Visualization of Micro-Neuromas by Using In Vivo Confocal Microscopy: An Objective Biomarker for the Diagnosis of Neuropathic Corneal Pain?

Abstract

Purpose:

The diagnosis of neuropathic corneal pain (NCP) is challenging, as it is often difficult to differentiate from conventional dry eye disease (DED). In addition to eye pain, NCP can present with similar signs and symptoms of DED. The purpose of this study is to find an objective diagnostic sign to identify patients with NCP, using in vivo confocal microscopy (IVCM).

Methods:

This was a comparative, retrospective, case-control study. Patients with clinical diagnosis of NCP (n=25), DED (n=30), and age and sex-matched healthy controls (n=16), who underwent corneal imaging with IVCM (HRT3/RCM) were included. Central corneal IVCM scans were analyzed by 2 masked observers for nerve density and number, presence of micro-neuromas (terminal enlargements of subbasal corneal nerve) and/or nerve beading (bead-like formation along the nerves), and dendritiform cell (DC) density.

Results:

There was a decrease in total nerve density in both NCP (14.14±1.03 mm/mm²) and DED patients (12.86±1.04 mm/mm²), as compared to normal controls (23.90±0.92 mm/mm²; p <0.001). However, total nerve density was not statistically different between NCP and DED patients (p=0.63). Presence of nerve beading was not significantly different between patients and normal controls (p=0.15). Interestingly, micro-neuromas were observed in all patients with NCP, while they were not present in any of the patients with DED (sensitivity and specificity of 100%). DC density was increased significantly in both NCP (71.89±16.91 cells/mm²) and DED patients (111.5±23.86 cells/mm²), as compared to normal controls (24.81±4.48 cells/mm²; p<0.05). However, there was no significant difference in DC density between DED and NCP patients (p=0.31).

Conclusion:

IVCM may be used as an adjunct diagnostic tool for the diagnosis of NCP in the presence of neuropathic symptoms. Micro-neuromas may serve as a sensitive and specific biomarker for the diagnosis of NCP.

INTRODUCTION

Neuropathic pain is defined by the International Association for the Study of Pain as ‘pain caused by a lesion or disease of the somatosensory system’.¹ The prevalence of neuropathic pain is estimated to be around 7-10%.² Neuropathic pain can occur in the cornea, and has been called “neuropathic corneal pain” (NCP), ^3–5 “corneal neuralgia”,⁶ or “ocular neuropathic pain”³. The prevalence and etiology of NCP are currently not known; however, dry eye disease (DED), infectious keratitis, and refractive surgeries are among the most commonly associated causes.⁵

Accurate diagnosis of NCP is paramount as most patients with NCP do not respond well to conventional DED therapies and thus the disease causes a significant decrease in patients’ quality of life.^5,6 Currently, the diagnosis of NCP is based on detailed history, lack of ocular surface signs on ophthalmic slit-lamp examination, validated pain questionnaires, and evidence of nerve injury.^5,7,8 Additionally, corneal esthesiometry and functional corneal nerve testing have been proposed tools to aid in diagnosis,^5,7 but are not applicable yet.

There is currently no specific objective criteria or sign for the diagnosis of NCP.^4,9 In vivo confocal microscopy (IVCM) is a useful tool, which provides a quasi-histological visualization of corneal nerves and dendritiform cells (DC), and can assist clinicians in diagnosis and management of ocular surface diseases.^10,11 Moreover, IVCM has been used in the assessment of systemic neuropathies, in particular for diabetic neuropathy.^2,12 Using IVCM, we have previously demonstrated a decrease in nerve density, presence of micro-neuromas (terminal enlargements of subbasal corneal nerve) and an increase in beading (small beads along the nerve fibers) in NCP patients.^13–15 However, the specificity and sensitivity of these findings have not been studied to date. We hypothesized that in patients with clinical symptoms of NCP we can find specific signs by IVCM to aid in the diagnosis of NCP. Therefore, using IVCM, we aimed to investigate whether corneal nerve and DC changes in NCP patients in comparison with DED patients and normal controls may aid in the identification of a specific and/or sensitive objective sign for the diagnosis of NCP.

