[Morphological possibilities for objectification of the prognosis for invasive ductal carcinoma of the breast].

Tissue sections were taken from 187 cases of invasive ductal mammary carcinoma (among them 94 cases with lymph node metastases, average follow-up period being 44 months) and, after Feulgen staining, were investigated by automated microscopic image analysis. A comparison was made, with reference to survival periods, between the relevance to prognosis of nuclear image analysis and results obtained from histopathological grading, according to Bloom and Richardson (1957), as well as findings recorded from image analysis of DNA. Histopathological grading was found to be of poor reproducibility (71.1%). While statistically significant differences were found to exist between survival curves of Grades I and III patients, grading proved to be of minor importance to prognostic assessment of the individual case, with the number of G-II cases being relatively high (42.7%). Results of better reproducibility were achievable from the quantitative morphological method of nuclear image analysis. Its forecasting value with regard to good or poor prognosis was clearly higher than that of histopathological grading, and the number of cases with obscure or dubious prognosis was clearly lower (18.9%). DNA ploidy measurement, using DNA histogram types according to Auer (1980), has proved to be another method by which prognostically favourable and unfavourable cases can be distinguished from each other with statistical significance, though its classification quality (70.5%) still was unambiguously below that achievable by nuclear image analysis (79%).