Central Cornea Changes on Anterior Segment OCT and In Vivo Confocal Microscopy After Autologous Limbal Epithelial Cell Transplantation

Abstract

Purpose:

To describe 1-year changes in the cornea as assessed by anterior segment optical coherence tomography (AS-OCT) and in vivo confocal microscopy (IVCM) for participants receiving a tissue graft generated from a new manufacturing process using cultivated autologous limbal epithelial cells.

Methods:

Cultivated autologous limbal epithelial cell grafts were produced in a 2-stage manufacturing process following a good manufacturing process-compliant protocol. AS-OCT and IVCM were completed at baseline and 12 months after the treatment in subsets of these participants. Secondary efficacy outcomes were determined based on improvement of central corneal epithelial morphology and thickness [corneal epithelial thickness (CET)] and presence of conjunctival or corneal cells in central cornea.

Results:

Among 14 participants, 13 (93%) were male, 12 (86%) were white, the mean age was 46 ± 16 years. At baseline, CET was 53 (range: 34, 64) microns, and epithelial basal cell density was 3964 (range: 822–5788) cells/mm²; the ratio of the cells at central cornea was 20% corneal and 90% conjunctival epithelial cells. At 12 months, the mean changes were 3 μm in CET (P = 0.67), and 1967 cells/mm² in epithelial basal cell density (P = 0.02); the proportion of the central cells improved to 75% corneal and 38% conjunctival epithelial cells.

Conclusions:

The AS-OCT and IVCM findings are consistent with the clinical improvement, indicating the reconstitution of the corneal phenotype and clearing of the optical axis. Nevertheless, IVCM is notably more effective for in-depth analysis of the epithelial phenotype and thickness than AS-OCT.

Limbal stem cell deficiency (LSCD) is characterized by dysfunctional limbal cells, resulting in alterations to the corneal surface, including conjunctivalization, neovascularization, inflammation, and scarring.^1–6 In cases of complete LSCD, the corneal surface undergoes a full circumferential encroachment by conjunctival epithelial cells. This invasion is subsequently followed by the development of subepithelial fibrosis and fibrovascular pannus, culminating in visual impairment.⁷

The goal of LSCD management is to restore the health of the ocular surface by repairing the adjacent limbal segment and supporting the growth of healthy limbal epithelial stem cells.^8,9 In the early stages of LSCD, medical therapy often involves the use of lubricating eye drops and anti-inflammatory medications to alleviate symptoms like pain and photophobia.^1–4,10 In advanced stages, therapeutic interventions are directed toward restoring limbal functionality and may involve cell-based therapies that use transplantation of allogeneic limbal stem cells (LSCs) procured from donor corneal tissue or autologous LSCs from an unaffected contralateral eye of the patient.^8,9 There are a few surgical strategies, including direct grafting of autologous tissue [conjunctival limbal autograft¹ and simple limbal epithelial transplantation (SLET)¹¹] and grafting of ex vivo proliferated autologous tissue [cultivated limbal epithelial transplantation (CLET)].¹²

We previously described our novel and rigorously standardized CLET two-stage manufacturing technique termed cultivated autologous limbal epithelial cell transplantation [cultivated autologous limbal epithelial cells (CA-LEC)].^13,14 In the final stage of this process, limbal epithelial cells are expanded on denuded human amniotic membrane using GMP-compliant manufacturing in serum-free media that is free of xenogeneic feeder cells. The development and validation of the CALEC manufacturing process and the feasibility, safety, and efficacy outcomes of the phase I/II trial have been previously reported.^13,15

The primary efficacy outcomes in our phase I/II trial were based on improvement in corneal epithelial surface integrity (complete success) or improvement in extent of corneal vascularization and/or participant symptomatology (partial success). Of 14 CALEC recipients, 93% were a complete or partial success at 12 months. Previous studies also show the utility of anterior segment optical coherence tomography (AS-OCT) and in vivo confocal microscopy (IVCM) in evaluating, diagnosing, and comparing management strategies for LSCD.^9,16–18 This paper evaluates changes in the central cornea on AS-OCT and IVCM parameters as secondary efficacy outcomes in participants with LSCD, 12 months after CALEC transplantation.

