Anterior Segment Optical Coherence Tomography Epithelial Thickness Mapping for Lesion Delineation in Recurrent Corneal Erosion: A Paired-Method Study

Abstract

Purpose

To investigate disparities in lesion areas measured using external eye photography with corneal fluorescein staining (CFS) and anterior segment optical coherence tomography (AS-OCT)–based corneal epithelial thickness (CET) mapping in recurrent corneal erosion (RCE), aiming to provide a basis for laser and surgical planning.

Methods

Fifteen RCE patients underwent same-day external eye photography and AS-OCT. CFS lesions were manually delineated, and CET lesions were identified by marking regions with a thickness ≥ 60 µm. The two methods were compared retrospectively. Patients with extensive CET mapping–identified lesions underwent phototherapeutic keratectomy (PTK) after manual epithelial debridement, with clinical outcomes and CET map findings evaluated.

Results

The lesion area defined by CET mapping was significantly larger than that measured by CFS, with a significant median difference of 9.26 mm² (95% confidence interval [CI], 5.87–14.42; P < 0.0001). Although the two methods were strongly correlated (r = 0.904; P < 0.0001), Bland–Altman analysis revealed a consistent mean bias. In 10 cases where CFS revealed only inferior involvement, CET mapping detected superior extension in five eyes, enhancing treatment planning accuracy. PTK was performed on nine patients, with no recurrences occurred (1 loss to follow-up). Smooth corneal epithelium with normalized thickness were observed in CET maps at a median of 7 months (range, 3–10; n = 5).

Conclusions

CET mapping is a sensitive and reliable tool for detecting larger lesion areas in RCE compared to CFS, enabling more accurate treatment planning and improved management outcomes.

Translational Relevance

AS-OCT–based corneal epithelial thickness mapping can refine clinical decision-making by revealing subclinical epithelial changes, supporting personalized treatment strategies for RCE.

Introduction

Recurrent corneal erosion (RCE) is a common corneal disorder characterized by repeated spontaneous breakdown of corneal epithelium due to its abnormal adhesion to the underlying basal lamina.¹ It typically presents with a sudden onset of ocular pain, is mostly encountered upon initial awakening, and is associated with photophobia, blurred vision, lacrimation, and redness. The primary cause of RCE is mechanical trauma, and the second most prevalent etiology is epithelial basement membrane dystrophy (EBMD).² Additionally, diabetes, dry eye syndrome, and ocular rosacea have been identified as potential risk factors.¹ Diagnosis of RCE typically begins with a comprehensive history to identify any past ocular trauma or the presence of EBMD. A thorough physical examination, particularly through slit-lamp biomicroscopy, is essential for assessing the condition of the corneal epithelium. Direct illumination can highlight loose or eroded epithelium, as well as cystic and map-like changes associated with EBMD. The application of topical fluorescein solution, referred to here as external eye photography with corneal fluorescein staining (CFS), facilitates the staining of irregular and loose epithelium and the visualization of non-staining lesions that protrude through the tear film, a phenomenon commonly observed in EBMD.³ Using CFS allows for a clearer distinction of the extent of RCE lesions compared to direct illumination with a slit-lamp biomicroscope without staining. Anatomically, the inferior half of the cornea is more commonly affected by RCE, as the upper half benefits from the protective coverage of the upper eyelid and the additional safeguard provided by Bell's phenomenon during injury.^1,2,4

In addition to slit-lamp biomicroscopy, recent studies have increasingly highlighted that anterior segment optical coherence tomography (AS-OCT) is an excellent tool for detecting corneal epithelial cell lesions, particularly for observing the epithelium and its adhesion to the stroma, where the primary changes associated with RCE occur.^5,6 AS-OCT–based corneal epithelial thickness (CET) mapping can provide en face thickness measurements of the corneal epithelium across different areas and is commonly used in the latest commercially available models with high precision, offering a resolution of 5 µm vertically and 15 µm horizontally. This provides a level of precision in assessing the extent and severity of epithelial abnormalities that cannot be achieved with conventional methods.^5,7–12

