Sub-basal corneal nerve plexus analysis using a new software technology

Purpose: To study sub-basal corneal nerve plexus (SCNP) parameters by in vivo corneal confocal microscopy using a new software technology and examine the effect of demographics and diabetes mellitus (DM) on corneal nerves morphology. Methods: A Confoscan 4 (Nidek Technologies) was used in this crosssectional study to image the SCNP in 84 right eyes at the Miami Veterans Affairs eye clinic. Images were analyzed using a new semiautomated nerve analysis software program (The Corneal Nerve Analysis tool) which evaluated 9 parameters including nerve fibers length (NFL) and nerve fibers length density (NFLD). The main outcome measure was the examination of SCNP morphology by demographics, comorbidities, and HbA1c level. Results: Interoperator and intraoperator reproducibility were good for the 9 parameters studied (Intraclass Correlations [ICCs] 0.73-0.97). Image variability between two images within the same scan was good for all parameters (ICC 0.66-0.80). Older individuals had lower SCNP parameters with NFL and NFLD negatively correlating with age (r=20.471, and 20.461, respectively, P<0.01 for all). Patients with diabetes had lower mean NFLD 10987.6 mm/mm2 (±3,284.6) and NFL 1,289.5 mm/frame (±387.2) compared with patients without diabetes (mean NFLD 15077.1 mμ/mm² [±4,261.3] and NFL 1750.0 mμ/frame [±540.7]) (P<0.05 for all). HbA1c levels in patients with diabetes were inversely correlated with NFL and NFLD (r= 20.568, and 20.569, respectively, P<0.05 for all). Conclusions: The Corneal Nerve Analysis tool is a reproducible diagnostic software technique for the analysis of the SCNP with confocal microscopy. Older age, DM, and higher level of HbA1c were associated with a significant reduction in SCNP parameters.