Comparative quantitative assessment of the human corneal sub-basal nerve plexus by histologic staining and in vivo confocal microscopy

Abstract

Purpose: The corneal sub-basal nerve plexus (SBP) offers a potential biomarker for early diagnosis of various neurodegenerative diseases such as diabetic peripheral neuropathy. Changes in SBP can be assessed in vivo by non-invasive confocal laser scanning microscopy (CLSM) observations and quantified using specific morphometric features. However, the ability of CLSM to adequately resolve intimate details of SBP anatomy remains incompletely understood. The present study is designed to compare data from histological staining of the SBP with data from in vivo CLSM to validate the differences in parameters related to SBP quantification in healthy humans.

Methods: The present study was conducted on 3 large scale SBP images (area=1.68 mm²) of healthy subjects generated using a CLSM mosaicking technique described earlier and 3 large scale histologically stained regions of human corneal whole mounts (area=1.65 mm²) prepared by beta-III tubulin staining method. A set of parameters for each image section including corneal nerve fiber length (CNFL), corneal nerve fiber density (CNFD), average corneal nerve single fiber length (CNSFL), average weighted corneal nerve fiber tortuosity (CNFTo), corneal nerve branch density (CNBD), corneal nerve connection points (CNCP) and average weighted corneal nerve fiber thickness (CNFTh) was calculated using a dedicated, published algorithm and compared.

Results: Experiments showed considerable increases in CNFD (259.6 vs. 1345.9 nerve fibers/mm²), CNFL (20.33 vs. 44.03 mm/mm²), CNBD (140.6 vs. 768.3 branches/mm²) and CNCP (33.43 vs. 56.86 connections/mm²) values derived from histological staining compared with values derived from CLSM. In contrast, CNFTh (2.22 vs. 1.35 µm) and CNFSL values were higher (74.76 vs. 33.16 µm) in the CLSM data. No significant difference was observed in CNFTo (0.085 vs. 0.091 µm^-1).

Conclusions: We conclude from these data that the high parameter bias for many quantitative SBP features is due to the higher resolution of the histological staining method compared to in vivo CLSM (1-2 µm). The latter method enables only imaging of SBP nerve fiber bundles but not of single fibers, which are only displayed in histology. In vivo CLSM facilitates SBP in vivo quantification for clinical staging, but the resulting values are biased due to inherent limitations in image resolution.