Algorithms for morphometric measurements on cancer cells in electron microscopy. Pilot tests

Automated procedures were designed for handling quantitative data derived from a morphometric study of neoplastic cells on electron micrographs. Pilot studies were carried out on 3 breast carcinomas. Each tumor cell was coded by the sequence of 30 stereological parameters that described the general organization of subcellular constituents. Wide standard deviations, as large as 90% of the mean, were observed in some parameters. The practical application of correlation analysis between stereological parameters turned out to be of limited value, since only parameters related to the same subcellular structure showed correlation.