A phase II study of repetitive cycles of dose-intense carboplatin plus paclitaxel chemotherapy and peripheral blood stem cells in metastatic breast cancer.

To assess the feasibility of administering sequential cycles of dose-intensive therapy, 14 patients without prior chemotherapy for metastatic breast cancer were registered to be treated with paclitaxel (Taxol; Bristol-Myers Squibb Company, Princeton, NJ) at an initial dose of 250 mg/m2 over 24 hours (day 1), followed by carboplatin dosed to an area under the concentration-time curve of 16 (calculated according to the Calvert formula), every 3 weeks for four cycles. This combination was supported with peripheral blood stem cells collected following granulocyte colony-stimulating factor with or without cyclophosphamide and paclitaxel. One patient failed to peripheralize CD34 cells after cyclophosphamide/paclitaxel therapy and was taken off protocol. The remaining 13 patients entered the paclitaxel/carboplatin phase of the program, and nine completed all four cycles. The median duration of severe neutropenia (absolute neutrophil count < 100/microL) was 6 days, despite the absence of routine use of granulocyte colony-stimulating factor. Only five of a total of 42 chemotherapy cycles (12%) were associated with febrile neutropenia requiring hospitalization. Most patients did not require platelet transfusions. The most significant nonhematologic toxicity was gastrointestinal (grade 3 in three patients, two of whom had received local radiation for relapse before chemotherapy). Most patients developed grade 1 or 2 sensory neuropathy by the final cycle. Of the nine patients who entered the paclitaxel/carboplatin phase and were evaluable for response, five achieved a complete remission. This doublet of high-dose therapy can be given in an entirely ambulatory setting and is associated with modest hematologic toxicity. The value of this option in the treatment of metastatic breast cancer compared with more conventional approaches to high-dose therapy will require a greater number of patients evaluable for response and longer follow-up.