Imaging Lid-Parallel Conjunctival Folds with OCT and Comparing Its Grading with the Slit Lamp Classification in Dry Eye Patients and Normal Subjects

Abstract

Purpose.: To visualize and describe the morphologic appearance of lid-parallel conjunctival folds (LIPCOFs) by using optical coherence tomography (OCT) and to relate it to dry eye signs and symptoms.

Methods.: The LIPCOF grade, noninvasive tear film breakup time (NIBUT), lipid layer interference pattern, and dry eye symptoms were recorded in 17 normal subjects and 33 patients with dry eye. LIPCOFs were evaluated with a slit lamp and visualized by OCT. Three different algorithms for OCT were developed to grade LIPCOFs according to tear meniscus height or the covering tear film on the folds.

Results.: The three OCT methods showed significant correlation with the slit lamp method (r = 0.470–0.473, P < 0.01). The OCT LIPCOF imaging methods were independent of NIBUT. The Dry Eye Questionnaire (DEQ) scores correlated with the height of the folds and the absence of tear film coverage of the folds (r = 0.574, P < 0.001 and r = −0.527, P < 0.001, respectively). The OCT LIPCOF grades correlated with the DEQ scores (r = 0.494, P < 0.001 and r = 0.310, P = 0.029, respectively). The slit lamp grade did not correlate with the DEQ scores in the whole population, but did in the normal group (r = 0.458, P = 0.024). The OCT LIPCOF grades showed inverse correlation with lipid pattern in the normal group (r = −0.422–0.481, P = 0.05); however, this association disappeared in the dry eye group.

Conclusions.: OCT enabled a noninvasive, high-resolution method of imaging, evaluating, and classifying LIPCOFs. These new classifications correlated well with the slit lamp grade and the DEQ scores, promising a new, more objective evaluation of dry eye.

Dry eye is a multifactorial disease of the ocular surface that causes symptoms of discomfort, visual disturbance, and tear film instability with potential damage to the eye. ^1,2  

Conjunctival folds were termed conjunctivochalasis by Hughes, ³ and Höh et al. ^4,5 described a detailed classification of LIPCOFs that compose a subtype of conjunctivochalasis. ⁶ The description of different degrees of LIPCOFs uses the patient's tear meniscus height as a comparator. ¹ The LIPCOF classification evokes a clinical interpretation, as it provides information on the severity of the dry eye syndrome and has a positive predictive value of 93.09% and a negative predictive value of 75.95%. ^1,7  

Optical coherence tomography (OCT) is a noninvasive method also suitable for the examination of the anterior segment of the eye with high magnification. Its advantage is that it generates little or no reflex tearing when compared with the slit lamp examination, which induces reflex tearing and changes in the tear menisci. ⁸  

To our knowledge, no studies have been published on the imaging of LIPCOFs by OCT. We intended to establish new algorithms for LIPCOF classification on the basis of OCT scans of conjunctival folds and tear menisci and to examine the relationship between this grading and classic dry eye parameters. 

Methods

The study adhered to the tenets of the Declaration of Helsinki. Informed consent was obtained from the subjects after explanation of the nature and possible consequences of the study. The research was approved by the Semmelweis University Regional and Institutional Committee of Science and Research Ethics. 

Inclusion and Exclusion Criteria

Dry eye patients randomly selected from the outpatient department at our ophthalmology clinic and healthy volunteers were asked to participate in the study. 

The subjects were grouped into dry eye and normal control patients according to their response to the Dry Eye Questionnaire ⁹ (DEQ) and a noninvasive test that determined tear film breakup time (NIBUT). 

Exclusion criteria were ocular surface disease, history of ocular surgery, treatment with a medication known to influence the ocular surface or tear film, and pregnancy. 

Signs and Symptoms of Dry Eye

The examinations took place in a dimly lit consulting room where the temperature and humidity were controlled. During the study visit, we did not use any type of eye drops. The schedule of the investigation allowed sufficient time for the ocular surface to recover after different types of illumination and to avoid reflex tearing and surface drying. 

