Reproducibility and comparison of quantitative DNA histogram features obtained with a scanning microdensitometer and...

A pilot study compared DNA histogram features of ten invasive ductal breast cancers measured by three different techniques: in cytologic and histologic samples using a scanning densitometer and in suspensions of single nuclei using a flow cytometer. The reproducibility of measurements pertaining to quantitative ploidy factors was better for flow cytometry (FCM) than for static cytometry and was worst of all using histologic samples. The use of external reference cells for the precise determination of the diploid value in measurements of paraffin-embedded single nuclei proved to be pointless. Of all static histogram descriptors, the modal value was the most reproducible; however, this feature could not distinguish between diploid and near-diploid tumors. In contrast, using the descriptors of (1) diploid peak width, (2) coefficient of variation of the integrated optical density values and (3) percentage of cells exceeding 2.5c from histograms derived from static measurements, it was possible to distinguish between tumors that were characterized as diploid and near-diploid by FCM (chi-square P = .01). This is the more important because subjective analysis of the static histograms of these tumors failed to identify a cell population deviating from the diploid value in spite of the fact that the FCM DNA index (DI) varied from 1.12 to 1.23. Thus, quantitative histogram descriptors, as evaluated in this study, may add objective information to the measurement of DNA ploidy when using static DNA cytometry, especially in the case of near-diploid tumors. Duplicate FCM assessments of the DI showed the highest correlation of all features investigated (r = .99).