Quantitative Assessment of Lid Margin Vascularity Using Swept-Source Optical Coherence Tomography Angiography

Abstract

Purpose: To evaluate the ability of swept-source optical coherence tomography angiography (SS-OCTA) to assess lid margin vascularity.

Methods: This prospective, cross-sectional trial enrolled 125 participants, including 15 control subjects and 110 meibomian gland dysfunction (MGD) patients. Lid margin blood flow density (LMBFD) was obtained using SS-OCTA. LMBFD was assessed for repeatability in 54 of 125 participants and for reproducibility in 23 of 125 participants. The efficacy of LMBFD was validated in the 125 participants, who were divided into mild (n = 46), moderate (n = 42), and severe groups (n = 37) according to the lid margin vascularity severity shown in the slit-lamp photographs. Correlations between LMBFD and MG-related parameters, such as ocular surface disease index (OSDI), fluorescein tear break-up time (FTBUT), cornea fluorescein staining (CFS), lid margin score (LMS), and meibomian gland expressibility (ME), were analyzed in all 125 participants.

Results: Repeatability and reproducibility coefficients were satisfactorily high in the scan mode with a scan area of 6 mm × 6 mm (intraclass correlation coefficient [ICC] repeatability = 0.905; ICC reproducibility = 0.986) and a scan area of 9 mm × 9 mm (ICC repeatability = 0.888; ICC reproducibility = 0.988). The LMBFD gradually increased in the mild, moderate, and severe groups (P < 0.001). LMBFD was significant correlated with OSDI (r = 0.290, P = 0.001), FTBUT (r = −0.195, P = 0.030), CFS (r = 0.352, P < 0.001), ME (r = 0.191, P = 0.033), and LMS (r = 0.370, P < 0.001).

Conclusions: LMBFD may be a noninvasive, repeatable, reproducible, and efficient index for the quantitative evaluation of eyelid margin vascularity in the future.

Translational Relevance: We demonstrated that SS-OCTA has the potential to evaluate the eyelid margin vascularity in MGD patients and guide future treatment strategies in clinics.

Introduction

Meibomian gland dysfunction (MGD) is the major cause of dry eye disease worldwide. MGD is characterized by alterations in the quality of MG secretion and abnormalities of the lid margin, including prominent telangiectatic blood vessels, lid margin irregularities, hyperplasia/metaplasia, and pouting of the MG orifices.¹ Lid margin abnormalities are detectable with a slit-lamp microscope and are important for the diagnosis of MGD or the determination of disease progression. Telangiectasia or vascularity at the lid margin is a key clinical sign associated with ocular surface inflammation and usually appears in more severe MGD.² The rise in the vascularization of the lid margin may also increase the oxygen content of the MG and lead to the deterioration of MG function.³ Some treatments to improve lid margin telangiectasia or neovascularization, such as intense pulsed light and anti-vascular endothelial growth factor (anti-VEGF) agent injection or topical application, have been confirmed to be effective against MGD.^4,5 However, assessment of the efficacy of treatment requires precise evaluation of changes in lid margin telangiectasia. Currently, lid margin vascularity is mostly qualitatively or hierarchically evaluated based on slit-lamp microscopy.^6,7 Tantipat et al.⁸ used image analysis software to morphometrically analyze the area of lid margin telangiectasia in digital photographs, but this quantitative analysis requires tedious manual work and is not suitable for clinical application. 

Swept-source optical coherence tomography angiography (SS-OCTA) is a recent and improved noninvasive approach utilizing equipment that offers a longer wavelength, stronger penetration, and clearer vessel imaging. SS-OCTA can be used for three-dimensional imaging and quantitative analysis of the cornea, conjunctiva, iris, retinal, and choroidal vessels through optimization algorithms.^9,10 Thus, we performed a pilot study to explore an optimized scan technique for the detection of lid margin vasculature and to investigate whether the SS-OCTA tool has the potential to characterize lid margin vascularity in MGD patients. The repeatability and validity of OCTA in assessing lid margin blood flow density (LMBFD) were evaluated in this study. 

Methods

Study Design and Subjects

This prospective, cross-sectional study was conducted at Wenzhou Medical University Eye Hospital. The study adhered to the tenets of the Declaration of Helsinki and was approved by the Research Ethics Committee of the Eye Hospital, Wenzhou Medical University (Ethics No. H2023-044-K-41-01). All subjects provided written informed consent before entry into the study. The study was registered at ClinicalTrials.gov (NCT06091722). 