Slit Lamp Examination.

LIPCOF Grading.

The participants were instructed to look straight ahead, and after some blinking, the LIPCOFs were evaluated with slit lamp at the temporal lower quadrant of the eye fissure, according to grading scale published by Höh et al. ⁴  

Ocular Surface Disease.

According to the description of LIPCOFs, conjunctival folds disappear when the lower eyelid is pulled away from the eyeball. ⁴ Therefore, all eyes were evaluated by a clinical ophthalmologist before grading, to exclude ocular surface disease causing similar conjunctival folds (e.g., pemphigoid). 

Dry Eye Questionnaire.

The DEQ ⁹ form measures the prevalence, frequency, and diurnal intensity of ocular surface symptoms of eye discomfort, dryness, visual disturbance, soreness and irritation, grittiness and scratchiness, foreign body sensation, burning and stinging, light sensitivity, and itching. 

We calculated the product of frequency and daily average intensity scores of each symptom. Finally, the nine individual subscale scores were summarized, adding up to a maximum score of 135 (3 × 5 × 9). 

Ocular and Nonocular Medical History.

We recorded the presence of autoimmune disease, skin disease, neuronal disease, general therapies (i.e., antidepressants and antihistamines), and ocular surgery. 

Optical Coherence Tomography.

Design of OCT Examinations.

One randomly selected eye of each subject was imaged by high-speed, high-resolution Fourier domain (FD)-OCT (RTVue-100; Optovue, Inc., Fremont, CA). The system performs 26,000 axial scans per second with an axial resolution of approximately 5 μm. Vertical scans were performed with a short-focus corneal adaptor module (S-CAM). 

For the visualization of the lower tear meniscus, the scan axis was centered at about the middle of the lower eyelid. During the imaging of the LIPCOFs of the right eye, the patients were instructed to position their heads in a 45° turn to the left. In the case of the left eye, the image was captured with the head turned the opposite way (Fig. 1). In primary gaze, the scan axis was centered on the temporal one third of the lower lid. 

Figure 1.

 

A patient in the proper position during LIPCOF imaging with OCT.

The raster imaging created 17 cross sections in a distance of 2 mm along the margin. 

Evaluation of the OCT Measurements.

In the vertical OCT scans, the existence, number, and height of the conjunctival folds and their coverage by tear film were detected. The tear meniscus area was outlined with a mouse pointer to select the three corners of the tear prism on the linear map. The area of the triangle was calculated by using the OCT system's incorporated software. According to Bitton et al. ¹⁰ the tear meniscus was assumed to be generally constant over a 1-mm distance; therefore, the tear meniscus volume was calculated as the product of the tear meniscus triangle's area and 1 mm (Fig. 2). 

Figure 2.

 

OCT line scan of the inferior tear prism taken at the middle of the lid margin in the line of the pupil center, delineated by lines measured in micrometers. The area is used for calculating the tear meniscus volume in microliters per one millimeter of the lower lid.

Lower tear meniscus height (LTMH) measurements were obtained immediately and 3 seconds after voluntary complete blinking. On the linear and the raster scans, we measured the LTMH in the pupillary plane. 

Development of a New OCT-Based Classification.

We established three algorithms for classification of LIPCOFs (Table 1):

1.  
   OCT method I: a count of the folds on the raster OCT scans and correlation of the height of the folds to the average LTMH.
2.  
   OCT method II: grading based on the LTMH measured 3 seconds after blinking.
3.  
   OCT method III: a test of whether the folds are covered by the tear film in the lower temporal position.

Table 1.

