Corneal Confocal Microscopy Features and Tear Molecular Profile in Study Participants with Discordance between Ocular Surface Disease Clinical Signs and Discomfort

Various ocular surface conditions such as dry eye disease can present with severe discomfort and pain. However, it is clinically challenging to establish etiology and prescribe correct treatment in patients who have a lot of discordance between symptoms and signs. To understand the basis of such discordance, we stratified subjects with ocular surface pain based on concordance between the severity of signs and symptoms and evaluated corneal structural features and tear molecular factors. All subjects underwent slit lamp examination, dry eye evaluation, and ocular surface disease index (OSDI) scoring. Subjects were stratified into group 1—without symptoms or clinical signs; group 2—without symptoms but with signs; group 3—with similar severity of symptoms and signs; and group 4—with symptom severity greater than that of the signs. Laser scanning in vivo confocal microscopy (IVCM) and tear fluid analysis for soluble factors by multiplex ELISA was performed for all subjects. Patients with a higher grade of symptoms and signs showed increased corneal dendritic cell (cDC) density (p < 0.05) which was more pronounced in subjects with discordance between the symptoms and signs (group 4). A significantly higher proportion of microneuroma-like structures and cDC were observed in group 4. IL-17A levels were significantly elevated in the tears of subjects with more discomfort. Our results demonstrate that corneal IVCM and the measurement of tear film factors can help clinicians improve diagnosis and treatment choice. Stratifying patients with ocular surface discomfort on the basis of discordance between symptoms and clinical signs may help identify patients who need additional adjunctive targeted therapy to resolve their condition.


Introduction
Ocular pain and discomfort are common presenting symptoms for patients visiting the outpatient clinic and can significantly deteriorate a patient's quality of life and their activities of daily living [1,2]. Infections, inflammation, and raised intraocular pressure are some common aetiologies of ocular pain in general, while conditions such as dry eye disease (DED), corneal infections, or injury are often related to ocular surface pain [3]. Structural factors such as the dense innervation of the cornea as well as the immune, inflammatory, 2 of 18 and neural interconnections of the ocular surface have an important role to play in ocular surface pain [3,4]. In some cases, the pain may be disproportionately higher than the clinical signs, thereby complicating the diagnosis and making management challenging [5]. Exaggerated ocular surface pain can be secondary to inflammation, increased nociception, neuropathy, or a combination of these factors that may co-exist with various ocular surface conditions. Nociceptive pain usually occurs due to an alteration in a non-neural tissue resulting in an increase in nociceptive molecular factors [1], whereas neuropathic corneal pain (NCP) is secondary to the involvement of the somatosensory nervous system [6]. However, significant overlap between these entities can occur [7]. It is important to identify nociceptive and neuropathic aspects of pain so as to treat them appropriately [8]. Neuropathic cases can have symptoms of pain and discomfort despite a complete absence of clinically visible ocular surface abnormalities [9,10]. Overlap between conditions and this discordance between signs and symptoms are major challenges in treating these patients even in those with an established diagnosis [11][12][13]. In addition, many patients have an assumed diagnosis of DED, and do not respond to therapy as expected. These patients are often treated with multiple different topical and systemic medications with suboptimal relief and may also get referred to the pain management clinic or for a psychological assessment [14].
A possible reason for this exaggerated pain response without significant clinical features may be subclinical inflammation [15,16], which may involve the corneal nerves leading to complaints of pain, discomfort, and a burning sensation secondary to polymodal nociceptor and thermoreceptor activation [17,18]. Additional tests such as laser scanning in vivo confocal microscopy (IVCM) and tear molecular analysis can help identify these hidden causes for pain. IVCM is an excellent tool to assess both inflammation and nerve abnormalities by assessing the density and morphology of corneal dendritic cells (cDC) and the alterations in the corneal nerve structures that are known to be associated with ocular surface pain [19,20]. Chronic dry eye and ocular surface inflammation can also have recalcitrant symptomatology even with an established diagnosis due to changes in corneal nerves structure and function resulting in NCP [21]. Therefore, evaluating patients with poor correlation between symptoms and signs would aid in stratifying treatment. As disparity between symptoms and signs is an important feature of NCP, we have used this as the main criteria to divide patients into four groups. We therefore evaluated the corneal IVCM feature changes (cDC, sub-basal nerve plexus features-SBNP, and microneuromalike structures) and tear molecular factors. We have then analysed the variations between these groups to identify specific subgroup characteristics that would help treatment.

