Optimization of goblet cell density quantification methods.

ABSTRACT

The purpose of this study was to develop a standardized, accurate and efficient method for estimating conjunctival goblet cell density (GCD) via optimizing sample storage conditions and quantification methods. Conjunctival impression cytology (CIC) membranes were collected from both eyes of 32 participants and were randomized to two storage durations (2-3 weeks, 6-7 weeks) and two storage container types (microcentrifuge tube, flat histology cassette). The CIC membranes were stained and subdivided into 25 areas (5 mm × 5 mm) for imaging and the GCs were counted under 200X magnification using three different methods: (1) full CIC membrane GC count of the 25 images with cell-counting software ("full"; reference method), (2) partial membrane GC count of 9 images with cell-counting software ("partial"), and (3) manual counting of the 25 images ("manual"). In all cases, GCD was determined by dividing the GC count by the counting area. The average time required for quantification was recorded to gauge efficiency. Results showed no significant difference in GC count between the two storage durations (p = 0.745) or storage container types (p = 0.552). The median (interquartile range (IQR)) time required to quantify a CIC membrane for the full, partial, and manual methods of GC counting, was 14.8(17.6), 4.6(5.2) and 5.0 (5.0) minutes, respectively. The agreement of GCD values between the full and manual methods (bias 0.4, 95% LOA [-4.6, 5.5]) was stronger than that comparing the full and partial methods (bias 0.5, 95% LOA [-18, 17]). All together, through systematic examination of key procedural variables, an optimized method for GCD quantification within 7 weeks of sample collection was outlined. Adaption of procedures described in this paper to facilitate accurate and efficient GCD quantification may serve as a valuable step in clinical trials investigating DED pathophysiology and/or novel DED treatment strategies.