Corresponding Author: Kyoung Woo Kim, MD, PhD. Department of Ophthalmology, Chung-Ang University Hospital, Chung-Ang University College of Medicine, 102 Heukseok-ro, Dongjak-gu, Seoul 06973, Korea. Pterygium is characterized by conjunctival fibrovascular growth extending onto the cornea, where it consequently distorts the anterior corneal surface [1]. As the lesion advances over the cornea, it creates localized surface irregularities and alters the corneal curvature, thereby increasing astigmatism and various wavefront aberrations [2-4]. Such optical changes can reduce visual quality and this may occur even when refractive error itself is relatively small [5]. Recent studies have demonstrated that the degree of corneal optical distortion correlates with the morphological parameters of the pterygium, such as its size, thickness, and the horizontal invasion length (HIL) of pterygium, representing the extent of corneal invasion [6,7]. Understanding the early morphological changes that precede these optical deteriorations is crucial for timely intervention and accurate prediction of visual outcomes.
Recent advances in anterior segment (AS) swept-source optical coherence tomography (SS-OCT) have enabled high-resolution visualization of the pterygium-cornea junction, overcoming the resolution limitations of conventional optical coherence tomography or Scheimpflug camera-based imaging [2,8]. This technology allows for the precise identification of microstructural changes at the interface between corneal stroma and the pterygium head. Through this enhanced visualization, we observed a linear, subepithelial hyperreflective structure originating at the distal epithelial edge of the pterygium head and extending further beneath the corneal epithelium. We have termed this structure the “subepithelial pterygium pedicle (SPP)” (Fig. 1). While clinicians may have empirically observed this subepithelial extension during surgical excision, it has neither been formally named nor quantitatively analyzed in previous literature. Therefore, this study aims to introduce the concept of the SPP, quantify its length using AS SS-OCT, and evaluate its independent effect on anterior corneal curvature and aberrations.
Materials and Methods
Ethics statement
The research protocol received approval from the Chung-Ang University Hospital Institutional Review Board (No. 2510-007-19596). The requirement for informed consent was waived due to the use of deidentified data and the retrospective nature of the study. The entire process followed the Declaration of Helsinki principles.
Study design and subjects
This study was designed as a single-center retrospective cohort based on review of clinical chart records. This study included 72 eyes from 72 patients with primary nasal-only pterygium which was evaluated using AS SS-OCT (Anterion, Heidelberg Engineering) between September 2021 and June 2023. Patients with double-headed pterygium, isolated temporal pterygium or recurrent pterygium were excluded, as were those with a history of corneal refractive surgery. To ensure statistical independence, for patients presenting with bilateral pterygium, one eye was randomly selected for analysis.
Study protocol
This study was performed as follows: (1) data acquisition: measurement of anterior keratometry and anterior corneal aberration values using AS SS-OCT; (2) morphologic analysis (quantification of the HIL of the pterygium and SPP length using AS SS-OCT images); (3) intergroup comparison (group 1, demographics and pterygium severity between the non-SPP group vs. group 2, the SPP group); (4) correlation analysis (evaluation of the impact of SPP length adjusted for HIL on the anterior corneal keratometry and aberrations); and (5) cutoff determination (identification of the SPP length cutoff value associated with significant deterioration in anterior corneal aberration by receiver operating characteristic [ROC] analysis).
Clinical severity grading
All eyes with pterygium were graded for clinical severity based on the T grades, which considers the translucency of the pterygium body stroma [1]. T grading system classifies pterygium as follows: T1 with clearly visible underlying episcleral vessels, T2 with episcleral vessels partially covered by the pterygium body, and T3 with episcleral vessels completely covered by the pterygium body.
Anterior corneal keratometry and wavefront aberrations based on AS SS-OCT
Using AS SS-OCT, the anterior simulated keratometry (Sim K) average was measured at the 3-mm ring. Anterior corneal wavefront aberration parameters, including root mean square (RMS) lower-order aberration (LoA), RMS higher-order aberration (HoA) and first- to fourth-order individual aberrations, including vertical and horizontal tilt, defocus, oblique and the ratio of with the rule (WTR) to against the rule (ATR) astigmatism, oblique trefoil, vertical and horizontal coma, horizontal trefoil and spherical aberration, were analyzed within a 6-mm diameter anterior corneal zone.