The treatment options for RCE include both medical and surgical approaches. The goal of surgical treatments is to induce a healing response in the anterior stroma, which ultimately contributes to the formation of a stronger epithelial adherence. For small lesions outside the visual axis, anterior stromal micropuncture and yttrium–aluminum–garnet (YAG) laser stromal puncture are commonly performed due to their efficiency and suitability for an office setting.^1,2,5,13,14 Lesions that are extensive or located within the visual axis require more rigorous and invasive treatment, such as debridement and superficial keratectomy.^1,4,14 Phototherapeutic keratectomy (PTK) is recommended for macroerosions or when other treatments fail. For large or central corneal lesions, anterior stromal micropuncture and YAG laser pose the risks of scarring or incomplete treatment, potentially causing recurrence. PTK, however, minimizes scarring risk and improves outcomes.^1,4,5

Accurately identifying the extent of abnormal epithelium in patients with RCE is crucial for selecting the appropriate surgical approach and preventing postoperative recurrence. Traditionally, slit-lamp biomicroscopy, with or without the use of CFS, has been used to assess these lesions; however, recurrence remains a possibility even after surgical or laser treatments.¹⁵ This study aimed to compare the regions of loosened epithelium identified using CFS and CET mapping. We hypothesized a priori that CET mapping would identify significantly larger lesion areas than CFS and reveal a different spatial distribution of abnormal epithelium. The comparison is focused on lesion delineation rather than diagnostic accuracy, as a definitive gold-standard reference for the exact boundaries of epithelial loosening is currently lacking. We specifically seek to evaluate whether CET mapping provides greater precision in identifying lesion areas compared to CFS, thereby offering a more optimal method for defining the RCE lesion area to assist in treatment planning.

Materials and Methods

This retrospective case series was approved by the Institutional Review Board of the National Taiwan University Hospital (IRB no. 202307100RINA), and the study was conducted in accordance with the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants who underwent surgical treatment. The study design and reporting followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for observational studies.

Subjects

We retrospectively reviewed patients 20 years of age and above who were diagnosed with RCE at National Taiwan University Hospital between January 2020 and June 2023. RCE was confirmed by (1) clinical history of at least one previous episode of corneal erosion, and (2) corneal erosion and/or epithelial loosening presented in CFS or under slit-lamp examination. Patients were included only if they received CET mapping on the same day of an eligible external eye photography. Exclusion criteria included (1) ocular surface diseases other than RCE, EBMD, or dry eye disease that may affect corneal epithelial homeostasis (e.g., limbal stem cell deficiency, corneal ulcer); and (2) prior ocular surface surgeries that may affect corneal epithelial attachment or thickness (e.g., PTK, photorefractive keratectomy, corneal cross-linking, stromal puncture). Clinical histories of these patients, including previous trauma history and other potential etiologies, were recorded. Treatment modalities provided to the patients were recorded, and PTK was recommended unless non-medical factors prevented it, such as fear of pain, financial concerns, or inability to take extended leave from work. Although CFS and CET mapping results were not masked to the primary care surgeon, the decision to undergo PTK was mainly based on clinical judgment and the patient's preference.

Acquisition and Analysis of Corneal Images

Corneal images were acquired by CFS and CET mapping. For external eye photographs, the lesion areas were measured by manually drawing on the CFS images by three independent researchers (W-LH, J-HC, L-YC) (Figs. 1, 2) and then automatically calculated using ImageJ (National Institutes of Health, Bethesda, MD). To ensure methodological rigor, measurements were repeated three times at 1-week intervals using images in a randomized order. Observers were masked to clinical data and previous results until all measurements were completed. Intra-observer reliability was then calculated based on these repeated sessions.

Figure 1.

Case 9 was a 58-year-old patient with a 3-month history of recurrent corneal erosion who had previously undergone corneal epithelial debridement twice at another clinic. (A) External eye photography of the affected eye, taken with direct illumination during the first visit to our clinic. (B) External eye photography of the affected eye with CFS during the first visit. (C) AS-OCT–based CET mapping of the affected eye during the first visit. (D) CET mapping of the healthy contralateral eye during the first visit. (E) Manual delineation of the lesion areas identified by CFS (green) and by CET mapping (orange). CET mapping delineated a larger lesion area compared to CFS, encompassing the central, inferior, and superior cornea, whereas CFS identified the lesion only in the inferior half of the cornea. Observers were masked for clinical data during manual delineation of CFS and CET.