Study Cohort, Clinical Examination, and Study Groups
This cross-sectional study has received approval of the Narayana Nethralaya Institutional ethics board. Subject recruitment, clinical information, and sample collection procedures were conducted in accordance with the tenets of the Declaration of Helsinki. Written informed consent was obtained from each subject prior to inclusion in the study. A total of 145 subjects (287 eyes) were included in the study with 115 patients (229 eyes) having the primary symptom of ocular surface pain or discomfort referred to the Cornea Clinic of Narayana Nethralaya, located in Bangalore, India, and 30 control subjects (58 eyes) with no symptoms of ocular pain or discomfort and no clinically visible ocular abnormality on examination. Patients were evaluated for possible causes for ocular discomfort using a detailed clinical history, including history of systemic illness and ocular examination. Visual acuity assessment, refraction, detailed slit-lamp examination including ocular surface evaluation for signs of inflammation and eyelid and meibomian gland assessment, and fundus evaluation were performed to exclude other ocular comorbidities. Exclusion criteria included the use of contact lenses, ocular infection in the last six months, uveitis, ocular trauma, history of any ocular surgery including refractive surgery, cicatrizing dis-ease, lid abnormalities and those on topical steroids and topical anti-inflammatory or immunomodulatory medication including cyclosporine A in the last 3 months.
As there are many possible causes for ocular surface pain, a detailed assessment is required to arrive at the possible underlying diagnosis and plan the appropriate management [22]. The presenting symptom of pain or discomfort was assessed using the ocular surface disease index (OSDI) questionnaire (Allergan, Dublin, Ireland) [23,24], and the severity graded based on the standard scores as follows: normal < 12; mild 13-22; moderate 23-32; and severe > 32 [25]. As DED is one the most common causes for ocular surface pain [8] and referral to the cornea clinic, detailed ocular surface and dry eye evaluation was done. Dry eye evaluation included Schirmer's test without anaesthesia, tear film breakup time (TBUT), and corneal and conjunctival fluorescein staining. Using sterile Schirmer's strips (Contacare Ophthalmics and Diagnostics, Vadodara, Gujarat, India), the Schirmer's test was performed. TBUT and ocular surface staining were performed using fluorescein strips (Contacare Ophthalmics and Diagnostics). Clinical tests such as the ocular surface staining with fluorescein and other dyes give us an insight into the health of the ocular surface. Tear film break up time assesses the tear film stability and the Schirmer's test measures the tear secretion rate. Subjects were divided into those without DED or having mild, moderate, and severe DED based on the Schirmer's and/or TBUT values as per standard guidelines from the TFOS DEWSII (Tear Film & Ocular Surface Society Dry eye workshop II) [26,27]. Further, DED patients have been classified as per TFOS DEWII into two predominant categories, i.e., aqueous deficient dry eye disease (ADED) or evaporative dry eye disease (EDED) based on the Schirmer's test I value and TBUT. Subjects with TBUT < 10 s but with a Schirmer's test I value > 10 mm/5 min were grouped as EDED. Subjects with a Schirmer's test I value < 10 mm/5 min were classified as ADED. Grade of tear film instability assessed by TBUT are as follows: normal ≥ 10 s, mild 7-9 s, moderate 5-7 s, and severe < 5 s. Grade of tear fluid secretion status based on Schirmer's test I metrics are as follows: normal ≥ 10 mm/5 min, mild 7-9 mm/5 min, moderate 5-7 mm/5 min, and severe < 5 mm/5 min.
To determine additional factors that may contribute to the ocular surface pain, the study participants were grouped based on the disparity between the severity of their symptoms and signs irrespective of their dry eye status, as described in discomfort concordance to signs (DCS) grouping ( Figure 1). Group 1 (D -S -)-presumed normal or control subjects no discomfort D -(OSDI score: <12) and no clinical signs S -(TBUT: ≥10 s, Schirmer's test: ≥10 mm/5 min, no ocular surface abnormality). Group 2 (D -S + ) includes subjects without discomfort D -(OSDI score: normal < 12) but with signs S + (TBUT: <10 s and/or Schirmer's test: <10 mm/5 min and/or ocular surface staining). Group 3 (D + S + ) includes subjects with similar grades of discomfort D + and signs S + as per the standard gradings discussed previously such as moderate symptoms on OSDI and moderate severity of signs on DED evaluation. Group 4 (D ++ S +/-) includes subjects with disparate symptoms and signs, having symptom grade very high D ++ and signs low or absent S +/-. IVCM parameters (cDC density, sub-basal nerve plexus morphologic characteristics, and microneuroma-like structures) and tear fluid soluble factor levels were studied across all 4 groups.

