Application of computer image analysis in endocrine cell quantification.

Computer image analysis was applied for quantifying endocrine cells by using an automatic standard sequence analysis operation. Two parameters were used, namely the number of cells per mm3 of epithelial cells and the cell secretory index (the volume of the immunoreactive secretory granules per cell). The first indicates the variation in the anatomical peptide-producing unit and the second the synthesis and secretion activity of the cell. The endocrine cells chosen in this study were chromogranin-immunoreactive cells and secretin-immunoreactive cells in the human duodenum. The measurements were made by five different investigators with different backgrounds in order to evaluate the effect of the intra- and inter-individual variation. This study showed that the intra- and inter-individual variation had no impact on the results. Comparisons with the classical point-counting method considered to be easy and most efficient in volumetry showed that the present approach is between two and three times faster and less strenuous for the performer. It is concluded that this approach seems to be suitable for adaptation in morphometric studies when information is required about the changes in the number of endocrine cells and about changes in secretory activities.