Figure 2.

Case 4 was a 19-year-old patient with a 3-month history of RCE who had previously undergone corneal epithelial debridement twice at another clinic. (A) External eye photography of the affected eye, taken with direct illumination during the first visit to our clinic. (B) External eye photography of the affected eye with CFS during the first visit. (C) AS-OCT–based CET mapping of the healthy contralateral eye during the first visit. (D) CET mapping of the affected eye during the first visit. (E) Manual delineation of the lesion areas identified by CFS (green) and CET mapping (orange). Observers were masked for clinical data during manual delineation of CFS and CET.

AS-OCT was performed using the Optovue OCT system (Visionix USA, Lombard, IL). CET maps (9 mm in diameter) were automatically generated following the AS-OCT scan. The regions of thickened epithelium (60 µm and above, color-coded as yellow in the preset of the Optovue OCT system) were manually marked by the same three independent researchers and then automatically calculated using ImageJ (Figs. 1, 2). This threshold was used as a pragmatic criterion based on the normative central epithelial thickness of 52 to 55 ± 3 to 4 µm reported in previous studies,^8,16,17 representing an increase of approximately 10% (or 1.5–2 SD) in CET. The involvement of RCE on the cornea recorded by both methods was further divided into subdivisions including a central region (central 2 mm in diameter) and 24 peripheral regions grouped into paracentral (2–5 mm), mid-periphery (5–7 mm), and periphery (7–9 mm) zones across eight directions. For patients who underwent PTK, CFS and CET mapping were repeated at least 2 months after surgery to evaluate the postoperative outcomes.

Phototherapeutic Keratectomy

Patients undergoing PTK first had all loose corneal epithelium manually scraped off using a no. 64 blade, ensuring complete removal over a central area of at least 8 mm in diameter. PTK was then applied to ablate 20 to 35 µm of corneal tissue, with artificial tears (Artelac eye drops; Bausch & Lomb, Vaughan, ON, Canada) applied between every 5- to 10-µm ablation. During the procedure, the patient's head was gently rotated to ensure the preset 6.5-mm ablation zone could cover as much of the corneal surface as possible. The PTK treatment was performed using the VISX Star S4 IR Excimer Laser System (Johnson & Johnson Vision, Jacksonville, FL). The corneal surface was then irrigated with cool balanced salt solution, without the use of mitomycin C. Postoperatively, a bandage contact lens was applied, and topical levofloxacin and betamethasone eye drops were used. The bandage contact lens was replaced and changed every 2 weeks and was worn for a total of 6 weeks for corneal protection.

Statistical Analysis

Statistical analyses were performed using R 4.3.2 (R Foundation for Statistical Computing, Vienna, Austria). Intra-observer and interobserver reliability was assessed using the intraclass correlation coefficient (ICC) based on a two-way mixed-effects model, with values above 0.9 indicating excellent reliability. Paired-sample analyses were performed using Wilcoxon signed-rank tests for ordinal or continuous variables. Spearman's rank correlation analysis was performed to assess the correlation between the lesion size determined by CFS and CET mapping. To assess agreement between measurement methods, a Bland–Altman analysis was performed.¹⁸ Effect sizes (Cohen's d) with 95% confidence intervals (CIs) were calculated for lesion area comparisons. Additionally, a sensitivity analysis was conducted using alternative CET map thresholds of 55 and 65 µm to test the robustness of the 60-µm definition. P < 0.05 was considered statistically significant.

Results

Patient Characteristics

A total of 15 patients with unilateral RCE were included in this study, consisting of nine females and six males. The demographic data are listed in Table 1. The underlying etiologies of RCE in these patients included seven cases of trauma-induced RCE with a clear history of triggering events and eight cases without a significant history of trauma.

Table 1.

