We appreciate the correspondence from Dr. Rushing and de Vries and colleagues. We understand Dr. Rushing's point that some physicians believe that there is adequate evidence from phase II trials to justify the use of a new therapy without confirmation from a phase III randomized trial comparing the current standard therapy to the new treatment. However, we believe it is premature to assume that high-dose chemotherapy is superior to conventional-dose chemotherapy for women with high-risk early-stage breast cancer, despite the encouraging preliminary data cited by Peters and colleagues [1]. Few long-term outcome data (survival) or toxicity data for treatment with high-dose chemotherapy in women with high-risk, early-stage breast cancer are available. We believe it is prudent to proceed cautiously because an emerging literature describes the long-term toxicities (such as myelodysplasia and acute leukemias) associated with high-dose chemotherapy and autologous stem cell transplantation [2, 3].

Because the number of newly diagnosed cases of breast cancer is high (184 000 per year), the number of potential candidates for this therapeutic strategy is large. Widespread use of a very expensive therapy before adequate long-term efficacy or toxicity has been defined may be a disservice to patients. We also believe that it is unfair and inaccurate to state that conventional-dose chemotherapy has no chance of curing patients who have high-risk early-stage breast cancer. Even among patients with meta-static disease there is a subset of patients who have prolonged disease-free survival and may be cured after attaining complete remission with conventional-dose chemotherapy [4]. Dr. Rushing cites examples of the unacceptably long time needed to meet goals of accrual to phase III studies after exciting preliminary data suggest a therapeutic advance (as was seen, for example, with lymphoma and testicular cancer). We do not believe that this diminishes the importance of the phase III confirmatory study. Examples of the importance of randomized studies include Grever and colleagues' comparison of interferon- 2b and 2-deoxycoformycin in hairy cell leukemia [5]. Phase II data suggested an advantage for patients treated with 2-deoxycorformycin with respect to disease-free survival; this was also demonstrated in the randomized study. However, the randomized study also showed that overall survival was not significantly different between the two treatment groups. In breast cancer, the failure to rapidly complete high-priority randomized studies is even more troubling because the number of patients who are potential candidates for these studies dwarfs the other disease sites mentioned above. One would hope that the large number of patients with breast cancer would facilitate completion of important phase III trials in a timely fashion.

We applaud the efforts and success of the Dutch experience in reaching consensus among all university hospitals and cancer centers on the design and implementation of a study comparing high-dose chemotherapy to conventional-dose chemotherapy in patients with four or more involved axillary lymph nodes. We remain convinced that randomized clinical trials are the gold standard for defining benefit of one treatment compared with another. This is particularly true when a state of clinical equipoise exists and well-informed, well-intentioned clinicians have opposing approaches for the same problem.