The prognostic value of classical prognosticators in 92 invasive breast cancers with long follow-up has been compared with a number of DNA-related image cytometric features (2c deviation index (2cDI)), Auer (histogram) type, 5c exceeding rate (5cER), DNA malignancy grade (MG), and the morphometrically assessed mitotic activity index (MAI) and mean nuclear area (MNA). The quantitative analyses were carried out without knowledge of the outcome nor of clinicopathological features. The cytometric analyses and assessments of MNA were performed on Feulgen stained, 4-microns thick tissue sections in Dresden. The MAI was assessed in Amsterdam. Thereafter, the survival data were matched with the classical variables, the cytometric features, the MAI and the MNA in univariate and multivariate analyses. Lymph node status (LN) and tumour stage were significantly different, tumour stage being the strongest discriminator between survivors and non-survivors (Mantel-Cox value = MC = 30.7, P < 0.0001). Grade was (just) significant (P = 0.05). The cytometric features, the MAI and the MNA were all highly significant prognosticators. The strongest cytometric feature was the Auer type followed by the 5cER, and the 2cDI. The MNA was also significant. The MAI was the strongest single prognostic factor in this blind analysis (MC = 50.8, P < 0.0001), and the MAI combined with the lymph node status, tumour stage and the cytometric features Auer type and 2cDI was the best multivariate combination (MC = 94.0). Separate analyses of the lymph node negative (LN-, n = 39) and positive (LN+, n = 53) groups showed that the 5cER and MAI were the most important in the LN- and the MAI in the LN+ patients.