In Vivo Confocal Microscopy (IVCM) Imaging
IVCM is a non-invasive imaging modality that can study the corneal microscopic structure, nerves, and cDC density and morphology [28,29]. IVCM was performed using the Rostock Corneal Module/Heidelberg Retina Tomograph ll (RCM/HRT ll; Heidelberg Engineering GmBH, Dossenheim, Germany). Proparacaine drops (0.5%) were instilled prior to the procedure to anaesthetize the ocular surface. The captured ICVM images that passed image quality control were used for analysis. cDC were identified based on their morphology and categorized into mature and immature [30] as shown in Figure 2a-c. cDC density (cDCD) expressed in cells/mm 2 was quantified using Cell Count software (Heidelberg Engineering GmbH) [31,32]. Sub-basal nerve plexus (SNBP) features (Figure 2d-f) in the study subjects were also determined as previously described [32]. Corneal nerve fibre area (CNFA) per square millimetre, corneal nerve fibre length (CNFL) in millimetres per square millimetre, corneal nerve fibre density (CNFD) per square millimetre, corneal nerve fibre width (CNFW) per square millimetre, corneal nerve branch density (CNBD) per square millimetre, and total branch density (CTBD) per square millimetre were analysed in IVCM images using Automatic CCMetrics software, version 1.0 (University of Manchester, Manchester, UK) [31,32]. Microneuroma-like structures in the nerve plexus were identified (Figure 2g,h) as described in existing literature [1,[33][34][35].

In Vivo Confocal Microscopy (IVCM) Imaging
IVCM is a non-invasive imaging modality that can study the corneal microscopic structure, nerves, and cDC density and morphology [28,29]. IVCM was performed using the Rostock Corneal Module/Heidelberg Retina Tomograph ll (RCM/HRT ll; Heidelberg Engineering GmBH, Dossenheim, Germany). Proparacaine drops (0.5%) were instilled prior to the procedure to anaesthetize the ocular surface. The captured ICVM images that passed image quality control were used for analysis. cDC were identified based on their morphology and categorized into mature and immature [30] as shown in Figure 2a-c. cDC density (cDCD) expressed in cells/mm 2 was quantified using Cell Count software (Heidelberg Engineering GmbH) [31,32]. Sub-basal nerve plexus (SNBP) features ( Figure  2d-f) in the study subjects were also determined as previously described [32]. Corneal nerve fibre area (CNFA) per square millimetre, corneal nerve fibre length (CNFL) in millimetres per square millimetre, corneal nerve fibre density (CNFD) per square millimetre, corneal nerve fibre width (CNFW) per square millimetre, corneal nerve branch density (CNBD) per square millimetre, and total branch density (CTBD) per square millimetre were analysed in IVCM images using Automatic CCMetrics software, version 1.0 (University of Manchester, Manchester, UK) [31,32]. Microneuroma-like structures in the nerve plexus were identified (Figure 2g,h) as described in existing literature [1,[33][34][35].

Tear Fluid Collection
Tear fluid samples were collected from the study subjects using Schirmer's strips a per standard protocol [36] and stored in 1.5 mL microcentrifuge tubes at −80 °C unt further processing. On the day of analysis, the tear fluid was extracted from Schirmer' strips by agitation in 300 µL sterile 1× PBS for 2 h at 4 °C. The tear fluid proteins wer eluted by centrifugation and immediately used for the downstream experiment a

Tear Fluid Collection
Tear fluid samples were collected from the study subjects using Schirmer's strips as per standard protocol [36] and stored in 1.5 mL microcentrifuge tubes at −80 • C until further processing. On the day of analysis, the tear fluid was extracted from Schirmer's strips by agitation in 300 µL sterile 1× PBS for 2 h at 4 • C. The tear fluid proteins were eluted by centrifugation and immediately used for the downstream experiment as mentioned in Section 2.4 to measure the various soluble factors by multiplex ELISA.

