In this study, the expression of cyclin B1 and p34cdc2 in neoplastic and non-neoplastic breast lesions was evaluated by immunohistochemistry and quantitative analysis in relation to cellular kinetic parameters such as Mitotic Index (MI), Anatelophase Index (ATI), and Apoptotic Index (AI). The percentage of cyclin B1 and p34cdc2-positive cells was significantly higher in neoplastic glands than in their normal counterparts. This finding was paralleled by significantly higher values of MI, ATI, and AI in breast cancer than in normal glands. Furthermore, two groups with different cytokinetic characteristics were identified among infiltrating ductal carcinomas by an unsupervised learning technique of cluster analysis using the percentages of cyclin B1 and p34cdc2 positive cells and the cellular kinetic parameters (MI, ATI and AI) as variables. The final clusters, groups I and II, consisted of 42 and 13 cases respectively. The first cluster (group I) was characterized by a significantly linear correlation between the percentages of cyclin B1 and p34cdc2-positive cells. On the contrary, the second cluster (group II) revealed no correlation between these two proteins and was characterized by values of p34cdc2 largely exceeding those of cyclin B1. A positive correlation between the expression of these two proteins and the cellular kinetic parameters (MI, ATI and AI) was also found in group I but not in group II. These observations suggest that a disturbed nuclear translocation of Mitosis Promoting Factor (MPF) components is present in group II cases, resulting in a defective cellular division cycle. In fact, group I cases showed lymph node metastasis more frequently than group II cases. Our results suggest that the analysis of the cell cycle "machinery" components, such as the cyclins and their dependent kinases, can identify tumors with different levels of aggressiveness.