 

Descriptions of the Three Algorithms Developed to Grade the LIPCOF Imaging by OCT

Degree of Intensity of LIPCOF	Method I	Method II	Method III
Basis of description	Description based on average LTMH	Description based on LTMH measured 3 seconds after blinking	Description based on tear film coverage of the folds itself
Degree 0	No permanently present fold	No permanently present fold	No permanently present fold
Degree 1	Single fold, smaller than the patient's average LTMH	Single fold, smaller than the patient's 3-second LTMH	Single fold, covered with patient's tear film
Degree 2	a. Single fold of up or higher than the patient's average LTMH	a. Single fold up to the patient's 3-second LTMH	a. Single fold not covered with patient's tear film
	b. Multiple fold not higher than the patient's average LTMH	b. multiple fold lower than the patient's 3-second LTMH	b. Multiple fold covered with patient's tear film
Degree 3	Single or multiple folds higher than the patient's average LTMH	Single or multiple fold higher than the patient's 3-second LTMH	Multiple folds not covered with patient's tear film

Noninvasive Tear Film Breakup Time.

The noninvasive tear film break-up time (NIBUT) was measured with an interferometer (Tearscope Plus; Keeler Ltd., Windsor, UK). ¹¹ The mean NIBUT was calculated from three different measurements. 

Interference Pattern of the Tear Film Lipid Layer.

The tear lipid pattern was graded according to the instructions in the interferometer users' manual. ¹¹ Grades 1 and 6 showed abnormally thin and abnormally thick tear film lipid layers, respectively. Grades 2, 3, 4, and 5 referred to a lipid layer thickness within normal range but gradually thickening. 

Statistical Analysis

The statistical analysis was performed and the figures were composed with commercial software (SPSS 16.0; SPSS, Chicago, IL, and Statistica 8.0; StatSoft, Tulsa, OK, respectively). The Mann-Whitney U test was used to compare the dry eye signs determined by slit lamp or OCT. We analyzed the correlation between measured parameters with Kendall's τ-b or Spearman correlation tests. 

Results

Patients' Characteristics

Fifty patients completed the study. According to the clinical data, 33 patients were classified in the dry eye group (mean age, 63.0; range, 25–89 years; 31 women) and 17 in the normal control group (mean age, 35.6; range, 20–79 years; 11 women). 

In the dry eye group, 10 of the 33 patients had medical conditions associated with dry eye syndrome (Sjögren's syndrome and rosacea). The mean NIBUT was significantly shorter, and the sum of the nine individual DEQ subscales' scores was significantly higher in the dry eye group than in the normal control group (both P < 0.001). The distribution of the different lipid pattern did not show a significant difference between the groups (Table 2). 

Table 2.

 

Comparing the Tear Lipid Pattern Distribution, NIBUT, and DEQ Score of the Normal and Dry Eye Groups

	Normal Group (n = 17)	Dry Eye Group (n = 33)	P
Tear lipid pattern, %
Grade 1	35.29	18.18	NS
Grade 2	5.88	18.18
Grade 3	23.53	33.33
Grade 4	11.76	0.00
Grade 5	23.53	30.30
NIBUT, s*	22.67 ± 14.68	7.52 ± 2.47	<0.001
DEQ score, s*	7.2 ± 5.1	29.7 ± 12.7	<0.001

*  Mean ± SE.

General Considerations of LIPCOF Imaging with FD-OCT

We were able to visualize the LIPCOFs with the help of OCT imaging. According to the serial OCT cross-sectional scans, some folds were covered by the tear film and some were not (Fig. 3). 

Figure 3.

 

(A) OCT raster image of multiple (n = 4) lid parallel conjunctival folds covered by tear film. (B) Lid parallel conjunctival folds not covered by tear film. Under the folds with medium reflectivity, an area was detected with lower inner reflectivity (arrow) than the folds themselves.