Soluble Factors Level Measurement
The concentration of secreted factors in the tear fluid that was measured include Interleukin (IL)-1α, IL-1 β, IL-2, IL-6, IL-8, IL-10, IL-17A, IFNγ (Interferon gamma), MCP1/CCL2 (Monocyte chemoattractant protein 1), RANTES/CCL5 (Regulated upon Activation, Normal T cell Expressed), sICAM1 (soluble Intercellular adhesion molecule 1), and TNFα (Tumor necrosis factor alpha and VEGF-A (Vascular endothelial growth factor-A). The levels of these secreted factors in the tears were measured by multiplex ELISA using cytometric bead array (BD Biosciences, San Jose, CA, USA) on a flow cytometer (BD FACSCantoII, BD Biosciences, San Jose, CA, USA) [36]. BD FACSDiva software (BD Biosciences, San Jose, CA, USA) was used to acquire the bead-antibody conjugate-analyte complexes and record their signal intensities. FCAP array version 3.0 (BD Biosciences, San Jose, CA, USA) was used to determine absolute concentration of the analytes using respective standards. The wetting length of the Schirmer's strip noted during tear collection and tear protein elution buffer volume were used for calculation of the dilution factor to derive the normalized concentration values of tear analytes.

Statistical Analyses
The normality of data was assessed by the Shapiro-Wilk normality test. The Kruskal-Wallis test with Dunn's multiple comparison test and the Mann-Whitney test were used to analyse the differences in the variable between the study groups in the datasets. Statistical analyses were performed with GraphPad Prism 6.0 (GraphPad Software, Inc., La Jolla, CA, USA). p < 0.05 was considered statistically significant.

Ocular Surface Clinical Parameters, IVCM Features, and Tear Soluble Factors in Different Groups
Patients were divided into the four groups based on the parameters discussed previously. Group  higher symptom grade compared to both the D -S -(group 1) and D -S + (group 2) (Figure 3a). The tear break up time (TBUT) and the Schirmer's test values were significantly lower in D + S + and D ++ S +/-compared to the normal D -Sgroup (Figure 3b,c). The TBUT and Schirmer's test values were significantly worse in the D + S + even though the symptoms were worst in the D ++ S +/-group (Figure 3b,c), which suggests that some patients in the highly symptomatic D ++ S +/-group could have only mild or no DED. A total of 201 out of 229 eyes had varying grades of DED. Ocular surface staining was noted in 13.8% of all eyes included in the study (35/287 eyes) of which most had varying grades of DED. Additional signs of ocular surface inflammation such as congestion or staining were also noted.     The OSDI scores were significantly higher in D ++ Scompared to D -S -(control) groups. (Figure 7a). The TBUT and Schirmer's test values were within the normal range ( Figure  7b,c). The density of total, immature, and mature forms of cDC were significantly higher in D ++ Scompared to the control D -Sgroup (Figure 8a). The corneal nerve fibre density was observed to be significantly higher in D ++ Scompared to D -S - (Figure 8c). No other significant difference in the various corneal sub-basal nerve plexus morphological parameters was observed between the groups (Figure 8b,d-g). The proportion of subjects and eyes with microneuroma-like structures were significantly higher in the D ++ Sgroup 4 compared to the D -S - (Figure 9). It is to be noted that the VEGF-A level was markedly lower (p = 0.07) in D ++ Scompared to D -Sgroups (Table 1). However, no significant differences in the tear soluble factors were observed between D ++ Scompared to D -Sgroups (Table 1).