Measurement of HIL and SPP length based on AS SS-OCT
HIL and SPP length were measured in images from the Metrics App in AS SS-OCT. The Metrics App acquires radial B-scans at fixed meridians (e.g., 0°, 30°, 60°, 90°, 120°, and 150°). Although the device’s Imaging App permits free meridian adjustment, it does not support the specific quantitative measurement tools required for this analysis. Therefore, to ensure standardized quantification, we utilized the Metrics App and selected the radial B-scan that most closely aligned with the central longitudinal axis of the pterygium (typically 0°, 30°, or 150° meridian) to capture the maximal extent of HIL and SPP.
HIL was measured as the horizontal linear distance representing the extent of pterygium invasion onto the cornea, defined from a tangent reference line at the limbal surface to the advancing edge of the pterygium head, in accordance with previously described methods [6,7,9]. SPP was defined as a linear subepithelial hyperreflective structure that originates at the distal epithelial edge of the pterygium head and extends further beneath the corneal epithelium as seen in both slit-lamp biomicroscopy (Fig. 1A) and AS SS-OCT (Fig. 1B). The SPP length was measured on the image that best delineated the pterygium morphology, using the Metrics App of the AS SS-OCT. To clarify the measurement protocol, a step-by-step illustration is provided in Fig. 1C-1E. We verified the SPP by magnifying the pterygium head (Fig. 1C) and identifying the region of interest where the SPP starts and extends distally (Fig. 1D). The measurement landmarks were defined based on tomographic features: the proximal start point (arrow a) corresponds to the epithelial edge of the pterygium head, and the distal end point (arrow b) corresponds to the termination of the linear high-signal SPP (Fig. 1D). The SPP length was measured as the perpendicular distance between two virtual reference lines passing through these points (Fig. 1E). For comparison, a representative optical coherence tomography image of a pterygium without SPP is shown in Fig. 1F, where the white arrow indicates the absence of the subepithelial structure emanating from the epithelial end. All SPP length values were measured by a single experienced examiner (KWK).
Statistical analysis
All statistical analyses were performed with Prism ver. 10 (GraphPad). Parametric or nonparametric tests were selected based on the normality of the data distribution. Comparisons between groups classified according to the presence of SPP were performed using the Student t-test and chi-square test. The correlation between HIL and SPP length was assessed with Spearman rank correlation test. To evaluate the correlation of SPP length adjusted for HIL with anterior corneal keratometry and wavefront aberrations, a partial correlation analysis was conducted while controlling for the effect of HIL. To determine the cutoff value of SPP length for predicting significant anterior corneal horizontal tilt, ROC curve analysis was performed. The Youden Index (sensitivity + specificity − 1) was calculated to identify the threshold that maximizes the combined sensitivity and specificity. Finally, analysis of covariance (ANCOVA) was performed with HIL as a covariate to compare subgroups divided by the cutoff value of SPP length. All values were expressed as mean ± standard deviation, and statistical significance was defined as p-value of < 0.05.
Results
Demographic and clinical characteristics of pterygium
A total of 72 eyes from 72 subjects with primary nasal pterygium were included in this study. Table 1 presents the demographic information and clinical profiles of subjects with pterygium. The subjects consisted of 34 men (47.2%) and 38 women (52.8%), with an mean age of 59.5 ± 11.9 years. In terms of the T grade classification, 21 eyes (29.2%) were classified as T1, 40 eyes (55.6%) as T2, and 11 eyes (15.3%) as T3. The mean HIL was 4.16 ± 1.41 mm.
Comparison of baseline characteristics between the groups with and without SPP
The subjects were classified into two groups based on the presence of SPP, group 1 (without SPP) and group 2 (with SPP). In group 2, where SPP was identified on AS SS-OCT, the mean SPP length was 842.3 ± 408.1 μm (Table 2). The proportion of female patients was 26.7% in group 1 and 59.6% in group 2, which showed a statistically difference (p = 0.023). The mean HIL was 3.37 ± 1.64 mm in group 1. This value was significantly lower than that in group 2, where the corresponding value was 4.37 ± 1.28 mm (p = 0.014). However, no significant differences were observed between the two groups in terms of mean age or T grade (p = 0.247 and p = 0.925, respectively).
Correlation of SPP length with HIL, anterior corneal keratometry, and wavefront aberrations
There was a statistically significant positive correlation between HIL and SPP length (r = 0.376, p = 0.001). Consequently, we analyzed the correlation between SPP length and anterior corneal refractive and aberration profiles both before and after controlling for HIL to investigate the SPP-specific effect.