Demographic Data of Patients in the Study

Case No.	Age (Y)	Sex	Laterality	Trauma History	Pre-Exam Episodesa	Subsequent PTK
1	34	F	OS	Yes	2	No
2	40	F	OS	Yes	3	Yes
3	23	F	OS	No	4	Yes
4	19	F	OS	Yes	5+	Yes
5	22	M	OS	Yes	4	No
6	57	M	OS	No	2	Yes
7	35	M	OS	No	3	No
8	44	F	OD	Yes	3	Yes
9	58	M	OD	No	3	Yes
10	28	F	OS	Yes	3	Yes
11	25	F	OD	Yes	4	Yes
12	40	F	OS	No	2	No
13	28	M	OD	No	5+	No
14	27	M	OD	No	4	No
15	43	F	OD	No	2	Yes

F, female; M, male; OS, left eye; OD, right eye.

^aPre-exam episodes are the number of recurrent corneal erosion episodes before receiving examinations.

Comparison of Lesion Areas Defined by CFS and CET Mapping

The lesion areas were delineated by three observers with excellent intra-observer reliability and interobserver agreement. The intra-observer ICCs for all observers were excellent for both methods (range, 0.985–0.999). The interobserver ICC for CFS was 0.979 (95% CI, 0.951–0.992), and that of CET mapping was 0.999 (95% CI, 0.998–1.000), both demonstrating excellent agreement.

Figures 1 and 2 present representative images of patients where CFS and CET mapping were used to identify lesion areas. Figure 3 provides a comparison of corneal epithelial lesion sizes determined by these two methods. The data show that the lesion areas defined by CET were consistently equal to or larger than those defined by CFS in all 15 patients. Specifically, the areas of fluorescein staining were contained within or overlapped with regions of increased epithelial thickness on CET maps. No case demonstrated fluorescein staining outside the CET map–defined abnormal region. Statistical analysis revealed that the median lesion size measured by CFS was 5.84 mm² (interquartile range [IQR], 2.13–9.52; range, 0.55–24.02), whereas CET identified significantly larger lesion sizes, with a median of 12.61 mm² (IQR, 8.25–27.18; range, 1.35–44.96). The median difference of lesion size measured was 9.26 mm² (95% CI, 5.87–14.42; P < 0.0001) with a large effect size (d = 1.412; 95% CI, 0.697–2.127). The Bland–Altman plot revealed a mean bias of 10.04 mm² (95% limits of agreement, –3.89 to 23.96). Furthermore, Spearman's rank correlation analysis demonstrated a strong positive correlation between lesion sizes obtained from CFS and CET mapping (r = 0.904; P < 0.0001).

Figure 3.

Comparison of lesion areas measured from AS-OCT–based CET mapping and external eye photographs with CFS. (A) Scatterplot. The red dotted line represents the linear trendline (slope = 1.631; 95% CI, 1.170–2.092; P < 0.001). (B) Bland–Altman plot. The gray dotted line represents the linear trendline. (C) Paired plots with per-eye lines.

Sensitivity analysis demonstrated that the 60-µm threshold provided the strongest correlation with clinical findings. Although the 55-µm threshold also showed significant differences compared to CFS (P < 0.05), it appeared to over-delineate areas beyond the zone of epithelial instability. The 65-µm threshold proved too restrictive, failing to identify the full extent of the loose epithelium.

Comparison of Involved Corneal Subdivisions Identified by CET Mapping and CFS

We further analyzed and compared the location of RCE lesions in both CET mapping and CFS images (Fig. 4). CET mapping detected more lesions in the mid-periphery and periphery zones, regardless of etiology. This was particularly evident in the inferotemporal peripheral subdivision for cases with traumatic etiology, where CET mapping was able to identify lesion involvement in four cases when CFS detected none. Inferior corneal involvement was more prevalent than superior involvement across all etiologies. Spatial analysis revealed that CET mapping consistently identified more extensive epithelial involvement across the 25 predefined corneal sectors. On average, CET mapping identified 4.5 ± 3.7 additional sectors per eye that were characterized as normal by CFS (CET mapping vs. CFS: 11.7 ± 6.3 vs. 7.3 ± 4.5 sectors; P < 0.001). This quantitative extension was most prominent in the paracentral and superior zones, providing a preliminary clue that subclinical disease often extends well beyond the area of active staining. Specifically, for the 10 cases where CFS revealed involvement only in the inferior half of the cornea, five of them (50%) were found to have superior extension of their lesions by CET mapping. This spatial reclassification highlights that a significant proportion of patients may have subclinical superior involvement that is not visible with conventional staining, providing a preliminary clinical clue that the true extent of epithelial instability often exceeds what is visible on CFS. By providing a more precise assessment of lesion involvement, these CET map findings could enhance the accuracy of treatment planning.

