PubMed1985
A mathematical model for the stereological analysis of parallel-oriented, spheroidal nuclei has been applied to the basal layers of respiratory, metaplastic and dysplastic nasal epithelia. Nuclear profiles seen on tissue sections were measured with an automatic image analyser (IBAS). Nuclear profile area distributions were used to assess possible polyploidies. The result are compared to the histopathological grading, consisting of pseudostratified (respiratory), cuboidal, mixed cuboidal/squamous, squamous, and dysplastic epithelia. The estimated nuclear axis lengths, volumes, surface areas and volume/surface ratios are, in the great majority of cases, significantly smaller in pseudostratified or cuboidal epithelium than in squamous metaplasia and dysplasia. Correspondingly, the numerical density of the nuclei is lower in the latter cases, in which is also noted a smaller nuclear eccentricity. No significant difference is found between the nuclei of squamous and dysplastic epithelia. There is a clear correlation between the mean profile areas and the stereologically estimated volumes of the nuclei. When classifying the mean profile areas into four classes, a progressive shift from the smaller towards the larger size classes is observed when passing from pseudostratified to dysplastic epithelia, through the different metaplastic stages. The nuclear profile area plots often show several peaks, even in some pseudostratified and cuboidal epithelia, probably reflecting polyploidy. There is a marked tendency towards larger profile areas in squamous metaplasia and dysplasia. One histopathologically typical dysplasia showed only small-sized nuclei and another had a notable contribution of those in its profile area plot, in contrast with all the other dysplasias and squamous metaplasias, which displayed practically no small nuclei. The possible implications of the existence of "small-sized-nuclei-dysplasias" are discussed. The literature dealing with nuclear volumes or DNA quantities in putative preneoplastic situations in reviewed. Our method adds some valuable objective criteria to those used in classical histopathological grading. It should be possible to apply this method to other epithelial tissues.