We determined the mitotic and apoptotic index through the spectrum of pre-invasive ductal breast lesions to invasive carcinoma in search of disturbances in the proliferation/cell death balance in breast carcinogenesis. Seventy-two pure pre-invasive ductal breast lesions (without invasive carcinoma) and 103 invasive breast carcinomas were used. The numbers of mitotic and apoptotic cells were microscopically counted in hematoxylin and eosin stained sections (MI and AI, respectively), and the ratio of the values of MI and AI was calculated for each individual case (M/A index). A distinction was made between well differentiated and poorly differentiated breast lesions, based on histological type and nuclear grade, to arrive at two plausible progression models for breast carcinogenesis. For the well differentiated breast lesions, the MI was rather equal for hyperplasias and well differentiated DCIS, but increased 6-fold from DCIS to well differentiated invasive carcinoma. The AI remained in the same range, resulting in a 4-fold increase of the M/A index. For the poorly differentiated breast lesions, a significant increase in MI and AI was found from hyperplasia to poorly differentiated DCIS. From DCIS to poorly differentiated invasive carcinoma, the MI increased significantly and the AI decreased 2-fold (n.s.), resulting in a 2.5-fold significant increase of the M/A index. In conclusion, the net increase of the number of cells in the transition from well differentiated pre-invasive to well differentiated invasive carcinoma is accompanied by an increase of cell proliferation rather than decrease in apoptosis, suggesting that in these lesions, proliferation related mechanisms are most important in carcinogenesis and progression. In contrast, in poorly differentiated breast lesions, decreased apoptosis seems to be also important in carcinogenesis and progression. At present, we are gathering patients with invasive breast cancer who had a previous biopsy with a pre-invasive lesion to obtain further more direct evidence for this hypothesis.