METHODS

Study Design and participants:

This is a comparative, retrospective, chart review, case-control study that was conducted between 2016-2018 at Tufts Medical Center, Boston, MA. The study was approved by the Institutional Review Board (IRB) of Tufts Medical Center. The study was compliant with Health Insurance Portability and Accountability Act, adhered to the tenets of the Declaration of Helsinki. Electronic medical records of individuals, with clinical diagnosis of DED or NCP who had IVCM performed for clinical purposes, were studied. Age- and sex-matched reference controls in this study were healthy, asymptomatic individuals, with no ocular pathology, absent ocular surface staining, and tear film break-up time of more than 10 seconds. Controls had no history of infectious or inflammatory eye diseases, diabetes, systemic inflammatory diseases, ocular surgery, and no current use of any eye drops or contact lens wear. All reference controls were selected from an IRB-approved prospective normative study database that enrolled healthy subjects after having a complete history and ocular examination. DED patients were identified as patients with DED related symptoms of more than 6 months (including dryness, soreness, discomfort, foreign body sensation, blurred vision, irritation, itching etc.), but absence of ocular pain (evaluated using a validated Ocular Pain Assessment Survey questionnaire⁸), along with, presence of 2 or more clinical signs of DED, including corneal fluorescent staining ≥ 1+ (using 0-5 Oxford scale), tear breakup time (TBUT) < 10 sec, or Schirmer’s score II<10 mm. Patients with the chief complaint of ocular pain and/or light sensitivity, TBUT ≥ 10 sec or Schirmer’s score II>10, and with or without a maximum of trace corneal fluorescent staining, were included in NCP group. Clinical data was retrieved from electronic medical records and relevant IVCM images were retrieved.

In Vivo Confocal Microscopy

The central cornea was scanned with laser in vivo confocal microscopy (Heidelberg Retina Tomograph (HRT) 3 and the Rostock Cornea Module (RCM), Heidelberg Engineering GmbH, Heidelberg, Germany) in all subjects at the time of diagnosis upon presentation, as previously described.¹⁶ This microscope is equipped with a 63X objective immersion lens with a numerical aperture of 0.9 (Olympus, Tokyo, Japan) and uses a 670-nm red wavelength diode laser source. The microscope records a 400 X 400 μm coronal section of the cornea, which is 160,000 μm², at a selectable corneal depth and is separated from adjacent images by approximately 1 to 4 μm lateral resolution and an axial resolution of 1 μm/pixel.

A disposable sterile polymethylmethacrylate cap (Tomo-Cap; Heidelberg Engineering GmbH, Heidelberg, Germany), filled with a layer of hydroxypropyl methylcellulose 2.5% (GenTeal gel, Alcon, Ft. Worth, TX), was mounted in front of the Rostock Cornea Module optics for each examination. One drop of topical anesthesia, 0.5% proparacaine hydrochloride (Alcaine, Novartis Ophthalmics, East Hanover, NJ), was instilled in both eyes, followed by application of GenTeal gel in both eyes. GenTeal gel was also placed on the outside tip of the Tomo-Cap to improve optical coupling, and manually advanced until the gel contacted the central surface of the cornea.

A total of 6-8 volume and sequence scans per time point were obtained from the center of each cornea, mapping the central cornea with at least 6 sequence scans, acquiring non-overlapping areas with particular focus on the subepithelial area and the subbasal nerve plexus, typically at a depth of 50 to 80 μm.

Image selection and Analysis

One random eye was selected for image analysis for each studied group. Three separate representative central corneal IVCM images were selected each for analysis of DCs density and corneal nerve density. The criteria for image selection were best-focused images, comprising one single layer, no overlapping images, without motion folds, from the layer immediately at or posterior to the basal epithelial layer and anterior to the Bowman’s layer by one observer (G.D.). Then selected images were masked (using numerical and alphabetical letters for each image) (A.J.) and 2 masked observers (H.M. and A.A.) analyzed the images for DC density and nerve number/density. Bright dendritiform hyperreflective cell bodies were identified as DCs. DC density was calculated manually using ImageJ¹⁷ (National Institutes of Health, Bethesda, MD). Corneal nerve analysis was performed as described before.^15,18 In brief, corneal nerves were traced using NeuronJ, a free semi-automatic plug-in for ImageJ (http://www.imagescience.org/meiiering/software/neuroni). Main nerve trunks were defined as nerves in one image that did not branch from other nerves. Nerve branches were counted as other remaining nerves. Total number of nerves was calculated as the sum of nerve trunks and branches in each image. Nerve density were calculated and reported as total nerve length/frame (mm/mm²). The average of 2 observers’ values was used for final report and analysis. In case of discrepancy by more than 10%, a third senior observer (GD) analyzed the images and images were adjudicated.