METHODS

Trial Conduct and Oversight

Funding was provided by a grant from the National Eye Institute of the National Institutes of Health, which participated in main study oversight but not directly in this imaging study design, data collection, study management, analysis, or interpretation of the data. Participant enrollment, biopsy, and transplant were completed at a single clinical center, Massachusetts Eye and Ear. The Jaeb Center for Health Research was the Coordinating Center responsible for the database, data validation, analyses, and trial coordination. Manufacturing procedures were performed at the Connell and O’Reilly Families Cell Manipulation Core Facility, Dana-Farber Cancer Institute.

The protocol adhered to the tenets of the Declaration of Helsinki and the Health Insurance Portability and Accountability Act and was approved by the Massachusetts Eye and Ear Institutional Review Board. Informed consent was obtained from all participants. The US Food and Drug Administration approved an Investigational New Drug Application (#16102), and the trial is registered on clinicaltrials.gov as NCT02592330.

Trial Design and Participants

Fourteen participants from 18 to 90 years of age with unilateral LSCD (as determined by conjunctivalization of the cornea defined by fibrovascular pannus more than 2 mm from the limbus into the cornea for ≥6 clock hours) completed a CALEC transplant. The complete eligibility criteria and the production of cellular grafts were described in the prior publications.^13,15

A comprehensive eye exam, including best-corrected vision acuity and slit-lamp examination, was performed by one of 2 independent clinical examiners (L.J. and J.Y.) who were not involved in the CALEC procedure. Slit-lamp pictures, AS-OCT images, and IVCM images were obtained by a certified ophthalmic technician at baseline and 12 months after the treatment. IVCM was deferred if applanation microscopy was clinically contraindicated (pain, presence of epithelial breakdown/defect, or bandage contact lens).

In Vivo Confocal Microcopy

Laser IVCM (Heidelberg Retina Tomograph 3) was conducted with the Rostock Cornea Module (Heidelberg Engineering GmbH, Heidelberg, Germany). Images were obtained with the help of a 63 × objective immersion lens with a numerical aperture of 0.9 (Olympus, Tokyo, Japan). This microscope uses a 670-nm red wavelength diode laser source, producing images representing a coronal section of the cornea of 400 × 400 μm (horizontal × vertical). The cornea module was mounted with a disposable, sterile polymethylmethacrylate cap (Tomo-Cap; Heidelberg Engineering GmbH), filled with a layer of hydroxypropyl methylcellulose 2.5% (GenTeal gel; Alcon), the gel was also applied to the surface of the cap. The epithelial layer images (50–100 images) were obtained via 6 to 8 sequence from central cornea and each quadrant of the limbus (superior, nasal, temporal, and inferior). Masked observers [L.Y.S, A.R.K., (third—in case more than 10% discrepancy between first 2 observers) L.K.J.] selected the 3 most representative images as best focused, single layer, minimum folds, and good contrast of the epithelial layers. Images were analyzed to determine the presence or absence of the following cell types or structures. In the central cornea, the presence of corneal epithelia was defined by polygonal cells with bright, well-defined borders, dark cytoplasm, and no visible nuclei. The cells with prominent nuclei and ill-defined borders, and cells with bright borders, dark cytoplasm, and no visible nuclei were determined as conjunctivalization at central cornea. In addition, the presence of goblet cells and cystic formations are further indicators of conjunctivalization. The cells of the limbus were defined as either epithelial cells as previously described or conjunctival cells where the cells have bright nuclei with ill-defined borders.¹⁹ The limbal images were evaluated for the presence or absence of palisades of Vogt, which appear as hyperreflective double contoured linear structures that alternate with islands of epithelial cells.