If the axial scans were conducted through a conjunctival vessel, a shadowing phenomenon appeared on the image (Fig. 4A). This typical image made it possible to differentiate vascularization from other phenomena. In 42 of 50 cases, a special area could be detected under the folds, with lower inner reflectivity than the folds themselves (Figs. 3B, 4A). In some cases, the dark area was round and belonged to one conjunctival fold, or it was oval and confluent belonging to more than one fold. The area may denote subconjunctival fluid under the LIPCOFs. In Figure 4B some rounded areas were detected with low inner reflectivity encircled by an annular area with higher reflectivity. The inner part's reflectivity was homogeneous and corresponded with the anterior chamber fluid's reflectivity. As conjunctival vessels showed a different image, this structure mostly resembled subconjunctival cysts. 

Figure 4.

 

(A) Blood vessels in the neighborhood of the LIPCOFs cause a shadowing effect in the underlying conjunctiva and sclera. (B) Fluid-filled cysts in the conjunctiva behind the LIPCOFs (arrows).

In the lower temporal lid margin, the heights of the LIPCOFs were determined. The conjunctival folds were counted on those raster scans that showed the highest number of folds. If the fold stretched over the tear meniscus at the point of the LIPCOF, the conjunctival fold was regarded as not being properly covered by the tear film (Fig. 3B). 

LIPCOF Classification Algorithms

All patients were classified for LIPCOFs with both slit lamp and OCT (Table 1). We established three different types of classification for LIPCOF grading with OCT and tested the distribution of the grades. The frequency of each grade was similar for all the OCT methods but different for the slit lamp method (Table 3). In the normal control group, LTMH measurements were performed according to OCT method II at 3 seconds after blinking. However, in the dry eye group, this measurement was encumbered by frequent involuntary blinking. Full agreement was found between results of the slit lamp and the OCT methods in 54.5% and 41.3% of the cases in dry eye and in normal groups, respectively. The slit lamp method mostly overestimated the LIPCOF grade compared with the OCT methods in cases with discrepancy in the grading (Table 4). 

Table 3.

 

Distribution of LIPCOF Grades with the Slit Lamp and the Three OCT Methods

LIPCOF Grade	OCT Method I	OCT Method II*	OCT Method III	Slit Lamp Method
0	1 (2)	1 (6)	1 (2)	1 (2)
1	10 (20)	5 (29)	11 (22)	6 (12)
2	29 (58)	9 (53)	27 (54)	15 (30)
3	10 (20)	2 (12)	11 (22)	28 (56)

 Data are expressed as the number of participants with each grade (percentage of the whole group). Agreement was found between the OCT methods and disagreement between the OCT and the slit lamp method (χ² with Bonferroni correction).

*  Normal group.

Table 4.

 

Agreements and Discrepancies between LIPCOF Gradings with the Slit Lamp and the OCT I Method in the Dry Eye and the Normal Group

	Dry Eye Group (%) n = 33	Normal Group (%) n = 17
Full agreement	54.5	41.2
Discrepancy = 1 grade	30.3	47.1
OCT < slit lamp
Discrepancy = 1 grade	9.1	5.9
OCT > slit lamp
Discrepancy > 1 grade	6.1	5.9

The dry eye sample was divided into two subgroups. In the mild subgroup (n = 18) the DEQ scores were lower than 30, and the symptoms were never constant or very intense. In the moderate subgroup (n = 15) the DEQ scores were higher than 30, and some dry eye symptoms may have occurred constantly or were described as intense or very intense. Both OCT algorithms resulted in grading that gave the patients with mild or moderate dry eye significantly different LIPCOF grades (P = 0.04); however, the slit lamp method did not result in different LIPCOF grades between the subgroups (Table 5). 

Table 5.