Ocular Surface Clinical Parameters, IVCM Features, and Tear Soluble Factors in Subjects with High Ocular Surface Discomfort but no Clinical Signs (D ++ S -)
The OSDI scores were significantly higher in D ++ Scompared to D -S -(control) groups. (Figure 7a). The TBUT and Schirmer's test values were within the normal range (Figure 7b,c). The density of total, immature, and mature forms of cDC were significantly higher in D ++ Scompared to the control D -Sgroup (Figure 8a). The corneal nerve fibre density was observed to be significantly higher in D ++ Scompared to D -S - (Figure 8c). No other significant difference in the various corneal sub-basal nerve plexus morphological parameters was observed between the groups (Figure 8b,d-g). The proportion of subjects and eyes with microneuroma-like structures were significantly higher in the D ++ Sgroup 4 compared to the D -S - (Figure 9). It is to be noted that the VEGF-A level was markedly lower (p = 0.07) in D ++ Scompared to D -Sgroups (Table 1). However, no significant differences in the tear soluble factors were observed between D ++ Scompared to D -Sgroups (Table 1).  . Corneal dendritic cell density and sub-basal nerve plexus feature profile in subjects with discomfort but no signs. Bar graphs represent (a) profile of total, immature and mature forms of corneal dendritic cell density (cDCD), (b) corneal nerve fibre area-CNFA, (c) corneal nerve fibre density-CNFD, (d) corneal nerve fibre length-CNFL, (e) corneal nerve fibre width-CNFW, (f) corneal nerve branch density-CNBD, and (g) corneal total branch density-CTBD determined in laser scanning in vivo confocal microscopy images in the study subjects. D -Sindicates subjects without discomfort/symptoms and signs and D + Sindicates subjects with discomfort/symptoms but       The adjacent table provides the absolute number of eyes with and without microneuroma-like structures in the different groups. D -Sindicates subjects without discomfort/symptoms and signs and D + Sindicates subjects with discomfort/symptoms but no signs. Chi-square test was performed to determine the statistical significance of the difference in the frequency of microneuroma-like structures between the groups. p < 0.05 is considered to be statistically significant. MN-microneuroma-like structures.

Ocular Surface Clinical Parameters, IVCM Features and Tear Soluble Factors in Subjects with DED (Evaporative or Aqueous Deficient)
The age distribution of subjects in the various groups are as follows: median (range)controls 32.5 (25-73) years, evaporative dry eye disease-EDED 35.5 (20-65) years, and aqueous deficient dry eye disease-ADED 46 (23-72) years. The sex distribution of subjects in the various groups are as follows: controls-M/F 16/14, EDED-M/F, 35/39, and ADED-M/F 13/11. The OSDI scores were significantly higher in both EDED and ADED subjects compared to the controls, and the OSDI score in ADED was significantly higher compared to EDED as well (Figure 10a). The TBUT and Schirmer's test values were significantly lower in EDED and ADED subjects compared to the controls (Figure 10b,c). The density of total, immature, and mature forms of cDC were significantly higher in both EDED and ADED compared to the controls (Figure 11a-c). The density of the total and immature forms of cDC were significantly higher in ADED compared to EDED (Figure 11a,b). No significant difference in the sub-basal nerve plexus parameters in DED compared to the controls was found (Supplementary Figure S1). However, the proportion of subjects and eyes with microneuroma-like structures in corneal nerves were significantly higher in DED compared to the controls (Supplementary Figure S2). The levels of IL-6, IL-17A, RANTES, and MCP1 were significantly higher and VEGF-A significantly lower in EDED and/or ADED compared to the controls ( D -S -(subjects without discomfort/symptoms and without signs-OSDI score: <12, TBUT ≥ 10 s, Schirmer's test 1: ≥10 mm/5 min); D + S -(subjects with discomfort/symptoms and without signs-OSDI score: >12, TBUT ≥ 10 s, Schirmer's test 1: ≥10 mm/5 min); SD-standard deviation; SEMstandard error mean; D -S -(n = 28 eyes); D + S -(n = 8 eyes); p < 0.05 is statistically significant, Mann-Whitney Test.

Ocular Surface Clinical Parameters, IVCM Features and Tear Soluble Factors in Subjects with DED (Evaporative or Aqueous Deficient)
The age distribution of subjects in the various groups are as follows: median (range)-controls 32.5 (25-73) years, evaporative dry eye disease-EDED 35.5 (20-65) years, and aqueous deficient dry eye disease-ADED 46 (23-72) years. The sex distribution of subjects in the various groups are as follows: controls-M/F 16/14, EDED-M/F, 35/39, and ADED-M/F 13/11. The OSDI scores were significantly higher in both EDED and ADED subjects compared to the controls, and the OSDI score in ADED was significantly higher compared to EDED as well (Figure 10a). The TBUT and Schirmer's test values were significantly lower in EDED and ADED subjects compared to the controls (Figure 10b,c). The density of total, immature, and mature forms of cDC were significantly higher in both EDED and ADED compared to the controls (Figure 11a-c). The density of the total and immature forms of cDC were significantly higher in ADED compared to EDED (Figure 11a,b). No significant difference in the sub-basal nerve plexus parameters in DED compared to the controls was found (Supplementary Figure S1). However, the proportion of subjects and eyes with microneuroma-like structures in corneal nerves were significantly higher in DED compared to the controls (Supplementary Figure S2). The levels of IL-6, IL-17A, RANTES, and MCP1 were significantly higher and VEGF-A significantly lower in EDED and/or ADED compared to the controls ( Table 2).    Controls: n = 28 eyes; EDED (Evaporative dry eye disease): n = 45 eyes; ADED (Aqueous deficient dry eye disease): n = 6 eyes; SD-standard deviation; SEM-standard error mean; p < 0.05 is statistically significant, Mann-Whitney Test.