AS-OCT

The scans were performed using an Avanti xR optical coherence tomography device (Optovue Inc. Fremont, CA) following manufacturer instructions in the “AS-OCT” modes using the Line, Cross-Line with a scan length of up to 8 mm and an axial resolution of 5 μm, Pachymetry, and PachymetryWide. Manual thickness measurements were done in Line mode using the proprietary Avanti xR OCT program for corneal stromal and fibrovascular pannus of the affected cornea. Central epithelial thickness was calculated with the method previously described by Deng et al²⁰; the procedure used cross-line scans for the central cornea and single-line scans for the limbal quadrants, capturing images with the scleral spur visible. Inadequate images were retaken or discarded. Each area underwent at least 3 scans. Central epithelial thickness was measured as the distance from the tear film to the basement membrane, using both a single-point center method and a three-point method around the center. In severe LSCD cases with subepithelial fibrosis, the epithelial layer was indistinguishable, leading to a default corneal epithelial thickness (CET) value of 0 μm.²⁰

Statistical Methods

The secondary efficacy outcomes evaluated were change in central CET, change in deepest limbal epithelial thickness (LET) in each quadrant (superior, temporal, inferior, nasal), change in central corneal epithelial basal cell density (EBCD), presence of conjunctival and corneal cells in the central cornea, and palisades of Vogt in each quadrant (Table 1). Only eyes with both baseline and month 12 measurements were included in the calculation of mean differences, and the paired t test was used to evaluate the differences. Associations among change in imaging measures with change in other vision measures were assessed using Spearman correlation coefficients. Histograms were used for comparison of imaging measures per participant. Boxplots were used to summarize change in imaging measures between baseline and the 12-month visit. Analyses were conducted using SAS/STAT software version 9.4 (SAS Institute, Cary, NC).

TABLE 1.

Changes From Baseline to 12 Months in Characteristics on Anterior Segment OCT and IVCM

	Baseline	12-Month	Mean difference* [95% CI] (P)
Anterior segment-OCT results†
Central corneal epithelial thickness (microns)—mean (SD), median [range] (N = 6)	51 (10) 53 [34–64]	48 (17) 50 [22–79]	3 [−16 to 23] (0.67)
Deepest limbal epithelial thickness (microns)—mean (SD), median [range]
Superior (N = 5)	73 (16), 77 [46–89]	66 (19), 64 [34–89]	−14 [−47 to 20] (0.18)
Temporal (N = 7)	78 (17), 75 [52–98]	79 (30), 74 [37–147]	−13 [−33 to 8] (0.18)
Inferior (N = 5)	107 (25), 95 [89–148]	92 (43), 76 [49–200]	−17 [−53 to 18] (0.15)
Nasal (N = 6)	85 (31), 79 [52–135]	79 (27), 74 [46–150]	−6 [−60 to 47] (0.76)
In vivo confocal microscopy results‡
Central epithelial basal cell density (cells/mm2)—mean (SD), median [range]	N = 7	N = 6	1967 [564 to 3369] (0.02)
	3316 (2000) 3964 [822–5788]	4871 (1603) 5192 [2805–7015]
Presence of corneal and conjunctival cells and palisades of Vogt	N = 10	N = 9	P
Central cornea—n (%)
Corneal epithelial cells	2 (20)	6 (75)	0.25
Conjunctival cells	9 (90)	3 (38)	0.13
Superior limbus —n (%)
Corneal epithelial cells	3 (33)	6 (67)	0.25
Conjunctival cells	9 (100)	8 (89)	§
Palisades of Vogt	1 (11%)	3 (33%)	‖
Temporal limbus—n (%)
Corneal epithelial cells	3 (30)	7 (100)	§
Conjunctival cells	10 (100)	7 (100)	§
Palisades of Vogt	3 (30)	2 (29)	0.50
Inferior limbus—n (%)
Corneal epithelial cells	3 (30)	6 (75)	0.50
Conjunctival cells	10 (100)	6 (75)	§
Palisades of Vogt	1 (10)	1 (13)	§
Nasal limbus—n (%)
Corneal epithelial cells	3 (33)	7 (88)	1.00
Conjunctival cells	9 (100)	7 (88)	§
Palisades of Vogt	2 (22)	1 (13)	1.00

^*Mean difference is based only on eyes that had measurements at both baseline and 12 months.