To analyze nerve morphology (micro-neuromas and nerve beading), full-thickness confocal scans comprised of an average 300 images from all layers of the cornea were screened for each selected eye. Micro-neuromas were defined as nerve abnormalities that present as irregularly shaped, terminal enlargements of subbasal nerve ending(s) with variable hyperreflectivity.^13–15 Beading was defined as discrete small bead-like formations along the length of the nerves.^13–15

Statistical Analysis

Continuous variables were expressed as mean ± standard error of the mean. Normality of data was evaluated using Kolmogorov-Smirnov test. Chi-square test was used to compare qualitative data. Student’s t-test was used for comparison of two groups. One-way ANOVA with Bonferroni post hoc was used to compare more than two groups. Intraclass correlation coefficient (ICC) was calculated between two masked observers for DC density and nerve density. Moreover, inter-grader and intra-grader agreements were calculated for detection of micro-neuromas. Statistical analysis was performed using SPSS version 18 (SPSS Inc. Chicago, IL) and a p-value of less than 0.05 was considered significant.

RESULTS

Twenty-five patients (n=25 eyes) with NCP, 30 patients with DED (n=30 eyes), and 16 healthy volunteers (n=16 eyes) were included in the study. Demographics of studied subjects are presented in Table 1. Risk factors of NCP in our patients were as follows: DED (11 patients), blepharitis (4 patients), cataract surgery (4 patients), allergic conjunctivitis (3), photorefractive keratectomy (2 patients), laser in-situ keratomileusis (2 patients); meibomian gland dysfunction (2 patients), trigeminal neuralgia (2 patients), pterygium (1 patient), rosacea (1 patient), and myasthenia gravis (1 patient) with some patients having more than one risk factor.

Table 1:

Demographics and clinical signs of studied patients

	Neuropathic Corneal Pain	Dry Eye Disease	Controls	p
Number of patients/eyes	25/25	30/30	16/16	-
Age (years)	48.32±3.33	54.80±2.66	52.38±1.56	0.24
Sex (male/female)	8/17	6/24	5/11	0.54
Tear break-up time (Sec)	10.86±0.70	2.99±0.44	>10.00	<0.0001
Schirmer’s score II (mm)	14.11±1.46	4.88±0.61	>10.00	<0.0001
Corneal fluorescein staining (0-5 Oxford scale)	0.00	1.73±0.17	0.00	<0.0001
OSDI	46.22±6.72	35.09±5.10	<5.00	0.18

Values are presented as mean ± standard error of the mean. OSDI, Ocular surface disease index.

Corneal nerve density and numbers

Sample IVCM scans for NCP, DED, and the control group are represented in Figure 1. Total nerve density was lower in both NCP patients (14.14±1.03 mm/mm²) and DED (12.86±1.04 mm/mm²), as compared to normal controls (23.90±0.92 mm/mm²; p <0.0001). There was no statistically significant difference in total nerve density between DED and NCP patients (p=0.63). Main nerve trunk density was 6.52±0.75 and 5.44±0.64 mm/mm2 in NCP and DED patients, respectively (p=0.45). Branch density was 7.79±0.68 and 8.21±0.89 mm/mm² in NCP and DED patients, respectively (p=0.91, Fig. 2, Table 2).

Figure. 1.

Representative in vivo confocal microscopy (IVCM) images of patients and controls to demonstrate corneal nerve changes. A) Subbasal corneal nerve plexus in a neuropathic corneal pain (NCP) patient. White arrow demonstrates a micro-neuroma. B) Subbasal corneal nerve plexus in a patient with painless dry eye disease (DED). C) Subbasal corneal nerve plexus in a healthy control subject. D) Sample nerve tracing in a healthy subbasal corneal nerve plexus (C). Yellow lines represent main nerve trunks and red lines represent nerve branches.

