Reproducibility of nuclear morphometric measurements in benign human epithelial cells. Simulation of variations in...

The value of nuclear morphometric measurements in diagnostic pathology is determined largely by the reproducibility of the measurements. Although a variety of factors have been shown to affect tissues during processing, the regulation of fixative type and the avoidance of air drying in particular have been shown to avoid significant variations in nuclear measurements. The current study simulated routine tissue processing in order to identify any factors that may introduce variability of nuclear morphometric values in day-to-day processing if air drying is avoided and fixative type and pH are regulated. Samples of benign endometrium were collected from three uteri, fixed in phosphate-buffered formalin (PBF) from 2 hours to 15 days and dehydrated in an automated tissue processor on four occasions. In addition, tissue from one case was cut at 4, 6 and 8 microns, simulating the potential variations in section thickness that may occur during routine processing. Mean nuclear areas and shape factors of epithelial cells were then determined using computed planimetry. By analysis of variance, no significant differences were found in nuclear morphometric values in relation to time of fixation, dehydration runs or tissue section thickness; coefficients of variation for all variables were less than 7%. This study suggested that routinely processed tissues are adequate for morphometric analysis, including retrospective analysis, provided that tissues are fixed in a pH-regulated fixative such as PBF and air drying is avoided.