^†Data unavailable for some participants because of poor imaging quality (because of severity of LSC disease and thick pannus formation, some OCT images were not quantifiable).

^‡Participants with missing data because of poor image quality (because of severity of LSC disease and thick pannus formation, some confocal images were not quantifiable) or an inability to obtain the image are excluded from the denominators in the calculations of reported percentages.

^§Insufficient data to calculate P-values using McNemar exact test.

^‖No discordant pairs available to calculate P-values using McNemar exact test.

RESULTS

Detailed demographics and primary results were reported previously.¹⁵ In brief: Of the 14 participants receiving CALEC transplant, 93% were male, 86% were of White ethnicity, and mean age at enrollment was 46 years (range: 24–78). The primary underlying cause of LSCD was chemical burn [9 (64%)] or thermal burn [2 (14%)].

At the outset of the study, 71% of participants exhibited no epithelial defects, indicating the varied manifestation of LSCD. In 4 cases where an epithelial defect was present, the average size was 11.1 mm². The mean neovascular area constituted 7.8% of the total area, with the central 5-mm zone being affected in 86% of these participants. Furthermore, all participants experienced a reduction in epithelial defect size by 90% or more, whereas 71% saw a reduction in neovascular area by at least 60%.

IVCM

EBCD improved from median 3964 (range 822–5788) cells/mm² at baseline to median 5192 (range 2805–7015) cells/mm² at 12 months, a mean change of 1967 [95% CI: 565, 3369 (P = 0.02)] cells/mm². The ratio of corneal epithelial cell at central cornea increased from 20% at baseline to 75% at 12 months, and ratio of conjunctival epithelial cell at central cornea decreased from 90% at baseline to 38% at 12 months. At baseline, in each of the nasal, inferior, temporal, and superior limbal quadrants, 100% of participants had conjunctival epithelial morphology. Corneal epithelial cell morphology was present in the periphery in 30% to 33% of participants (Table 1, Fig. 1). In all quadrants of limbus, the presence of corneal epithelial cells improved to 60% or more at 12 months (Table 1, Fig. 1). There were no age-associated effects for changes in EBCD from baseline to 12 months.

FIGURE 1.

Demonstration of clinical improvement with slit-lamp microscopy pictures, IVCM, and AS-OCT of central cornea and limbus. (A, slit-lamp microcopy picture of the patients with LSCD at baseline and (F) 12th month after CALEC treatment. B, IVCM of the central cornea at baseline and (G) at 12th month after CALEC treatment. C, IVCM of the limbus showing the destruction of healthy limbal structure and fibrosis at baseline and (H) those showing the improvement of fibrosis and healthy transition of the limbal structure at 12th month after CALEC treatment. D, AS-OCT of central cornea of the patients with LSCD at baseline and (I) 12th month after CALEC treatment. E, AS-OCT of the limbus showing the heterogenous appearance of limbal structure and fibrosis at baseline and (J) and healthier limbal structure at 12th month after CALEC treatment.

AS-OCT

Median CET was 53 (range: 34–64) microns at baseline and 50 (range: 22–79) microns at 12 months postoperatively, with a mean change of 3 [95% CI: −16 to 23 (P = 0.67)] microns (Table 1, Fig. 2). Sample sizes for assessing correlations between changes in AS-OCT and IVCM were low (N = 3–5) because of missing measurement values at baseline or follow-up. No significant associations were observed between the changes measured by AS-OCT and IVCM across various parameters, including central epithelial corneal thickness (r = −0.50, P = 0.67), temporal LET (r = −0.40, P = 0.60), inferior LET (r = −0.50, P = 0.67), and nasal LET (r = −0.60, P = 0.28). Superior LET was associated with changes in IVCM (r = 1.00, P < 0.001) and showed age-associated differences between visits (P = 0.03).