Figure 2.

Corneal nerve density in patients and controls. Error bars represents standard error from the mean. *P= 0.001 compared to control group. **P<0.0001 compared to control group. ns, represents not statistically significant. ANOVA multivariate analysis.

Table 2.

In vivo confocal microscopy (IVCM) parameters in studied subjects

IVCM parameters	Neuropathic Corneal Pain N=25	Dry Eye Disease N=30	Controls N=16	P
Total nerve density (mm/mm2)	14.14±1.03*	12.86±1.04*	23.90±0.92	<0.0001
Main nerve trunk density (mm/mm2)	6.52±0.75*	5.44±0.64*	10.40±0.50	<0.0001
Branch nerve density (mm/mm2)	7.79±0.68*	8.21±0.89*	13.5±0.65	<0.0001
Total nerve number (n/frame)	11.13±0.86*	10.35±0.91*	28.78±1.40	<0.0001
Main nerve number (n/frame)	2.69±0.31*	2.17±0.26*	4.34±0.23	<0.0001
Branch nerve number (n/frame)	8.42±0.74*	8.14±0.78*	24.43±1.50	<0.0001
Dendritiform cell density (cells/mm2)	71.89±16.91	111.5±23.86*	24.81±4.48	0.02
Beading (%)	25/25 (100%)	30/30 (100%)	15/16 (93.7%)	0.175
Microneuroma (%)	25/25 (100%) *†	0/30 (0%)	0/16 (0%)	<0.001

P values are calculated by on one-way ANOVA for all variables except beading and neuroma, which were calculated by chi-squared test.

^*Statistically significant difference as compared to control group

^†Statistically significant difference as compared to dry eye disease

The total number of nerves was 11.13±0.86, 10.35±0.91, and 28.78±1.40 n/frame in the NCP, DED, and control group respectively (Fig. 3). The nerve numbers were significantly lower in both DED and NCP patients, as compared to normal controls (p<0.0001). However, there was no significant difference in corneal nerve numbers between DED and NCP patients (p=0.83, Fig. 3). The mean number of main nerve trunks was 2.69±0.31, 2.17±0.26, and 4.34±0.23 n/frame in the NCP, DED, and control group respectively. Main nerve trunk number was significantly lower in NCP and DED as compared to control group (p=0.001 and p<0.0001, respectively). However, there was no significant difference between NCP and DED (p=0.35; Fig. 3). The mean number of nerve branches was 8.42±0.74, 8.14±0.78, and 24.43±1.50 n/frame in the NCP, DED, and control group respectively. Similarly, nerve branch number was also significantly lower in NCP and DED as compared to control group (p<0.0001 for both). There was no significant difference between NCP and DED (p=0.97; Fig. 3, Table 2).

Figure 3.

Comparison of nerve numbers between patients and controls. Error bars represents standard error from the mean. *P= 0.001 compared to control group. **P<0.0001 compared to control group. ns, represents not statistically significant. ANOVA multivariate analysis.

There was a high agreement between observes in calculating nerve density (Intraclass correlation coefficient (ICC) of 0.98).

Corneal nerve morphology

Micro-neuromas were not observed in IVCM images of controls (Fig. 1A) or DED patients (Fig. 1). However, micro-neuromas, which can be very rare (as low as 1-2 in all acquired scans per patient), were observed in all patients with NCP (p<0.0001). In contrast, nerve beading was present in 15 out 16 normal control subjects, as well as in all DED and NCP patients (p=0.15; Fig. 1, Table 2). Sensitivity and specificity of micro-neuromas to detect NCP was 100% in this study. Inter-observer and intra-observer reliability for detection of micro-neuroma were good (kappa= 0.79 and 0.86, respectively). There was a high agreement between observes in calculating nerve density (ICC= 0.94).

Dendritiform cell density

There was no significant difference in DC density between NCP (71.89±16.91 cells/mm2) and controls (24.81±4.48 cells/mm²; p=0.30). However, DC density was significantly higher in the DED group (111.5±23.86 cells/mm²) as compared to normal controls (p=0.01). There was no significant difference between DED and NCP patients (p=0.31; Fig. 4 and Fig. 5) Although, the mean of DC density was higher than two standard deviations from the mean of DC density in control group in 53.3% of DED patients versus 28.0% of NCP patients. There was a high agreement between two observers in analysis of IVCM scans for DC density (ICC= 0.94)

Figure 4.