Discussion
Our understanding of acute and chronic ocular surface discomfort and pain continues to evolve. In patients who have a clinically attributable and comparable symptoms, the treatment is more straight forward and directed at the underlying aetiology. However, at the cornea clinic, we observe a large number of patients who have a disparity between clinical signs and symptoms where the severity of the symptoms cannot be explained by clinically visible signs. This can be due to an underlying nociceptive and neuropathic pain component [35]. Neuropathic corneal pain may not have clinically evident signs and has been classically referred to as "pain without stain"

Discussion
Our understanding of acute and chronic ocular surface discomfort and pain continues to evolve. In patients who have a clinically attributable and comparable symptoms, the treatment is more straight forward and directed at the underlying aetiology. However, at the cornea clinic, we observe a large number of patients who have a disparity between clinical signs and symptoms where the severity of the symptoms cannot be explained by clinically visible signs. This can be due to an underlying nociceptive and neuropathic pain component [35]. Neuropathic corneal pain may not have clinically evident signs and has been classically referred to as "pain without stain" [12]. Thus, the pathophysiology of such exaggerated discomfort could be related to ocular surface inflammation, altered nociception, or neuropathy [1,37]. The perception of pain on the ocular surface is likely an interplay between the structural, epithelial, neuronal, molecular, and immune cell changes in the eye and neuronal connections to higher centres in the thalamus and somatosensory cortex [38,39]. Structurally, as the cornea has the highest nerve density in the body, it is particularly sensitive to alterations in the local molecular factors and environmental influences [40]. The status of features such as the cDC density, the sub-basal nerve plexus, microneuroma-like structures, and tear fluid factors needs to be characterized comprehensively in subjects with discordant signs and symptoms. The current study addresses this knowledge gap by objectively stratifying patients based on the discordance between the severity of ocular surface signs and symptoms and determining how cornea-specific features vary within groups.
DED status, and ocular surface clinical status and discomfort (OSDI) scores for each patient were evaluated. The chronicity of the pain and the patient's description of the discomfort or pain were also taken into consideration [14]. When evaluating the severity of symptoms across the groups we found that the OSDI score was significantly higher in D + S + (group 3) and D ++ S +/-(group 4) as compared to D -S -(group 1, normal) subjects. The cause for the increased pain in this group of patients could be due to the inflammation, altered nociception, or neuropathy. Chronic DED and the associated inflammation are common causes for increased nociception or NCP. The clinical signs were most significantly altered in the D + S + group 3 (where the severity of symptoms and signs are proportionate). In the D ++ S +/group 4, where symptoms are out of proportion to signs, the alteration to DED clinical parameters were relatively less severe. These findings also corroborate those seen in other studies that state that increased nociception and NCP can be associated with different ocular and systemic conditions [41,42]. This reiterates the hypothesis that there may be additional factors driving the discomfort in this discordant group, hence we analysed IVCM and tear molecular factors across the cohort.
We found a significant increase in cDC density in D + S + group 3 and D ++ S +/group 4 compared to normal D -S -. The highest cDC density was seen in the D ++ S +/group, thereby implying that the cDC density correlates positively with the patient's discomfort. This finding is supported by previous studies that reported an association of cDC density with increased inflammation and discomfort [36,43]. The cDC may proliferate within a tissue and secrete a variety of inflammatory factors that can raise the overall nociceptive response [44]. Corneal sub-basal nerve fibre features have been studied in ocular surface conditions including DED with varied observations [45]. A significant increase in cDC density has been reported in a metanalysis on IVCM features similar to the observations made in the current study [46]. However, this meta-analysis reports conflicting observations with reference to corneal nerve features between different studies. Though corneal nerve feature-related changes such as increased tortuosity, beading, looping, and decreased nerve fibre density, etc., have