Figure 4.

Distribution of lesion areas defined by CFS and AS-OCT–based CET mapping among all patients, trauma-related patients, and non–trauma-related patients. The regions of involvement were divided into central (2 mm in diameter), paracentral (2–5 mm), mid-periphery (5–7 mm), and periphery (7–9 mm) zones across eight directions. The number of affected eyes in each region is represented using varying shades of color to indicate the prevalence of abnormal corneal epithelium. In both the trauma-related and non–trauma-related groups, a higher proportion of lesions were observed in the central or paracentral lower cornea. However, some lesions were also present in the upper cornea.

Treatment Responses After PTK

Of the 15 cases included in this study, nine patients subsequently underwent PTK. Keratometric data, surgical parameters, and clinical outcomes are summarized in Table 2. Upon the epithelial debridement procedure during PTK, we confirmed that the location of epithelial loosening was similar to the lesion shown on the CET. Laser ablation was applied to the entire area with poor epithelial–basement membrane adherence. After PTK, the operated eyes were covered with bandage contact lenses for 6 weeks, allowing complete re-epithelialization to take place. In the five cases where postoperative CET was obtained, smoothing of the corneal epithelium with normalized epithelial thickness was observed at a median of 7 months (range, 3–10) (Fig. 5). No recurrence of RCE, anterior stromal haze, or other adverse events were recorded in our cohort.

Table 2.

Clinical Data of Patients Receiving PTK Treatment

Case No.	Pre-PTK BCVA	Post-PTK BCVA	Pre-PTK K Value (D)	Post-PTK K Value (D)	Change in K Value (D)	Total Ablation Depth (µm)	Ablation Zone Diameter (mm)	Post-PTK CET	Follow-Up Duration (mo)	Outcome and Adverse Events
2	NA	0.5	NA	43.50/44.25	NA	35	6.5	Smooth at 9 mo	10	No recurrence, no haze, no AE
3	0.9	1.0	42.25/45.25	42.50/45.50	+0.25	23	6.5	Smooth at 10 mo	12	No recurrence, no haze, no AE
4	1.0	1.0	41.75/42.50	41.75/42.50	+0.00	30	6.5	NA	3	No recurrence, no haze, no AE
6	0.4	0.9	46.00/49.25	47.50/47.75	+0.00	25	6.5	Smooth at 7 mo	7	No recurrence, no haze, no AE
8	0.9	1.0	39.25/39.75	38.50/39.50	−0.50	25	6.5	NA	5	No recurrence, no haze, no AE
9	0.6	1.0	43.25/43.75	42.25/43.25	−0.75	25	6.5	NA	11	No recurrence, no haze, no AE
10	NA	NA	43.00/43.50	NA	NA	23	6.5	NA	1	Lost to follow up after 1 mo
11	0.4	0.8	42.00/43.75	43.75/44.75	+1.38	25	6.5	Smooth at 3 mo	6	No recurrence, no haze, no AE
15	0.3	1.0	NA	45.75/47.75	NA	20	6.5	Smooth at 4 mo	5	No recurrence, no haze, no AE

AE, adverse event; BCVA, best-corrected visual acuity; K, keratometry; NA, not available.

Figure 5.

Comparison of AS-OCT–based CET mapping in representative patients with RCE treated with PTK. The CET images from three representative patients (case 3, case 8, and case 11) were taken before and after they underwent PTK. Prior to PTK, the CET mapping images clearly showed areas of epithelial loosening, which resolved after the procedures, with CET returning to normal. Of the nine patients treated, five showed normalization of CET at a median of 7 months, whereas the remaining were lost to follow-up or did not receive post-PTK CET exams. None of the patients experienced a recurrence of RCE in the follow-up period.