Patients with MGD and normal subjects were enrolled from March 1 to May 15, 2023. The MGD patients were ≥20 years old and were diagnosed according to Chinese MGD diagnostic criteria,¹¹ including dry eye symptoms and at least one of the following signs: (1) poor meibum expressibility or quality, or (2) at least one eyelid margin abnormality (lid margin irregularity, plugging of the gland orifice, vascular engorgement/telangiectasia, or anterior replacement of the mucocutaneous junction). Normal subjects were defined as adults without any dry eye symptoms or lid margin abnormalities. The exclusion criteria were as follows: (1) subjects with ocular diseases or conditions known to affect the anatomy of the anterior segment, such as acute ocular inflammation, a history of eyelid surgery, contact lens wear within 2 weeks, and/or eyelid trauma; (2) subjects with a history of systemic or topical medications that might cause ocular congestion; (3) subjects with severe systemic illness or pregnancy; or (4) subjects who had difficulty cooperating with the examination, resulting in the obtained images not being sufficiently clear for automatic analysis. Only the right eyes of all subjects were included. 

Clinical Assessment

Ocular surface assessments were performed from least to most invasive as follows: fluorescein tear breakup time (FTBUT), cornea fluorescein staining (CFS),¹² meibomian gland expressibility (ME), meibum quality (MQ), and ocular surface disease index (OSDI).¹³ Upper lid ME was assessed by finger based on five central glands using a four-point scale (0, all expressed; 1, three or four expressed; 2, one or two expressed; 3, no glands expressible).¹⁴ MQ was also recorded on a 4-point scale (0, clear; 1, cloudy; 2, granular; 3, toothpaste), and an average score of the 8 upper lid central glands was recorded.¹⁴ Tear meniscus height (TMH) and images of upper MGs were measured with a Keratograph 5M (OCULUS, Wetzlar, Germany). The upper eyelid meiboscore ranged from 0 to 3 (0, no loss; 1, ≤1/3 deletion; 2, 1/3–2/3 deletion; 3, >2/3 deletion).¹⁵ The lid margin score (LMS) including lid margin thickening, irregularity, gland orifice plugging, vascular engorgement/telangiectasia, and anterior or posterior replacement of the mucocutaneous junction, was assessed on a scale ranging from 0 to 5.¹⁴ 

Digital Photograph and Lid Margin Vascularity Grade Scales

The center of the upper eyelid margin was marked with a sterile skin marker pen, and then the central upper eyelid margin was photographed by high-definition photography under a slit-lamp microscope (DED-3H; Chongqing Kanghua Ruiming, Chongqing, China). The severity of lid margin vascularity shown in high-definition eyelid margin photographs was evaluated by two experienced clinicians (YF, MH). The evaluation criteria were defined based on the number of visible capillaries at the middle third of the upper lid margin: grade 0, no capillaries; grade 1, zero to five capillaries; grade 2, six to 10 capillaries; and grade 3, more than 10 capillaries (Fig. 1). 

Figure 1.

 

Images of lid margins with different severities of vascularity: (A) grade 0, no capillaries visible; (B) grade 1, zero to five capillaries visible; (C) grade 2, six to 10 capillaries visible; (D) grade 3, more than 10 capillaries visible.

All 125 participants were grouped according to the grade of capillaries at the lid margin assessed by two clinicians as follows: (1) mild group, for which the scores of both clinicians were ≤grade 1; (2) moderate group, for which one doctor scored the capillaries as grade 2 and the other scored them as 1 to 2; and (3) severe group, for which one doctor scored the capillaries as grade 3 and the other doctor scored them as ≥2. 