In the unadjusted analysis, SPP length showed significant positive correlations with multiple parameters, including RMS LoA, RMS HoA, horizontal tilt, oblique astigmatism, WTR/ATR astigmatism, oblique trefoil, horizontal coma, and horizontal trefoil (Table 3). However, after controlling for HIL, most of these associations lost statistical significance. Notably, the anterior horizontal tilt remained significantly correlated with SPP length even after adjustment (p = 0.026). In contrast, no significant correlations were observed between SPP length and anterior Sim K or other wavefront aberration profiles (Table 3).
Determination of the cutoff value of SPP length for anterior horizontal tilt
Given that anterior horizontal tilt was the only parameter significantly correlated with SPP length after controlling for HIL, we next sought to identify the SPP length cutoff value that best discriminates anterior corneal horizontal tilt (Table 3). First, we analyzed the distribution of horizontal tilt values across the entire study cohort and identified that the threshold for the upper quartile (top 25%) was 3.65 μm. Accordingly, we performed ROC curve analyses to determine the SPP length cutoff for predicting horizontal tilt values ≥3.65 μm.
For predicting horizontal tilt ≥3.65 μm, the area under curve was 0.740 (95% confidence interval [CI], 0.604-0.875; p = 0.002) (Table 4). An SPP length cutoff of >635.5 μm yielded a Youden J index of 0.3889, with a sensitivity of 83.3% and specificity of 55.6%. Alternatively, a cutoff of >853.0 μm showed the same Youden J index (0.3889) but provided a lower sensitivity (55.6%) but better specificity (83.3%), indicating a trade-off for clinical screening. Since lower specificity implies a higher rate of false positives, where significant tilt might be driven by factors other than SPP, we prioritized high specificity to define a distinct phenotype of subepithelial invasion. Then, considering that the cutoff of approximately 853 μm consistently demonstrated the best predictive performance or balance across both thresholds, we selected 850 μm as the final cutoff, approximating the statistical threshold of 853 μm for clinical practicality.
Table 5 presents the results of the ANCOVA, performed with HIL as a covariate, to compare the estimated marginal means of anterior corneal horizontal tilt between subgroups stratified by this 850 μm cutoff. A statistically significant difference was observed between the subgroups (F = 6.051, p = 0.016). The unadjusted mean horizontal tilt values were 1.81 ± 1.85 μm for the SPP <850 μm group and 3.75 ± 2.52 μm for the SPP ≥850 μm group. After adjusting for HIL, the estimated marginal means remained significantly distinct at 2.01 ± 0.25 and 3.22 ± 0.42 μm, respectively.
Discussion
Previous studies have demonstrated that morphological profiles such as pterygium size, thickness, and HIL influence anterior corneal astigmatism, irregularity and aberrations [2,3,6,10]. In this study, we identified a previously unreported subepithelial hyperreflective structure from the pterygium head, which we termed as SPP, using AS SS-OCT. This region represents a subepithelial microstructure of the pterygium, and we aimed to investigate this microstructure in detail and revealed its influence on the anterior corneal aberration, especially anterior horizontal tilt.
Ha and Kim [6] examined several morphometric profiles including HIL, height, thickness and residual corneal thickness/central corneal thickness to analyze the optical impact of pterygium and to predict surgical outcomes, identifying HIL as a key morphological determinant of AS distortion. Similarly, previous studies also demonstrated that horizontal or relative invasion length significantly correlates with corneal astigmatism and HoA [11,12]. However, these studies focused primarily on surface level morphology and did not assess the underlying subepithelial microstructures. In contrast, our study expands the analytical scope to the microstructural level, linking subepithelial morphology with optical aberrations thus deepening the pathophysiological understanding of pterygium induced visual distortion.
In the group analysis, HIL was significantly greater in group 2 (with SPP) than in group 1 (without SPP), and the sex-ratio difference likely reflects a sampling artifact due to the limited number of cases. A significant positive correlation between HIL and SPP length further suggests that both parameters represent markers of a shared fibrovascular proliferative pathway. To determine whether SPP represents a secondary by-product of HIL or an independent contributor to corneal distortion, we evaluated the relationships between SPP length and anterior corneal profiles while adjusting for HIL. Using ANCOVA with HIL as a covariate, SPP length remained significantly correlated with anterior horizontal tilt. Horizontal tilt, measured by AS SS-OCT, quantifies the degree of wavefront tilt in the horizontal direction and represents one of the anterior corneal wavefront aberration components. As a first-order Zernike term, horizontal tilt is categorized within LoA. Clinically, horizontal tilt reflects a deviation of the optical axis from the visual axis in the horizontal direction [13]. Although tilt is a LoA that can theoretically be corrected with spectacles, in the context of pterygium, excessive horizontal tilt signifies a gross asymmetric mechanical distortion. While this may not permanently limit visual potential like HoAs, it serves as a critical surgical indicator. A high tilt value suggests that the SPP is exerting significant traction. Therefore, identifying this change preoperatively may suggest the need for a more extensive subepithelial dissection to potentially release this mechanical stress. Mechanistically, we hypothesize that the SPP acts as a subepithelial fibrous anchor exerting tractional stress. Since the pterygium progresses horizontally, this tension vector is aligned with the horizontal meridian. Consequently, the SPP induces specific asymmetric flattening that manifests as horizontal tilt, while the vertical meridian remains relatively unaffected due to the lack of vertical tractional forces.