Discussion

RCE is a condition characterized by repeated episodes of corneal epithelial breakdown, leading to pain, photophobia, tearing, and visual disturbances. Its etiology often involves trauma to the corneal epithelium or underlying basement membrane dystrophies, such as EBMD. Anatomically, RCE is marked by poor adhesion of the epithelium to the basement membrane, often accompanied by microcysts, fingerprints, or map-dot patterns in EBMD. Management strategies include conservative treatments, such as lubricants and hypertonic saline, which promote healing and prevent epithelial sloughing.^1,2 More advanced options include therapeutic contact lenses, needle anterior stromal puncture, YAG laser anterior stromal puncture, and PTK.^4,19,20 Needle anterior stromal puncture is simple and can be performed in an outpatient clinic; however, it is not easily applicable on a large area and may lead to corneal scarring. If the lesion is located in the central cornea, it could affect vision, and the recurrence rate is relatively high.¹ YAG laser anterior stromal puncture is minimally invasive but carries a moderate recurrence risk. It is also difficult to apply on a large area.²¹ PTK offers high success rates by remodeling the basement membrane but may lead to haze and is more costly. Despite effective interventions, RCE recurrence rates vary widely, ranging from 13% to 36%,¹ emphasizing the importance of accurately assessing the extent of the lesion to avoid undertreatment and to select the appropriate treatment method based on the size and extent of the lesion.

AS-OCT has shown excellent results in diagnosing various corneal epithelial diseases.¹² Its benefits include being non-invasive, providing quick examinations, allowing comparisons between both eyes or over different time points, and offering both en face and cross-sectional views. AS-OCT also provides quantitative measurements across different corneal regions with high resolution, strong repeatability, and a large detection range.

AS-OCT has become an important tool in the diagnosis and management of corneal diseases, particularly ocular surface diseases, keratoconus, and other ectatic disorders.^7,8 This non-contact imaging modality allows for a detailed assessment of the corneal epithelium across a wide area—extending beyond 9 to 10 mm in diameter—with high resolution, revealing structural changes that may indicate lesion areas. Unlike slit-lamp biomicroscopy, which cannot provide high-resolution corneal images, or in vivo confocal microscopy, traditionally used for detecting corneal epithelial abnormalities but limited to small areas and requiring direct contact with the cornea,^8,13 AS-OCT offers a superior diagnostic solution. In recent years, advancements in AS-OCT technology have led to the development of increasingly high-resolution models,⁷ further enhancing its applications in the diagnosis and management of corneal diseases. This progress has solidified the potential of AS-OCT as a promising tool in the field of corneal imaging.

Our study builds upon the foundational work of Diez-Feijóo and Duran,¹⁴ who first characterized structural AS-OCT findings in RCE, such as epithelial edema and anterior stromal hyperreflectivity. Although their work established the diagnostic utility of cross-sectional imaging, our study extended this by utilizing en face CET mapping to quantify the total area of involvement. This novel approach aligns with the findings of Levy et al.,⁸ who demonstrated that CET mapping is highly sensitive for detecting localized epithelial thickening in various ocular surface disorders, often before they become clinically apparent under slit-lamp examination. Transitioning from purely diagnostic observation to quantitative area mapping provides a more objective basis for clinical decision-making and surgical planning in RCE management.

In this study, the threshold of 60 µm was not arbitrarily chosen but was based on the manufacturer's preset color scale of the Optovue epithelial thickness map, which corresponds to the upper limit of normal epithelial thickness reported in healthy corneas (typically 52–55 µm centrally).^8,16,17 Although the Optovue AS-OCT epithelial thickness map is generated from a 200 × 200 matrix (40,000 data points), there is no available method to modify the color scale based directly on numerical values, which further limits the feasibility of defining abnormal regions purely by manual numerical thresholding. Therefore, the 60-µm threshold is a practical easy-to-read cut-point for clinical use. Future studies using newer generations OCT or those incorporating artificial intelligence may benefit from more detailed delineation and provide a better subject-specific cut-point for RCE lesion detection.