been reported in various studies using IVCM [22,43], we did not find a significant difference across these nerve parameters assessed in the SBNP across the groups of our cohort. Possible ethnicity variation and measurement strategies (manual versus algorithm based) could contribute to differences in observations. Microneuromalike structures, which are terminal enlargements of nerve endings with variable shape and hyperreflectivity [47], can be observed in the sub-basal nerve plexus or stroma of the cornea [35,43]. We found a significantly higher frequency of microneuroma-like structures among the subjects in the highly symptomatic D ++ S +/-(group 4) compared to the other groups in our study. In subjects with the classic presentation of "pain without stain" consistent with NCP, the IVCM analysis revealed a higher proportion of microneuroma-like structures, similar to previous reports [34,35]. Our cohort also had a few asymptomatic subjects who were found to have microneuroma-like structures in their sub-basal nerve plexus, contrary to previous reports that did not show this feature in normal subjects [35]. This suggests that even though the microneuroma-like structures are strongly associated with NCP, the patient symptoms may be dependent on an interplay between inflammatory, structural, and neuronal components.
In addition to their role in modulating inflammation on the ocular surface, the dysregulation of certain molecular factors can have nociceptive potential as well. We have previously demonstrated that an altered balance between the pro-and anti-nociceptive tear soluble factors can contribute to the increased symptomatology in DED [36]. It has been shown that the ocular neurosensory pathway has various structural and molecular components and dysregulation of any of these can result in corneal hyperalgesia [48]. Inflammatory factors can sensitize thermoreceptors and mechano-nociceptors and reduce the threshold to pain stimuli [49]. Receptors of IL-17A, have been shown to be expressed by nociceptor neurons and, hence, play a role in pain perception by the altered expression of TRPV4 channels [50]. We found IL-17A, a pro-nociceptive factor, significantly increased in the D + S + and D ++ S +/groups, with much higher levels in the latter. VEGF-A, in addition to being an angiogenic factor, also has anti-nociceptive potential [51,52]. In our study we found that the levels of anti-nociceptive factor VEGF-A were reduced in the D + S + group, suggesting a tip in balance towards increased pro-nociception in these groups. The increased symptomatology in these groups could be due to this altered balance between pro-and anti-nociceptive factors on the ocular surface. It is pertinent that multicentric studies across different ethnicities and larger cohort be conducted to validate our findings to provide a clinically actionable algorithm that uses these patient specific corneal features for stratifying patients with ocular surface pain. Treatment targeted at improving the nociceptive balance could help improve patient symptoms and long-term comfort [34].
Currently, the treatment options for patients presenting with discordant signs versus symptoms are limited to the following: (a) artificial tear supplements along with anti-inflammatory agents such as topical steroids, (b) nonsteroidal medications such as cyclosporine A and tacrolimus, and (c) biologicals such as Anakinra (human IL-1 receptor antagonist). Further, autologous serum eye drops have also been shown to have a positive effect in some cases as they contain multiple factors such as nerve growth factor (NGF) and epidermal growth factor (EGF) that support nerve growth and epithelial health. In patients who have a central component of pain, systemic medications such as tricyclic antidepressants (e.g., amitriptyline, nortriptyline), anticonvulsants (e.g., Gabapentin, carbamazepine), and serotonin uptake inhibitors (e.g., Duloxetine and venlafaxine) may help relieve symptoms [34].

Conclusions
The assessment of ocular pain based on questionnaires and correlating it to clinical metrics and confocal features such as cDC density and microneuroma-like structures can help clinicians better classify the condition and aid in customized treatment planning. The discordance between patient reported ocular surface pain/discomfort and clinical signs is associated with enhanced cDC, presence of microneuroma-like structures imaged by IVCM, as well as increased IL-17A ( Figure 12). This study highlights the importance of scoring the disparity in symptoms and signs and IVCM as important clinical tools to stratify patients for targeted treatment.    Informed Consent Statement: Written informed consent was obtained from all subjects prior to inclusion in the study.