FIGURE 2.

Alterations of IVCM and AS-OCT of central cornea at baseline and 12th month of CALEC treatment. A–C, The changes in central epithelial thickness of patients with LSC deficiency measured with optic coherence tomography. D–F, The changes (1967 [95% CI: 565–3369 (P = 0.02)] cells/mm²) in central EBCD of patients with LSC deficiency measured with IVCM.

DISCUSSION

In this study, we sought to describe the alterations in the central cornea and limbus after 1 year using AS-OCT and IVCM to support the efficacy of a new surgical method, CALEC treatment, which was previously published.^13,14 Given that the central cornea serves as a marker for the overall activity of LSCs,^20,21 this study primarily concentrated on observing the central corneal alterations preceding and following CALEC therapy. The increased corneal epithelial morphology (with decrease in conjunctival epithelial cell morphology) and improved EBCD on IVCM corroborate the clinical efficacy of CALEC in participants with LSCD.^13,14

Recent studies have underscored the pivotal roles of both OCT and IVCM techniques in advancing our comprehension of structural and cellular alterations in individuals with LSCD and the transformations following various LSCD treatment modalities.^{19,20,22–25} Moreover, the utility for ancillary imaging in the diagnostic approach to LSCD has been emphasized by the International LSC Deficiency Working Group.¹ The complementary nature of these imaging modalities allows for a more comprehensive evaluation of the corneal status, including structural and functional aspects.^7,23–28 However, it is essential to recognize that both imaging techniques have limitations, and their utility in LSCD management may vary depending on the specific objectives of the evaluation.^{20,22–24,26,28–32}

The noninvasive and high-definition imaging features of OCT are especially advantageous for measuring corneal thickness, detecting changes in the epithelial and stromal layers, and assessing the health of the limbal structures in LSCD cases.^{19,20,22–24,33} Recent publications showed an observed average reduction in CET in cases of LSCD; Mehtani et al found a 22.2% decrease in CET in LSCD cases, whereas Chan et al noted a 20.2% decrease reflecting the adverse effect of a reduced number of functioning LSCs on epithelial renewal. In severe forms of LSCD, the tissue of the corneal pannus exhibits squamous metaplasia or subepithelial fibrosis, which is characterized by abnormal cell proliferation, increased layering situated above the Bowman layer, and identified as hyperreflective on AS-OCT.^23,34 The presence of hyperreflective subepithelial fibrosis, often seen in severe cases of LSCD, can complicate the precise assessment of CET and LET.^23,24 In such instances, IVCM becomes an essential follow-up diagnostic tool to closely inspect the epithelial phenotype and measure its thickness.^20,21

OCT provides detailed cross-sectional images that facilitate quantitative analysis of surgical results, including CET measurements and detecting epithelial irregularities.³⁵ Historically, such data have been essential for tracking the efficacy of interventions like conjunctival limbal autograft, SLET, and CLET in the reconstruction of corneal architecture.^35,36 Results from this study showed minimal change in CET (3 μm [95% CI: −16 to 23], P = 0.67) and patient by patient response was highly variable (Fig. 2). The findings in this study acknowledge that the variability in optical OCT assessments before and after treatment may be influenced by the presence of hyperreflective subepithelial fibrosis, a common complication in severe LSCD.^23,34 This condition impedes the accurate measurement of CET and LET, necessitating the use of IVCM for a more detailed examination of the epithelial phenotype, structure, and its thickness. Although AS-OCT offers greater accessibility and simplicity of operation compared with IVCM, the results were more variable compared with IVCM as we observed in our study (Fig. 2).