Representative in vivo confocal microscopy (IVCM) images of patients and controls to demonstrate dendritiform cell (DCs) changes. A) Presence of DCs in a patient with neuropathic corneal pain. B) DC density in a patient with dry eye disease shows a marked increase in comparison with normal controls. C) DCs are present as bright hyperreflective cell bodies in a normal control subject.

Figure 5.

Comparison of dendritiform cell density in patients and controls. Error bars represents standard error from the mean. *P=0.01 dry eye disease (DED) compared to control group. There was no significant difference between neuropathic corneal pain (NCP) patients and normal controls as well as NCP and DED patients. ANOVA multivariate analysis.

DISCUSSION

Currently, there is no available objective criteria or diagnostic biomarker to confirm diagnosis of NCP, as it exists for non-ocular neuropathic pain in the form of punch skin biopsies.^4,5,19 Rosenthal et al. previously proposed that corneal nerve changes by IVCM may potentially aid in the diagnosis of NCP.²⁰ However, they only presented sample IVCM images from one NCP patient, demonstrating loss of intact axons and increased mature dendritic cells as compared to a normal age-matched patient.²⁰ Similarly, our group has previously described corneal nerve changes, including the presence of micro-neuromas by IVCM scans of patients with NCP.^4,5,15 However, sensitivity and specificity of corneal nerve changes in NCP patients, as well as its comparison with DED patients have not been studied to date. Herein, we directly compare IVCM scans from a cohort of patients, clinically diagnosed as NCP (without clinical signs of DED) and DED (without pain), and compared these to each other and to healthy controls. We demonstrate that micro-neuromas as visualized by IVCM, may potentially serve as a highly sensitive and specific diagnostic biomarker to confirm the diagnosis of NCP. We have shown that while all NCP patients present with micro-neuromas in their central corneal IVCM scans, none have been observed in DED without pain, or in normal controls. Interestingly, micro-neuromas were observed in only 62.5% of patients with photoallodynia in our initial study.¹⁵ This discrepancy may be explained by a difference in methodology for micro-neuroma evaluation from the previous study. In the current study, we assessed all full-thickness images available per patient (average of 300 images of the subbasal nerve layer for each eye) to evaluate micro-neuromas; our former study only evaluated the three representative images for each eye that were selected based on corneal nerve density representation. Micro-neuromas were initially described as “tangled masses” formed following incomplete axonal growth at the end of the cut nerve after transection of sciatic nerve in mice and are known to be a source of chronic pain.²¹ Corneal micro-neuromas or neuromas were defined as terminal enlargements of subbasal corneal nerves, which are formed following axonal nerve regeneration secondary to any nerve damage as demonstrated before in rabbits and cats.^22,23 Immunostaining (acetylcholinesterase staining) of rabbit corneas after excimer laser photokeratectomy confirmed the presence of regenerative nerve sprouting.²³ Micro-neuromas have been described as a possible source of discomfort and abnormal sensation in patients with NCP.²² This has been based on the fact that morphologic nerve changes culminate in a constellation of molecular changes, such as modified genes and protein expressions, which change the overall excitability of neuromas.²² This excitability ultimately may cause hyperalgesia, allodynia, and other unpleasant symptoms of NCP. ²²

Herein, we demonstrate that corneal nerve beading is neither a specific, nor sensitive marker for NCP. Beading reflects increased metabolic activity (contains high densities of mitochondria).²⁴ Beadlike-formations were present along corneal nerves in almost all of the studied subjects including 15 out of 16 controls. This finding is consistent with a study by Oliveira-Soto et al., who demonstrated beadlike formation of corneal nerves in all of their normal studied subjects.²⁵ Similarly, our group previously showed that about 94% of patients with NCP had beading in their IVCM scans.¹⁵ The assessment and quantification of beading may, however, have value in monitoring patients over time.

