Three hundred sixty-four cases of invasive ductal breast cancer diagnosed during the years 1988 to 1991 were analyzed to determine quantitative thresholds for mitotic activity. Mitotic counts were calculated in each sample and expressed as standardized mitotic index (SMI) and mitotic activity index (MAI). Based on Kaplan-Meier curves, univariate and multivariate analysis of Cox's regression, and maximum efficiencies of ROC analysis, optimal thresholds were determined on the basis of survival and recurrence of disease. In our material, with a follow-up time of 5 years 9 months, we found two thresholds--a lower and a higher--for both SMI (17 mitoses/mm2 and 32 mitoses/mm2) and MAI (13 mitoses/10 HPF and 35 mitoses/10 HPF). The thresholds were the same in the whole material and in subgroups divided according to the patients' age and axillary lymph node status at the time of diagnosis, and tumor size. The thresholds clearly separated patients with favorable, intermediate, and unfavourable outcome of disease. In our material, the risk of breast cancer death associated with the determined thresholds (ranging from 4.7 to 3.8) clearly exceeded those of menopausal status, axillary lymph node status and tumor size. The risk of breast cancer death associated with the determined thresholds was still emphasized in the groups of premenopausal and axillary lymph node-negative patients, and with tumor size less than 2 cm in diameter (risk ratios, 11.8, 6.0, and 6.7, respectively). The results suggest that the presented quantitative thresholds could be applied in grading of invasive ductal breast cancer.