Résumé : Anthracycline-based regimens are among the most active but also with greater risk of both acute and long-term side effects, namely cardiotoxicity. Predictive markers of response to anthracyclines are therefore essential. Topoisomerase-IIalpha (topo-II) is the target of anthracyclines and preliminary data suggest its promising role as a predictive marker of sensitivity to these drugs. After screening a population of about 350 patients with locally advanced or metastatic breast cancer, two subgroups were selected for the present analysis: a study group (31 patients), composed of 14 complete responders (CR-a) and 17 true non-responders (PD-a) to anthracycline-based CT, and a control group (28 patients), composed of 7 CR (CR-t) and 21 true non-responders (PD-t) to taxane-based CT. True non-responders were defined as progressive disease (PD) within the first three cycles of CT. Archival tumor samples of these patients were collected, biological markers evaluated and their status correlated with response to therapy. HER-2 and topo-II gene status were evaluated by FISH (Vysis multi-color probe-positivity cut-off: >/=2 ratio for HER-2 and >/=1.5 for topo-II), topo-II protein was evaluated by IHC (positivity cut-off >10%). All cases in which HER-2 gene was non-amplified did not show topo-II gene aberrations. No association was found between HER-2 gene amplification and response to anthracyclines (5/14 (36%) CR and 5/17 (29%) PD to anthracycline-based CT were HER-2+). The topo-II gene was amplified in 3/14 (21%) CR but only in 1/17 (6%) PD to anthracyclines. Amplification of the topo-II gene was seen in 1/7 (14%) CR and in 3/21 (14%) PD to a taxane-based CT. Topo-II protein was overexpressed in 6/11 (55%) CR and in 2/17 (12%) PD to anthracyclines, while in the control group, overexpression was seen in 5/7 (71%) CR and 8/20 (40%) PD. In conclusion: i) HER-2 gene amplification did not seem to be correlated with response to anthracyclines. ii) Both topo-II gene amplification and protein overexpression seem to correlate with response to anthracyclines, although other factors, such as p53 and cell proliferation, are most likely to be involved. iii) The role of combined evaluation of several relevant markers and of potential 'molecular signatures' are currently being evaluated in prospective randomized clinical trials.