Moreover, we demonstrate decreased corneal nerve density in patients with NCP in comparison with normal control subjects. This finding is in line with our previous reports in patients with NCP and allodynia.^13,15 Previously, corneal nerve density has been shown to correlate with the severity of symptoms of NCP.¹⁵ However, decreased corneal nerve density cannot be used as a sensitive or specific criterion for diagnosis of NCP, as we observed similar findings in DED patients without pain. Furthermore, decreased corneal nerve density has been observed previously in DED patients and other ocular surface diseases such as herpes simplex keratitis, post refractive surgeries, or diabetic patients, among others, and thus is not a specific finding in NCP.^10,26

DC density was not significantly different in the NCP patients as compared to controls. This finding is consistent with the previously reported findings in case series of patients with NCP.^14,15 On the other hand, DC density was significantly higher in the DED group as compared to the controls. Interestingly, the number of DED patients who had higher than two standard deviations from the mean of DC density in control group, was 1.9 times of NCP patients. Inflammation, secondary to proinflammatory cytokines and neuropeptides, is common in pathogenesis of both DED and NCP.^5,7,9 However, inflammation (demonstrated as increased number of DCs) may play a more important role in the pathogenesis and symptoms of DED patients as compared to NCP patients. Our findings may also reflect that chronic state of NCP. Thus, while on average DC density changes were not significantly higher in NCP patients as compared to controls, it is possible that earlier in the disease, increased DC changes may have been present. Identifying the DC density by using IVCM could serve as an important tool to assess immune and inflammatory changes in the cornea for treatment stratification.^27,28

DED patients can present with overlapping symptoms as compared to NCP, such as ocular pain.⁹ Thus, in the current study, we excluded DED patients with ocular pain from this study. In addition, although NCP has previously been described as pain without signs of DED,⁹ we have recently shown that patients with NCP may also present with concurrent signs of DED;²⁹ hence, we excluded NCP patients who had signs of DED in their examination, including presence of corneal fluorescein staining, a low Schirmer’s score, and low TBUT. Therefore, we attempted to separate conventional DED from NCP for the current study.

Limitations of the current study are the evaluation of only the corneal center by IVCM for analysis, which cannot be necessarily extrapolated to the peripheral cornea. Further, the retrospective nature of the study, as well as the limited number of patients are another limitation of this study. Selected DED patients in this study do not necessarily represent all DED patients as DED encompass a wide range of variety in signs and symptoms, including patients with pain. However, given that NCP is rather rare, it would require multicenter studies for inclusion of larger cohorts. Nevertheless, the current study warrants additional prospective studies with larger number of patients to further validate the utility of micro-neuromas as a diagnostic biomarker.

Differentiating conventional DED from NCP can be challenging for clinicians, as they both can present with overlapping signs and symptoms and have similar underlying mechanisms.⁷ As other groups have hypothesized,^3,7 these two diseases could be different presentations of the same disease along a continuum. While DED has been shown to be one of the important underlying causes of NCP,^4,5 NCP can present with DED symptoms due to corneal nerve damage and its consequences.^7,22 Our findings may set the stage for future prospective or larger studies, would confirm reliable sensitivity and specificity for micro-neuromas in diagnosis of NCP and reveals IVCM changes throughout the course of the disease. Additional studies to confirm analytical and biological validation of this potential biomarker are warranted. In addition, detection of micro-neuromas requires training and application of standard operating procedures to map the central cornea, as well as training to distinguish them from immune cells or entry points of stromal nerves through the Bowman’s layer that occur in non-NCP patients. Current manual assessment of all IVCM images for patients for the detection of micro-neuromas is thus time consuming and utilization of automated software with artificial intelligence, which are currently underway, will be beneficial for clinical implementation. Additional clinical trials on functional sensory tests and future in office nerve function tests are warranted for more accurate diagnosis of NCP and differentiating it from DED and other conditions.

Conclusion:

In conclusion, IVCM may be utilized as a useful adjunct diagnostic tool for the diagnosis of NCP. Detection of micro-neuromas in corneal IVCM images may serve as a sensitive and specific sign and a diagnostic biomarker for NCP. This study warrants a larger prospective study to better investigate IVCM characteristics between DED and NCP patients for further biomarker validation.

Abbreviations

DED

dry eye disease

NCP

neuropathic corneal pain

IVCM

In Vivo Confocal Microscopy

DC

dendritiform cells

HRT3/RCM

Heidelberg Retina Tomograph 3 with the Rostock Cornea Module