SS-OCTA Measurement of the Blood Flow Density of the Lid Margin

An SS-OCTA system (VG200D; Intalight [formerly SVision Imaging], Shanghai, China) was used in this study. OCTA works at a central wavelength of nearly 1050 nm with a scan speed of 200,000 A-scans per second, providing an acquired image resolution of 384 × 384 pixels. Each OCTA image is generated from the stacking of multiple tomography images obtained from three automated scans. The scan area tracking was not used in this measurement, as it greatly extended the measurement time and made it difficult for many participants to cooperate and to obtain the image successfully. To ensure the consistency of the scanning area, scanning and image analysis were carried out according to the following protocol, as shown in Figure 2. Each patient was asked to look slightly down and stare at an external fixed light below. The center of image scanning and the center of image analysis were located at the same meibomian gland orifice below the black mark in the middle of the lid margin (Fig. 2C). Scan modes for the anterior segment with scan areas of 6 mm × 6 mm or 9 mm × 9 mm were used to assess the blood flow of the upper eyelid margin. The analysis mode for the conjunctival vasculature was used to analyze the lid margin vascularity (Figs. 2A, 2B). The grid diagram showed the blood flow density in an area of 6 mm × 6 mm or 9 mm × 9 mm, with each small square representing the value in an area of 1 mm × 1 mm (Fig. 2E). Only the average value of five adjacent small squares (5 mm × 1 mm) around the same scan and analysis center was recorded as the final result (Fig. 2E). Images with a signal strength of 8 to 10 were considered valid for analysis. Unclear images due to motion artifacts and insufficient exposure of the eyelid margin were judged to be of poor image quality. All images of poor quality were excluded. All participants were measured by two examiners (CW, YL). 

Figure 2.

 

SS-OCTA measurement of the blood flow density of the upper lid margin in an MGD patient. (A) The lid margin vascularity shown in the slit-lamp photograph was severe. (B) The blood flow shown in the OCTA diagram was consistent with the blood vessel shown in (A) and (C). (C) The analysis center was located at the palpebral margin below the yellow asterisk, as shown in the OCTA enface images. (D) The measured depth of vascularity was approximately 500 µm in the OCT image. (E) Vascular density map with a 1 mm × 1 mm grid.

The intraexaminer repeatability was defined as the repeatability in three quantitative measurements of the same eye using the same device under both 6 mm × 6 mm and 9 mm × 9 mm scan areas by the same examiner (CW). The interexaminer reproducibility was defined as the agreement in measurements of three images under both 6 mm × 6 mm and 9 mm × 9 mm areas taken for the same eye using the same device between results analyzed by two different examiners (CW, YL). The validity of LMBFD for assessing the severity of the palpebral margin vessels was determined by comparing the consistency of the LMBFD severity under the 6 mm × 6 mm scan area with the grade of lid margin vascularity assessed through slit-lamp photography. 

Statistical Analysis

Statistical analysis was conducted using SPSS Statistics 27.0 (IBM, Chicago, IL). The normality of the data was tested using the Kolmogorov–Smirnov test. Depending on the data distribution, data were described by the median (P₂₅, P₇₅). All data in this study were non-normal variables. The coefficient of variation (COV) and intraclass correlation coefficient (ICC) were used to evaluate the repeatability and reproducibility. The Mann‒Whitney U test with Bonferroni correction was used for comparisons among the three groups classified according to vascular severity. Spearman correlation analysis was used to evaluate the LMBFD and age or ocular surface parameters. Two-tailed P < 0.05 was considered statistically significant. 

Results

Demographics of the Study Participants

Ultimately, 125 participants were recruited for this study. Of these, 54 participants (39 MGD patients and 15 normal subjects) were scheduled for repeatability analysis, and 23 of the 125 participants were scheduled for reproducibility analysis. All 125 participants (110 MGD patients and 15 normal subjects) were enrolled for validity analysis. Table 1 summarizes the demographic data of allparticipants. 

Table 1.

 

Patient Characteristics

 

Repeatability Analysis and Reproducibility Analysis

Table 2 summarizes the repeatability values of the LMBFD under scan mode for the anterior segment with an scan area of 6 mm × 6 mm or 9 mm × 9 mm and the reproducibility values between the two examiners. The scan mode with 6 mm × 6 mm area showed good to excellent repeatability (ICC ≥ 0.9), and 9 mm × 9 mm area showed good repeatability (ICC ≥ 0.75). The interexaminer reproducibility between the two examinations also showed high consistency (6 mm × 6 mm, ICC = 0.986; 9 mm × 9 mm, ICC = 0.988) (Table 2). 

Table 2.

