Abstract
Purpose: :
Automated quantification of morphological features in different epithelial cell layers in human corneas in vivo.
Methods: :
In vivo confocal laser scanning microscopy (CLSM) was performed unilateral in the central cornea on six healthy volunteers using a dedicated HRTII+RCM hardware and software setup. Stacks of 160 images (400×400 µm, 384x384 pixels, 8 bit greyscale, 30 fps) with an inter-slice distance of 0.4 µm were used to sample the full epithelium (EP) thickness and generate volume data sets. An original image processing algorithm based on phase correlation was used to reduce motion distortions in volume scan image sequences (Allgeier 2011). EP thickness normalization was performed using SD-OCT data (Prakasam 2011). Size and shape factors of basal cell (BC) and intermediate cell (IC) layers of central corneal epithelium was quantified using appropriate image processing and analysis algorithms. Evaluated morphometric parameters include mean surface parallel cell area, cell compactness, cell solidity, cell major and minor diameter and cell maximum boundary distance.
Results: :
Mean cell area (A) measurement of basal and intermediate of EP cells of all subjects demonstrated a linear increase (A(d) = 124.7d + 76.7, d - normalized depth starting from Bowman) from 80 ± 10 µm² to 160 ± 13 µm². A similar trend was noted with major and minor cell diameter and maximum boundary distance. Major diameters of BC and IC measured between 13.2 ± 1.0 and 17 ± 1.4 µm while minor diameter of these cells measured between 8.6 ± 0.9 µm and 12.4 ± 1.4 µm. The maximum boundary distance of BC and IC ranged from 7.0 ± 0.8 to 9.1 ± 0.9µm. Compactness of epithelial cells clustered around 1.45 and 1.5 while cell solidity measured between 1 and 1.03 extending with no significant change except a small raise in solidity of IC.
Conclusions: :
: There are several characteristic morphological quantities which can be calculated from CSLM data sets using this methodology of cell quantification without manual intervention. Our study demonstrated promising results and suggests that this fully automated morphological quantification can be successfully applied to assess microstructral changes of the EP in normal and various corneal disorders.
Keywords: cornea: epithelium • microscopy: confocal/tunneling • imaging/image analysis: non-clinical