 

Differences between the LIPCOF Grades According to the Severity of Dry Eye

LIPCOF Grade	Mild Dry Eye Group, n (%)	Moderate Dry Eye Group, n (%)	P
OCT Method I
0	0 (0)	0 (0)	0.04
1	4 (22)	1 (6)
2	12 (67)	7 (47)
3	2 (11)	7 (47)
OCT Method III
0	0 (0)	0 (0)	0.04
1	7 (39)	0 (0)
2	7 (39)	10 (67)
3	4 (22)	5 (33)
Slit Lamp Grade
0	0 (0)	0 (0)	NS
1	3 (17)	1 (7)
2	5 (28)	6 (40)
3	10 (55)	8 (53)

In the whole study group, there was a strong and significant correlation between the OCT methods (n = 50, r = 0.648, P < 0.001), and all OCT LIPCOF grading methods showed a significant correlation with the slit lamp method (r = 0.470–0.473, P < 0.01 in all cases). 

Relationships between the FD-OCT-Measured Parameters of LIPCOFs and the Subjective Symptoms of Dry Eye

We selected the most pronounced LIPCOF in each subject and calculated the mean height in our population (370.6 ± 183.2 μm; range, 0–1110). In the dry eye group the mean LIPCOF height was 483.8 ± 189.8 (range, 140–1110) μm, and in the control group 151 ± 61.9 (range, 0–240) μm (Mann-Whitney P < 0.0001). The height of the LIPCOF and the DEQ scores showed a strong, positive correlation (r = 0.574, P < 0.0001, n = 50). In contrast, the presence of tear film covering on the LIPCOF showed a negative correlation with the DEQ scores (r = −0.527, P < 0.0001; n = 50). 

In the mild dry eye subgroup, the height of the LIPCOF did not differ from that in the moderate dry eye subgroups (P = 0.452). In addition, the LTMH and the tear meniscus area were similar in these two groups (P = 0.346 and P = 0.129, respectively). 

However, in the moderate dry eye group, significantly more people had LIPCOFs that were not covered with tear film than in the mild group (8/18 [44%] vs. 13/15 [86%] patients; P = 0.04). 

Relationships between LIPCOF Grades and the Symptoms of Dry Eye

In the dry eye group, the LIPCOF grades correlated with the DEQ scores (r = 0.494, P < 0.001 and r = 0.310, P = 0.029). Although the slit lamp grading scheme did not show any correlations with DEQ scores (r = 0.148, P = 0.516) in the whole population, the slit lamp grading and the symptoms correlated moderately in the normal control group (r = 0.458, P = 0.024; Fig. 5). 

Figure 5.

 

Association between the different grading schemes and the subjective dry eye scores.

Among cases with a single fold (n = 21, 42%), nine (42%) scans were captured without tear film coverage. Among the cases with multiple folds (n = 28, 56%), the percentage of noncoverage increased by 15% (n = 16, 57%). 

The OCT LIPCOF grade was independent of NIBUT, age, and sex. The OCT LIPCOF grades correlated negatively with the tear film lipid pattern in the normal group (r = −0.422–0.481, P = 0.05); however, this association disappeared in the dry eye group (r < 0.213, P = 0.234–0.612). 

Associations of Tear Volume Changes and LIPCOF Grading Schemes in Normal Controls

In the normal group, the mean ± SD change in tear volume was calculated during a blink. Immediately after the blink, the mean tear volume was 3.47 ± 1.80 (range, 1.00–16.50) × 10^–2 microliter/mm, and 3 seconds after the blink it was 2.69 ± 1.30 (range, 0.90–10.40) × 10^–2 microliter/mm. The mean change in tear volume measured with OCT was almost 10-fold higher in patients with multiple folds than in patients with single folds (mean ± SD, 9 ± 9.5 × 10^–4 vs. 1 ± 1.3 × 10^–4 microliter/mm). The LIPCOF grades classified with slit lamp showed moderate and significant correlation with the tear volume changes measured during a blink (Kendall's τ r = 0.448, P = 0.027). 

In normal controls we measured the LTMH in the pupillary plane immediately after blinking and 3 seconds later. The average LTMH immediately after blinking was 302.12 ± 91.2 (167–741) μm and 3 seconds later was 273.53 ± 76.8 (range, 137–735) μm. The LTMH measured at the two time points showed no statistically significant difference (P = 0.537). 