 

Lid Margin Blood Flow Density (%) Repeatability

 

Validity Analysis

Based on the results of the repeatability and reproducibility analysis, we recruited another 71 patients with MGD so we could evaluate the effectiveness of the scan mode with 6 mm × 6 mm area in measuring LMBFD. Finally, the results for a total of 125 participants were analyzed for validity. The mean LMBFD was 78.29% (range, 18.67%–94.19%). A total of 125 participants were divided into three subgroups according to the severity of lid margin vascularity assessed by two clinicians: mild group (n = 46), moderate group (n = 42), and severe group (n = 37). The average LMBFD increased significantly with the severity of eyelid margin vascularity, with values of 63.08% (range, 18.67%–88.18%) in the mild group, 78.69% (range, 69.73%–85.71%) in the moderate group, and 86.98% (range, 78.32%–94.19%) in the severe group (Fig. 3). 

Figure 3.

 

The average blood flow density increased significantly with the severity of eyelid margin vascularity (**P < 0.001, Mann‒Whitney U tests with Bonferroni correction).

Correlation Between Blood Flow Density and Clinical Parameters

The correlation analysis showed that there was no significant correlation between the LMBFD and age (Spearman's r = 0.042, P = 0.640), TMH (Spearman's r = −0.158, P = 0.078), upper meiboscore (Spearman's r = 0.140, P = 0.119), or MQ (Spearman's r = 0.053, P = 0.560). However, the LMBFD was positively correlated with the OSDI (Spearman's r = 0.290, P = 0.001), CFS (Spearman's r = 0.352, P < 0.001), ME (Spearman's r = 0.191, P = 0.033), and LMS (Spearman's r = 0.370, P < 0.001) and was negatively correlated with FTBUT (Spearman's r = −0.195, P = 0.030) (Fig. 4). Analysis of each lid margin abnormality showed that LMBFD was positively correlated with lid margin thickening (Spearman's r = 0.189, P = 0.035), irregularity (Spearman's r = 0.172, P = 0.045), gland orifice plugging (Spearman's r = 0.232, P = 0.009), and vascular engorgement/telangiectasia (Spearman's r = 0.328, P < 0.001) but was not significantly correlated with anterior or posterior replacement of the mucocutaneous junction (Spearman's r = −0.005, P = 0.954). 

Figure 4.

 

Correlation analysis between LMBFD and the ocular surface parameters CFS, FTBUT, ME, and LMS.

Discussion

MGD can act either as an inflammatory condition, such as blepharitis, or as a non-inflammatory condition, such as atrophic MGD.¹⁶ Currently, there is a lack of objective and quantitative indicators that can be easily manipulated clinically to assess the severity of MG inflammation. Qazi et al.^17,18 used palpebral conjunctival immune cells density based on in vivo confocal microscopy and ImageJ software (National Institutes of Health, Bethesda, MD) to relevantly evaluate the inflammation for MGD patients, which not only is time consuming and laborious but also requires professional ability for inspection and interpretation of results. Telangiectasia or lid margin vascularity is another important sign for the diagnosis of MGD, indicating an increase in inflammation.¹⁹ Some studies have confirmed that some therapies, such as intense pulsed light or intrameibomian gland injection of the anti-VEGF agent, which can reduce or close the excessive vessels of the lid margin, may significantly improve the symptoms and degree of MGD.^4,5 However, the measurement of neovascularization in these studies was mostly carried out using subjective two-value quality rating scales (normal/abnormal) or computer-assisted quantitative analysis through digital slit-lamp images, which is not accurate and convenient enough for extensive clinical or research evaluation.^4–6 Therefore, a noninvasive, convenient, and quantitative technique that can be widely used to accurately assess the severity of vessels at the eyelid margin is urgently needed. The present study is the first, to our knowledge, to propose the utilization of SS-OCTA for quantitative analysis of blood flow at the lid margin while also confirming the excellent repeatability of OCTA-measured blood flow at the lid margin and its correlation with clinical data related to the ocular surface. 

