Aclarubicin is an anthracycline antibiotic that differs from doxorubicin in its structure, mechanism of action, and preclinical toxicity profile, especially its reduced cardiotoxicity. We therefore conducted a side-by-side in vivo and in vitro trial of this agent in metastatic breast-cancer patients and their biopsied tumor specimens, respectively. Aclarubicin (100 mg/m2) was given by intravenous infusion every 3 weeks to 22 patients with objectively measurable metastatic breast cancer, 15 of whom had not previously received doxorubicin. The dose-limiting toxicity consisted primarily of leukopenia and severe nausea and vomiting. No objective response was observed in the 19 evaluable patients. After disease progression, 10 of the 15 doxorubicin-naive patients were treated with doxorubicin; 6 patients achieved a partial response, including 4 who responded to doxorubicin alone and 2 who responded to doxorubicin in combination with thiotepa and vinblastine. Tumor specimens were obtained from 14 of the 22 patients prior to the start of therapy and were tested for in vitro sensitivity to aclarubicin and doxorubicin using a soft agar colony-forming assay. Adequate colony growth occurred in 9 of 14 cultured tumor specimens. All 9 specimens, including 3 obtained from doxorubicin-naive patients, demonstrated in vitro resistance to aclarubicin. In all, 1 of 3 specimens taken from doxorubicin-naive patients demonstrated in vitro sensitivity to doxorubicin, whereas 6 tumor specimens obtained from patients who had undergone prior doxorubicin therapy demonstrated in vitro resistance. The patient whose tumor demonstrated in vitro doxorubicin sensitivity responded to a doxorubicin regimen after failing aclarubicin treatment; in vitro doxorubicin resistance correlated with clinical resistance in all cases. We conclude that aclarubicin is inactive in metastatic breast cancer at the dose and schedule used. Side-by-side in vivo and in vitro trials are feasible and could be useful in the development of investigational agents with activity greater than that of aclarubicin and, particularly, in the evaluation of analogs of clinically active drugs.