Validation of a novel confocal microscopy imaging protocol with assessment of inter-rater correlation in patients with and without dry eyes

Abstract

Purpose : Confocal microscopy is used to image the corneal sub-basal nerve plexus but lacks an accepted standardized method. Using a new grid-based visual imaging protocol, we assessed 7 nerve plexus metrics. Inter-rater reproducibility was evaluated. Confocal sub-basal nerve metrics between dry eye and normal patients were compared.

Methods : Patients were recruited from Proctor Foundation ophthalmology and optometry practices. Participants completed 3 surveys: Dry Eye Questionnaire 5, ocular surface disease index, and neuropathic pain symptom inventory. Using a Heidelberg Confocal Microscope (HRT3), confocal microscopy of the sub-basal nerve plexus was performed following a standardized protocol. The examination was repeated by a different examiner on the same eye using the same protocol on the same day. Images were analyzed using ACCmetrics (University of Manchester, UK) automated nerve analysis program. Statistical analysis was conducted using Stata version 15.0. Intra-class correlation (ICC) was used to assess inter-rater reproducibility. Linear regression was conducted to investigate the correlation between questionnaire score and nerve metric measurements. Student’s T-test was conducted to assess the difference in mean nerve metrics between patients with dry eyes and controls.

Results : 16 dry eye patients and 10 controls were included in our analysis. Each patient contributed two sets of 30 images for a total of 1560 observations. The confocal imaging protocol demonstrated fair agreement between microscopists (Figure 1). Nerve metrics were not significantly different between dry eye patients and controls (Figure 2), and were not correlated with the scores on any of the 3 surveys (P>0.05 for each).

Conclusions : Sub-basal nerve imaging by a standardized protocol demonstrated fair agreement when conducted by 2 different microscopists. Larger studies are needed to determine whether basal nerve imaging has useful clinical correlation.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.

 

Inter-microscopist intra-class correlation (ICC) of imaging results for 7 measured nerve metrics.

 

Comparison of nerve metrics measured by ACCmetrics in dry eye patients and controls.