Reliability of Corneal Subbasal Nerve Plexus Analyses Using Semi-Automated Software

Aim: The aim of this study was to evaluate the interobserver and intraobserver reliability of quantitative
corneal subbasal nerve plexus (CSNP) analyses using semi-automa ted software.
Patients and Methods: Forty volunteers who applied to the Ophthalmology Department of the Necmettin
Erbakan University Meram Medical Faculty between 20 December 2021 and 20 January 2022 were
enrolled in the study. Images showing CSNP were obtained from the right eyes of the participants by using
Heidelberg Retina Tomograph III with Rostock Cornea Module. Three best quality images were selected
from each case. NeuronJ plugin for ImageJ software was used to trace nerve fibers and calculate nerve
fiber length (NFL), nerve fiber density (NFD), and nerve branch density (NBD). All these measurements
were performed by two different observers, and repeated for the second time by one of the observers
with an interval of one week, and interobserver and intraobserver reliability were determined using the
intraclass correlation coef ficient (ICC).
Results: The mean age of 40 participants (12 female and 28 male) was 34.70±5.86 years. The ICCs
for interobserver reproducibility were 0.967 (95% CI 0.939-0.982) for NFL, 0.826 (95% CI 0.696-0.904)
for NFD, and 0.949 (95% CI 0.906-0.973) for NBD indicating good to excellent reliability. The ICCs for
intraobserver repeatibility were 0.964 (95% CI 0.932-0.981) for NFL, 0.803 (95% CI 0.657-0.891) for NFD,
and 0.890 (95% CI 0.802-0.941) for NBD indicating good to excel lent reliability.
Conclusion: Quantitative CSNP analyzes using semi-automated software have high interobserver and
intraobserver reliability and can therefore be used in both clinical practice and clinical studies for detection
and follow-up of corneal nerves’ well-being, damage, and evaluation of response to treatment.