Background. Several molecular genetic alterations in breast cancer, including aneuploidy, altered apoptosis, aberrant expression of p53, HER-2/neu and Bcl-2, have been associated with poor prognosis in breast cancer patients. To determine the importance of molecular-genetic factors relative to more traditional surgical- pathologic prognostic factors, multivariate analysis was performed. Materials and Methods. Ninety-four fresh tissue samples of primary breast carcinoma were studied with flow cytometry for DNA ploidy. On the same specimens steroid hormone receptors (ER and PR) were measured in the cytosol fraction using Abbott ELISA assays, HER-2/neu was determined in the membrane fraction and mutant p53 protein in the nuclear fraction, both, by Oncogene Science ELIZA procedures. Bcl-2 and apoptosis (Cell death) were measured in cell lysates by Oncogene Science & Boehringer Mannheim ELISA assays. In addition, information regarding surgical-pathological features of the tumor was obtained. Multivariate analysis using unconditional logistic regression model was done to identify variables predictive of poor prognosis. Results. Using univariate analysis, histological grade, tumor stage, lymph node status, HER-2/neu, and mutant p53, were predictive of poor short-term prognosis. By multivariate analysis, tumor stage, lymph node status, HER-2/neu were independent factors. Grade subgroup analysis versus time of relapse, illustrated a predictive value of Bcl-2 in only low-grade tumors while apoptosis was significant in high-grade type. Conclusion. Among a panel of moleculargenetic factors investigated, HER-2/neu was the most strongly predictive of poor shortterm prognosis in breast cancer. Patients with HER-2/neu positive tumors can benefit from Herceptin therapy.