Abstract
Purpose: :
To assess the reliability of morphometric parameters for normal human corneal epithelium using different numbers of cells from confocal microscopy images.
Methods: :
Contact confocal microscopy images were obtained from twenty subjects without any history of eye disease or surgery. The confocal microscope used was the Heidelberg Retina Tomograph II with the Rostock cornea module. For each subject, the clearest image of basal epithelial cells was recorded. Then outlines of 75 cells were traced manually using Image J software. Area, perimeter and circularity values were analyzed in progressive sets of 5 cells; each pair of values was compared to that obtained using all cells. The coefficient of variation of cell areas (COV) was calculated from the standard deviation (SD) of the cell area measures.
Results: :
An average cell area of 109 µm² ± 2 µm² (SD) was obtained of count of 75 cells / image from 20 normal corneas. For normal basal epithelial cells, 40 cells can usually provide average cell area values that are within 1 and 3.5% of the values estimated from all cells. For this group, the average cell perimeter was 40 µm ± 0.5 µm and the average circularity was 0.86 ± 0.01. 40 cells can usually provide average cell perimeter values that are within 1 and 1.7% of the values estimated from all cells. A similar reliability was observed for estimates of circularity.The calculated COV was 12.5%. With 40 cells used, the calculated COV values were closed to the overall estimates (≤ 1.5%), however, the deviation from the overall estimates was 24%. Even with 75 cells analyzed, the deviation of the COV from the overall estimates was 18%.
Conclusions: :
In assessing the corneal basal epithelial layer with confocal microscopy, cell count should be given and this should be ≥ 40/epithelium for an expected variance to be at a level close to that recommended for monitoring possible changes
Keywords: cornea: epithelium • microscopy: confocal/tunneling • imaging/image analysis: clinical