As dry eye patients had significantly shorter NIBUT and more difficulty in restraining voluntary blinking, in this group we performed only the LTMH measurements immediately after blinking (mean: 364.2 ± 173.21; range, 20.83–1041.67 μm). Furthermore, the LTMH values of the dry eye and the control patients did not differ significantly (P = 0.180). 

Although we used the OCT scans for the LIPCOF grading, we detected similar differences in the mean tear volume change measured on OCT scans between patients with multiple folds (LIPCOF method II, 3, n = 14) and in patients with single folds (LIPCOF method I, n = 3; mean ± SD: 11 ± 10.2 × 10^–4 vs. 1.4 ± 1.6 × 10^–4 microliter/mm), the mean tear volume change did not correlate with LIPCOF-OCT grading. 

Discussion

As far as we are aware, we are the first to report on the visualization of LIPCOFs with an FD-OCT system equipped with a corneal adaptor. 

The axials scans made it possible to visualize the conjunctival texture behind the LIPCOFs, and in some cases cysts, subconjunctival fluid, and vessels were detected. The conjunctival folds had medium reflectivity, and the areas with lower inner reflectivity possibly corresponded to subconjunctival fluid. There is no information available yet about the possible mechanism of such subconjunctival fluid accumulation (Fig. 3B, 4A). A follow-up study or a molecular analysis could get us closer to the explanation. 

According to the slip lamp grading method of Höh et al., LIPCOF grades were comparable with the degree of dry eye. ^4,6 As dry eye is for the most part a symptom-based disease, we described three new LIPCOF classification algorithms based on OCT scans affirming subjective symptom scores and potentiate the appreciation of the severity. 

Unlike previous studies ^4,6 we based our grading system on the patient's averaged tear meniscus height, measured immediately after blinking and 3 seconds later in the pupil line, instead of on a theoretically normal, but in reality continuously altering, tear meniscus. ^8,10,12 Furthermore, during the OCT grading, we took into account the coverage of the conjunctival folds by tear film which is clearly visible using OCT imaging. 

The importance of the knowledge of the folds' coverage by tear film is emphasized by our result, which showed that more people had conjunctival folds without tear film coverage in the moderate than in the mild dry eye group. Using high-speed video topography, Nemeth et al. ¹³ provided quantitative measurements of tear-film dynamics stating, that after a blink, it takes some seconds to reach the most regular state. Taking into account the so-called tear film buildup time, hypothetically the conjunctival folds' coverage by tear film would have a steady state time point, causing the least ocular surface irregularity or subjective dry eye symptoms. 

The LIPCOF gradings performed by slit lamp and by OCT are in strong statistical correlation. The two grading methods showed full agreement or maximum one-grade discrepancy in >80% of the individual cases. In only approximately 6% of cases, we found more than one-grade discrepancies between the OCT and the slit lamp grading. According to our results the slit lamp method tended to overestimate the seriousness of dry eye in these cases. 

In our work, using the high resolution of OCT imaging, we were able to objectively measure the height of the folds and the tear meniscus with micrometer-scale resolution. The mean LIPCOF height in dry eye patients was about three times higher than in normal controls. A comparison of this result with other authors cannot be achieved, as to our knowledge no data are available on the exact size of the folds. 

Johnson et al. ¹² noted first a rapid, near-exponential LTMH increase after blinking followed by a plateau phase as the tear film thins over the ocular surface. In our measurements, we did not find a statistically significant change between LTMH measured just after eye opening and 3 seconds later. One reason for that could be, as the literature pointed out, that the main LTMH changes fell on the first 2 seconds after blinking. According to Zhou et al. ¹⁴ FD-OCT measures LTMH dimensions and area with high between-visits reproducibility. Shen et al. ⁸ measured a cutoff value of 164 μm in the lower tear meniscus between normal and dry eye immediately after blinking by real-time OCT, suggesting a good diagnostic accuracy for dry eye. In our healthy population, the mean LTMH was approximately twice as high as the above-mentioned cutoff value and close to the average LTMH of Johnson et al. ¹² According to other studies ^15,16 our LTMH OCT measurements were reliable and noninvasive. 

