Multivariate quantitative histopathological assessment of prognostic factors in ductal infiltrating breast carcinoma

The aim of this study was to evaluate the power of quantitative pathology in improving the accuracy of prognosis in ductal infiltrating breast carcinoma. Ninety tumours were studied, with a diameter < or = 2.5 cm; positive (metastatic) axillary lymph nodes were found in 48 of the 90 cases; no patient had systemic metastases. Surviving patients had a mean follow-up of 106 months (minimum 69 months). At the end of the study, 45 patients were alive, 45 deceased. The histopathologic study of tumour nuclei and nucleoli included: a) geometric features, i.e., the mean and standard deviation of nuclear area, perimeter, diameter and form factor, nucleolar area and nucleolar/nuclear ratio, measured with a Kontron-IBAS automatic image analyser; b) immunohistochemical features, i.e., the percentage of PCNA (Proliferating Cell Nuclear Antigen)-positive nuclei, the expression (positive or negative) of vimentin (as a marker of tumour dedifferentiation), and the nuclear protein p53. Statistically significant differences were found between surviving and decreased patients for most features; multivariate analysis gave 92.2% accuracy in predicting outcome. Among the classes defined with multivariate analysis, it was possible to distinguish subsets of patients with bad prognosis by studying the expression of vimentin and p53, although these markers were positive in a small number of cases: 5 out of 6 patients with vimentin positivity and 4 out of 8 with p53 positivity died within 24 months of the diagnosis.