In 199 breast cancers, HER-2/neu amplification was analyzed by Southern blotting (149 cases) and by slot blotting (149 cases), with 99 cases studied using both techniques. There were 18.8% amplified tumors (> or = 2 copies) by Southern blotting and 15.4% by slot blotting. The difference in the percentages of amplified tumors was not statistically significant (p = 0.44). There was a correlation between HER-2/neu amplification and the SBR grade (p = 0.046): this correlation relied on the absence of amplification in the GI tumors. In a subset study, the negativity of the progesterone receptor content was correlated with HER-2/neu amplification in the node positive (p = 0.02), the pre-menopausal (p = 0.04) and the pre-menopausal node positive (p = 0.002) patients. In the literature as in our results, amplification appears to be correlated with poor prognostic factors. However, in a subgroup with most of the favorable prognostic factors: positive estrogen and progesterone receptor contents and node negativity, the frequency of amplified tumors (19.4%, n = 67) was the same as that observed in the whole group (16.6%, n = 199). This result may suggest that the HER-2/neu amplification could act as an independent prognostic factor.