation of this study is that ERM thickness and morphology were evaluated on fovea-centered B-scans rather than through full volumetric segmentation of the preretinal tissue. Three-dimensional segmentation approaches could enable more precise mapping of spatial heterogeneity and improve characterization of regional variation in ERM thickness, including coexistence of thin and thick membrane components within the same eye.

By integrating HR-OCT this study provides evidence that thin and thick idiopathic ERMs through the fovea represent distinct phenotypes associated with different clinical and ocular characteristics. The strong associations with pseudophakia and AMD support the concept that ERM thickness reflects factors other than vitreoretinal biology. Future longitudinal and surgical correlation studies will be essential to determine whether these phenotypes differ with respect to natural history, functional response to surgery, or intraoperative factors such as ERM friability, ERM adherence to the underlying retina, and/or visualization response to the use of adjuvant stains such as brilliant blue G or indocyanine green.