Objective: To assess whether computer-assisted quantitative morphological parameters can be an effective tool for objectively distinguishing reactive renal tubular cells from low-grade urothelial carcinoma cells (LG-UCs) in voided urine. Methods: Nuclear morphometry was performed by a computer-assisted image analyser system on Papanicolaou-stained cytological specimens. The circumference of reactive renal tubular cells (n = 40) or LG-UC (n = 20) nuclei were manually traced, and the following nuclear morphometric parameters were analysed: (i) area, (ii) perimeter, (iii) roundness factor, (iv) maximum length, and (v) linear factor. For each nuclear measurement, we calculated the maximum, minimum, mean and standard deviation. Results: The mean nuclear area and nuclear perimeter were higher in reactive renal tubular cells compared to the LG-UCs. The mean of roundness and linear factors (reflecting a tendency for the nuclear outline to be regular and oval, respectively) were higher in LG-UCs compared with reactive renal tubular cells. Among nuclear areas, the nuclear perimeter, roundness factors and maximum length did not show any significant differences between reactive renal tubular cells and LG-UCs. On the other hand, the linear factor showed a mean higher value among LG-UCs than reactive renal tubular cells (P = 0.023). Conclusions: Of five quantitative nuclear morphological parameters, only linear factor was statistically significant in differentiating reactive renal tubular cells in renal disease from LG-UCs.
ASJC Scopus subject areas