A key finding in our study is the significant discrepancy between lesion sizes identified by CET mapping and CFS, with CET mapping consistently identifying larger areas of pathology. This discrepancy likely relates to the detection of subclinical epithelial edema or uneven hydration rather than frank epithelial adhesion defects alone. In RCE, the primary pathology involves abnormal adhesion between the epithelium and the underlying basal lamina. We hypothesize that, in areas adjacent to active erosions, the epithelium may already be loosened or poorly anchored, leading to localized fluid accumulation and subsequent thickening on the CET map. Because these areas of subclinical edema have not yet resulted in a complete break in the epithelial surface, they do not take up fluorescein stain, and their thickness change may be too subtle to be identified as negative staining. Consequently, CET mapping provides a more comprehensive view of the “at-risk” epithelium, revealing a broader field of basement membrane instability that necessitates treatment to prevent future recurrence.

In this study, we also found differences between the two methods in identifying the spatial distributions of the abnormal corneal epithelium. Whereas CFS primarily showed involvement in the central and lower corneal regions, CET mapping detected more extensive lesions covering the upper regions of the cornea (Fig. 4). This pattern was consistent across both groups of disease entities, including trauma-related and non–trauma-related RCE in this study. Clinically, this is highly significant, as the location and extent of the lesions can substantially influence the choice of treatment approach and scope. For example, smaller lesions that do not involve the central cornea are more suitable for needle or YAG laser anterior stromal puncture. Identification of a more extensive epithelial involvement on CET mapping may favor surface-regularizing treatments such as PTK, which can address broader epithelial–stromal interface pathology. Conversely, when CET mapping demonstrates a localized lesion corresponding closely to clinical findings, less invasive or focal treatments may be considered appropriate. In this context, CET mapping may serve as an adjunctive imaging tool to support more individualized treatment planning, potentially reducing undertreatment related to lesion underestimation by slit-lamp examination alone.

PTK enables effective treatment of RCE involving extensive corneal areas. Our results confirm the effectiveness of PTK in managing RCE, as no recurrences were observed among the patients who underwent the procedure. The absence of recurrence and the normalization of CET map findings post-treatment underscore the importance of accurate lesion identification and proper treatment selection in effectively managing RCE. It is noteworthy that, in post-PTK eyes, the detection of the epithelium may be interfered with by the ablated anterior stroma, resulting in delineation error and artifacts. Further longitudinal studies regarding epithelial recovery after PTK may aid the understanding of the dynamic remodeling process of the corneal surface and help establish objective, time-dependent CET mapping criteria to distinguish physiological healing from true pathological recurrence.

While CET proves to be an advanced diagnostic tool in identifying abnormal corneal epithelial areas, it is not without limitations. The accessibility of AS-OCT may be limited in certain clinical settings due to financial reasons. The system provides a high axial resolution of approximately 5 µm, yet measurement precision can still be influenced by technical factors. Similar to CFS, CET map measurements of the corneal surface are influenced by the moisture level of the corneal surface and the adequacy of blinking. Ensuring that the tear film is sufficient during the examination and that blinking is adequate prior to the procedure will help obtain accurate and reliable results.

Although this study presents significant findings, confirming that CET mapping may detect lesion areas more accurately than CFS and aid in selecting appropriate treatment strategies, there are several limitations to our research. The retrospective nature and single-center design may have introduced selection bias and limited generalizability. The sample size was relatively small, and there was a lack of normative control. The lesion areas identified by the two different examination methods may have varied depending on the examiner's interpretation and manual, threshold-based CET segmentation, highlighting the need for a more precise method to define the boundaries between normal and abnormal corneal epithelial areas. Additionally, the follow-up period was insufficient to evaluate long-term outcomes. Future studies with a larger sample size, prospective cohort design, and longer follow-up periods, as well as the development of more objective and standardized statistical methods for defining corneal lesion areas utilizing newer generation or artificial intelligence-driven OCT, could provide more robust and reliable results.

Conclusions

Our study demonstrated that CET mapping provides a potentially more accurate and detailed assessment of RCE compared to CFS. CET mapping identifies larger lesion areas, particularly detecting abnormalities in the central and upper cornea that CFS may overlook. Based on these findings, we recommend incorporating CET mapping into the clinical workflow for pre-procedural planning. The findings support the use of PTK as an effective treatment option in most cases, reducing recurrence risk. Further prospective longitudinal cohort studies may be needed to validate these findings and to assess the feasibility of clinical decision making based on CET mapping.