A Grid-Based Nucleus Counting Method for Estimates of the Density of Superficial Conjunctival Cells from Impression Cytology Samples Taken from Normal Healthy Human Eyes.

Purpose: To develop and validate an easily procured objective estimate of the cell density of the superficial epithelial cells of the normal bulbar conjunctiva using conjunctival impression cytology (CIC).
Methods: From 20 healthy non-contact lens-wearing young adults, CIC was undertaken off the nasal aspect of the exposed bulbar conjunctiva, the filters stained with Giemsa, and images taken at 200× that only included an all-but-contiguous monolayer of epithelial cells. These images were projected at 1000× magnification and an overlay drawn of the location of the stained nuclei, from which two types of analysis were undertaken using either 100 micron circular or square regions of interest (ROI). From the former, the distances between all nuclei were measured and the number of nuclei/circle calculated to establish uniformity of cell nuclei distribution as a necessary prerequisite to using a grid-based counting method. From the latter, the number of nuclei in each square grid was manually counted and the grid-by-grid variability assessed.
Results: The average inter-nucleus distance was 16.1+/3.5 microns, with predictable positioning to successive nuclei in the circular ROI to give an estimated numerical density of 1944 ± 237/sq mm. Counting of nuclei in grid squares yielded numbers between 12 and 42 to yield an estimated cell density of 2390 +/ ̶330 /sq mm, with repeat counts off grid square ROIs being predictably within +/-15%.
Conclusions: These studies indicate that it should be fairly easy to use carefully selected CIC samples to make estimates of the density of the superficial conjunctival cells. Such a method should be useful and offer advantage over the subjective assignment of morphological grades to CIC samples which have been shown to be highly variable.