OCTA for anterior segment vasculature imaging, which is an emerging modality, includes the cornea, iris, sclera, and conjunctiva.^9,20,21 It works on the basis of low coherence interferometry and analysis of signal decorrelation between consecutive scans by comparing the changes in phase, amplitude, or the full OCT signal.²² In our study, we evaluated the repeatability of LMBFD obtained through OCTA. As there is no mode specially designed for the analysis of lid margin vessels, we used the scan mode designed for conjunctival vessels with an area of 6 mm × 6 mm or 9 mm × 9 mm. The scan mode with an area of 6 mm × 6 mm showed better repeatability (ICC > 0.90). This may be because the VG200D OCTA system for the anterior segment was equipped with projection artifact removal, but without motion correction, which may magnify the effect of eyelid flutter on the results to a certain extent. In the scan mode with 6 mm × 6 mm area, the scanning area is smaller, and the scanning time is shorter; consequently, the effect of eyelid shaking on the results is less, leading to better repeatability. Meanwhile, during our examinations, we also found that the mode with 9 mm × 9 mm area required a longer scan time, so it was more difficult for some patients to time, so it was more difficult for some patients to cooperate to obtain high-quality images. Due to the limitations of the analysis mode and the width of the palpebral margin, we selected only the middle 5 mm × 1 mm range of the lid margin for analysis. The reproducibility of image analysis between the two examiners was extremely high (ICC = 0.986 and 0.988). The absolute values of LMBFD (range, 18.67%–94.19%) seem to be quite high. As OCTA did not show the depth of the measurement, we used ImageJ software (Java 8.0_172, 64-bit) to measure some OCT images, and the results showed that the depth of blood vessels measured was approximately 500 µm (Fig. 2D). Therefore, the blood flow density of the lid margin measured by OCTA includes not only the blood vessels of the blepharon surface but also the blood vessels of the subcutaneous tissue. Meanwhile, the results of validity analysis showed that the value of LMBFD measured by OCTA was consistent with the severity of the lid margin vasculature shown in the photograph. These data support a role for OCTA as a commercial instrument in evaluating lid margin vascularity. 

Interestingly, our results demonstrated that the LMBFD values gradually increased with the degree of lid margin neovascularization and were significantly positively correlated with more severe clinical symptoms (OSDI), poorer tear film stability (FTBUT, CFS), and worse MG-related parameters such as LMS, ME, and MQ. Previous studies have reported that ocular surface inflammation was closely related to the deterioration of tear stability.^23,24 Long-term chronic ocular surface inflammation may be manifested in lid margin neovascularity, which also explains its significant correlation with worse tear film stability and similarly with poor MG function. However, as the study did not include inflammatory tear cytokines, its relationship with ocular surface inflammation should be further studied. Based on our results, we suggest that LMBFD may reflect eyelid inflammation in MGD patients. Meanwhile, the LMBFD was not statistically correlated with age in this study. Hykin and Bron²⁵ found that the lid margin vascularity severity, which was graded according to the involvement of vessels of the palpebral margin (1 point for each 25% increase in the range of involvement, with a total score of 0–4), appears to be affected not only by MGD but also by aging. The possible reason for different results may be that the depth and range of the vessels assessed by the two methods were not consistent. 

This research expands the clinical application of OCTA in MGD patients with lid margin vascularity. At present, it is the only viable, rapid, and noninvasive tool that can be used in the clinic. There are several limitations to this study. First, the sample size of the normal control and three subgroups was relatively small to accurately compare repeatability between subgroups. Second, a repeatability analysis between different examiners was lacking and should be conducted in the future. Additionally, although we strictly followed an established scanning and image analysis protocol, the scope of the assessment was limited only to the middle third of the palpebral margin. The repeat scan area tracking was not used for this measurement, which may affect the measurement accuracy to some extent. Thus, a more adaptive measurement mode to improve accuracy and convenience especially for the eyelid has yet to be developed. Moreover, its application in monitoring therapeutic effects requires further study. 

In conclusion, this study utilized a noncontact OCTA system for the first time, to our knowledge, in MGD patients to successfully delineate vascularity in the lid margin, with good repeatability, reproducibility, and efficacy. Although the indicator still must be updated with a more adaptive measurement mode and clinical validation in the future, it holds promise as a clinical metric for eyelid inflammation. 

Acknowledgments

The authors warmly thank all the individuals who participated in this study. 

Supported by grants from the Wenzhou Science and Technology Plan Project (Y2023820), the Provincial Construction Project of Zhejiang (WKJ-ZJ-2135) and the Pioneer and Leading Goose Project of Zhejiang Province (2022C03070). 

Disclosure: M. Hu, None; C. Wang, None; Y. Li, None; H. Zhang, None; H. Li, None; Q. Dai, None; H. Lian, None; Y. Zhao, None; Y. Fu, None