In the diagnosis of dry eye, Shen et al. ^8,17 reported that LTMH had a diagnostic sensitivity of 0.92 and specificity of 0.90, exceeding the upper TMH with a diagnostic sensitivity of 0.72 and specificity 0.85. Fodor et al. ¹⁸ found a highest reproducibility of the LIPCOF test in the lower temporal position. Therefore, we conducted our OCT scans at the lower temporal part of the eyelid. Schirra et al. ⁶ described a method for estimating the relative risk of dry eye with the help of the morphologic appearance of the conjunctiva. 

Watanabe et al. ¹⁹ suggested that LIPCOFs may be a result of mechanical forces during blinking with an overexpansion or lymphatic dilation. These folds present a large surface area of exposed epithelium, and thus may be more susceptible to drying that will further increase friction. ⁷ The mean tear volume change measured with OCT in our study was almost 10-fold higher in patients with multiple folds than in patients with single folds. This phenomenon may have evoked the folding of the conjunctiva results in ocular surface irregularity. Our results emphasize the microenvironmental process that occurs even in normal controls. A sharp decrease in tear volume occurs after blinking in the neighborhood of multiple folds compared with single folds. This early stage of tear film instability may refer to a future dry eye condition, although prospective studies have to confirm this hypothesis. 

The DEQ ⁹ gave the possibility of describing precisely the subjective symptoms. We found it very important to include the patient's perception and self diagnosis of dry eye when evaluating parameters of ocular surface imaging. The OCT grades, the height of the folds, and the existence of tear film coverage were in good accordance with the severity of dry eye measured by DEQ. 

In our study, the thinner lipid layer associated with higher LIPCOF grades with multiple and higher folds. The extremely thin lipid layer is characteristic of meibomian gland dysfunction, which is the most common cause of evaporative dry eye. ^1,11 Even with the preliminary small number of cases, the inverse correlation between the LIPCOF grade and lipid pattern supports the results of Dausch et al., ²⁰ who reported improvements in the LIPCOFs after treatment with phospholipid liposomes for dry eye. 

For the evaluation of LIPCOFs, the OCT scans outshone the slit lamp examination. Höh et al. ⁵ performed invasive methods, including the Schirmer test and impression cytology, to test the predictive value of LIPCOFs for dry eye. In contrast, in our study, we analyzed the dry eye signs noninvasively. Furthermore, taking into account the time schedule of our clinical investigation, the reflex tearing and surface drying did not disturb our analysis. The OCT classification was based on meticulous morphologic pictures with around 5-μm resolution. On the other hand, OCT imaging also gave the possibility of visualizing the direct connection and relationship between the conjunctival folds and the tear film, which is not possible using the slit lamp. 

In summary, we propose that FD-OCT could be an ideal tool for imaging LIPCOFs. Thus, it may provide a quick, noninvasive, quantitative method for the assessment and follow-up of dry eye patients. Our classification method describes the local anatomic situation, which has direct functional and clinical consequences. All the three OCT LIPCOF classification algorithms showed strong correlations with the classic slit lamp grading method and the subjective and objective signs of dry eye disease. The association of the lipid pattern of the tear film and the LIPCOF grades suggests a potential role in the etiology of the conjunctival folds and may put into focus the differential diagnosis of dry eye. The application of OCT to LIPCOFs could open a new field of exploration, to determine the effect of LIPCOFs on the ocular surface. 

Footnotes

 Disclosure: A. Veres, None; B. Tapaszto, None; K. Kosina-Hagyó, None; G.M. Somfai, None; J. Németh, None