Among the cutoff values assessed, an SPP length of 850 μm most clearly distinguished the subgroups with respect to anterior horizontal tilt. Importantly, this difference remained significant even after adjusting for HIL. Patients with an SPP length ≥850 μm exhibited significantly greater horizontal tilt values, suggesting that beyond this threshold, subepithelial invasion exerts a mechanical distortion on the optical axis independent of the superficial pterygium size. Conversely, if the SPP is shorter than 850 μm, its capacity to induce corneal tilt appears limited. Thus, 850 μm may represent a critical microstructural cutoff for monitoring optical degradation in pterygium progression.
It is noteworthy that pterygium invasion often extends into the 6-mm wavefront analysis zone, potentially confounding aberration measurements. However, by treating HIL as a covariate in our statistical models, we adjusted for the direct optical effect of the pterygium mass. The persistence of a significant correlation between SPP and horizontal tilt after this adjustment supports the hypothesis that the SPP contributes a distinct tractional component beyond the mere physical presence of the lesion.
From a surgeon’s perspective, after removing the epithelium over the pterygium head, the subjacent SPP embedded within the underlying tissue can be identified. Based on our finding that SPP length correlates with mechanical tilt, we hypothesize that incomplete removal of this structure could theoretically allow residual tethering forces to persist. As current pterygium surgery emphasizes micro-level anatomical dissection [14,15], the SPP may represent a relevant target for dissection to release this traction. However, it is important to note that this study did not evaluate postoperative outcomes. Therefore, while we suggest that more extensive dissection in eyes with SPP ≥850 μm might be beneficial to relieve mechanical stress, future longitudinal studies are required to confirm whether complete SPP excision actually translates to improved refractive stability or reduced recurrence.
This study has several limitations. First, SPP identification and length measurements were performed manually, which may inherently result in observer variability and potentially compromise reproducibility. While this limitation could be overcome in future studies by implementing interobserver and intraobserver variability assessments or using automated segmentation with quantitative validation tools, we strengthened measurement reliability in the present study by having all data collected by a single experienced examiner to minimize intraobserver variation. Second, due to the retrospective design, baseline anterior corneal keratometry and wavefront aberration data prior to pterygium development were unavailable. Furthermore, this study focused on preoperative morphological-optical correlations and did not assess postoperative outcomes. Thus, we could not provide direct evidence linking SPP removal to surgical prognosis, which remains a subject for future investigation. Consequently, direct prepterygium and postpterygium comparisons could not be performed, which limits the ability to determine the precise extent to which SPP alone contributed to anterior optical changes.
Another limitation is the lack of direct histopathological correlation. Due to the retrospective design and the tendency for delicate SPP tissue to distort during peeling, obtaining intact specimens for histological analysis was not feasible. Consequently, the SPP should be interpreted as a tomographic imaging biomarker of subepithelial invasion rather than a definitive histological entity. Future prospective studies are needed to clarify its precise cellular composition. Nevertheless, this study is the first to quantitatively evaluate subepithelial morphology using AS SS-OCT and to objectively establish the association between SPP and anterior corneal aberrations, thereby providing novel insights into the microstructural-optical interplay in pterygium.
In addition, it should be noted that the standard deviations for wavefront aberration metrics were relatively large compared to the means. This variability reflects the heterogeneous nature of pterygium morphology and individual differences in corneal biomechanical response. However, the fact that significant correlations and group differences were observed despite this variability reinforces the robustness of the association between SPP length and anterior corneal distortion.
In conclusion, this study demonstrates that SPP length independently correlates with anterior horizontal tilt after adjusting for HIL, identifying it as a significant tomographic biomarker of subepithelial invasion. Rather than merely affecting refractive error, SPP length may serve as a practical surgical indicator for predicting mechanical corneal distortion, suggesting the potential clinical value of identifying this tethering tissue during pterygium excision.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Full text via PMC
Download Citation
Print