IVCM provides a live, dynamic perspective on cellular activity and has previously been used in many studies.^37–47 It is instrumental in analyzing LSC density, the morphological features of limbal cells, and the regrowth of corneal nerves.^43,48 This technology grants critical information regarding the cornea’s functional recovery postsurgery. IVCM is especially useful for evaluating the sustained outcomes of treatments for LSCD, offering a window into the regenerated limbal epithelium’s health and vigor.⁴⁹ Wang et al demonstrated healthy epithelial regeneration with IVCM and impression cytology after SLET treatment, and they concluded that IVCM and impression cytology may reflect more accurately the phenotype of the epithelium; thus, both are considered more sensitive in diagnosis. In this study, central EBCD was improved by 1967 [95% CI: 564–3369 (P = 0.02)] cells/mm² at postoperative month 12 suggesting the success of the CALEC procedure in promoting stem cell regeneration. IVCM also showed an improvement in the presence of corneal epithelial cells, increasing to 75%, which is a positive sign of corneal healing. Conversely, conjunctival epithelial cells decreased to 38% in the central cornea, indicating the successful clearance of the conjunctival overgrowth from the optical axis and regression of the pannus.

In a study employing allogeneic bone marrow–derived mesenchymal stem cell transplantation and allogeneic CLET, initial central corneal assessments via IVCM preoperatively identified 15 eyes with a conjunctival epithelial phenotype and 8 with a mixed epithelial phenotype.^43,50 Notably, after 12 months, there was a significant shift in epithelial phenotype distribution, reflecting substantial improvements in both treatment groups. Specifically, 71.4% of mesenchymal stem cell transplantation cases and 66.7% of CLET cases demonstrated enhanced epithelial phenotype alignment with expected postoperative outcomes, with a statistical correlation of P = 1.000, paralleling the results observed in our study. This study demonstrates several strengths and limitations. The strengths include its longitudinal design, utilization of both AS-OCT and IVCM for comprehensive corneal assessment, and the use of a standardized manufacturing technique, CALEC, for LSCD treatment. The study population consisted of participants with varying LSCD severity, adding to the generalizability of the results. However, some limitations should be considered. The sample size for AS-OCT and IVCM assessments was small, potentially limiting the generalizability of the findings. In addition, confocal imaging was not able to be obtained in some participants because of contraindication, and this could have resulted in a selection bias. The lack of a control group makes it challenging to attribute observed changes solely to CALEC treatment. Further research with larger cohorts and controlled designs is needed to validate these findings and establish the long-term safety and efficacy of CALEC for LSCD management. Despite these limitations, this study provides valuable insights into the use of imaging modalities in evaluating LSCD treatment outcomes and underscores the potential benefits of CALEC therapy.

In summary, OCT and IVCM are valuable in evaluating and benchmarking surgical interventions for LSCD. The synthesis of OCT’s high-resolution anatomical insights and IVCM’s dynamic cellular imaging affords a comprehensive method for gauging surgical success. In our study, IVCM stands out as the superior imaging modality for a detailed assessment of epithelial phenotype and thickness, offering enhanced accuracy and precision that is critical in such complex cases. Continued research is anticipated to hone these imaging modalities, enhancing patient prognosis and contributing to a nuanced understanding of corneal regeneration dynamics in LSCD.

In this study, by comparing the outcomes assessed through OCT and IVCM, we were able to gain a comprehensive understanding of the effectiveness of the newly developed technique, CALEC, in aspects of change in CET and basal cell density, improvement of corneal pannus, and clearance of optic axis (Figs. 1 and 2; Table 1). The data herein indicate that the CALEC procedure is a promising and efficacious modality for LSCD management, evidenced by marked gains in clinical parameters confirmed with AS-OCT and IVCM. Although additional studies are imperative to substantiate these preliminary findings, this research underscores the utility of OCT and IVCM in empirically evaluating surgical outcomes, propelling the knowledge base and therapeutic strategies for LSCD.