To the Editor:

Data from multiple clinical trials have indicated that pathologic complete response (pCR) after primary chemotherapy in patients with breast cancer (BC) (ie, no residual cancer in the breast or lymph nodes) is associated with an excellent long-term prognosis.¹ These results have suggested that pCR can be used as a surrogate marker of efficacy. Many primary chemotherapy trials have been designed with this as the primary end point. It should be noted however, that while this data has demonstrated pCR as a strong prognostic parameter, this does not necessarily imply that it is also a valid surrogate end point of efficacy. pCR, in fact, may have selected a subgroup of patients already destined to a better outcome even without treatment. Prentice provided a definition and a set of criteria to be used to identify a surrogate end point of treatment efficacy.² Accordingly, to prove that pCR is a surrogate parameter of efficacy of primary chemotherapy, it is not sufficient to demonstrate that it occurs with treatment and correlates with overall outcome. We should also demonstrate that any difference in survival disappears once we adjust for pCR. In other words, prognosis, once the surrogate is known, should be independent of treatment, and responders to more and less active treatment should have the same survival.

The validation of pCR after primary chemotherapy as surrogate end point of treatment efficacy would therefore require a meta-analysis comparing two different regimens administered preoperatively, large in size (more than 2,000 patients), in which a large difference in pCR can be obtained (>10%), and this would further improve, if a significant difference in survival has been observed. Of course we do not have such data yet. On these grounds should we accept the results of randomized primary chemotherapy trials having pCR as primary end point? Buzdar et al³ provide interesting information in this respect. Forty-two patients randomly assigned to receive four cycles of fluorouracil, epirubicin, and cyclophosphamide followed by four cycles of paclitaxel with or without trastuzumab, showed a dramatic increase in pCR rate in favor of patients treated with trastuzumab (65.2% v 26.3%). Based on these results, the authors interrupted the randomized trial early and they prosecuted it as a single arm trial. The efficacy of adding trastuzumab to cytotoxic chemotherapy in terms of survival prolongation was demonstrated in adjuvant setting by the combined analysis of two multicenter American studies, the National Surgical Adjuvant Breast and Bowel Project B31 trial and the North Central Cancer Treatment Group N9831 trial, recently presented at the 41st Annual Meeting of the American Society of Clinical Oncology in Orlando, FL, May 13-17, 2005. The trial has randomly assigned 3,351 HER+ patients to receive adriamycin plus cyclophosphamide (AC) followed by paclitaxel compared to the same regimen plus trastuzumab, which was administered concomitantly with paclitaxel and prosecuted for 12 months after the completion of chemotherapy. It is impressive to note that the results of large clinical trials, which have involved thousands of patients followed for 24 months on average, were anticipated by the results obtained in a few months of a small single institution study. These data have several implications: first of all, they strongly support the notion of pCR as a surrogate parameter of treatment efficacy. Secondly, in the near future different breast cancer subsets will be identified and treated along distinctive pathways. It may therefore become more difficult to conduct large-scale randomized clinical trials.

Primary systemic treatment represents the ideal model for studies aiming to develop new treatment strategies in early breast cancer patients in the targeted therapy era. Buzdar et al provided the first application of this model in this respect.³ On the basis of a dramatic improvement in pCR, trastuzumab-based chemotherapy was considered as a standard treatment, leading to the early interruption of the clinical trial, after only approximately 20 patients were randomly assigned onto each arm.

It should be noted, however, that in randomized trials a difference in pCR did not always translate into an outcome improvement.

The National Surgical Adjuvant Breast and Bowel Project B-27 trial randomly assigned 2,411 stage I-III patients to one of three treatment groups: four cycles of primary AC before surgery, four cycles of primary AC followed by four cycles of docetaxel before surgery, and four AC cycles followed by surgery and then four cycles of docetaxel.⁴ The results showed a clear advantage of adding docetaxel to AC in terms of pCR (26.1% v 13.6%, respectively). At a recent update, however, this trial failed to demonstrate a significant improvement in overall survival of docetaxel treated patients despite a mature follow-up and an adequate number of events.⁵

How could we explain these discrepancies? In our opinion the patient selection could have made the difference. Several recent papers have demonstrated that patients with poorly differentiated, highly proliferating, estrogen receptor-negative primary tumors have the greatest chance of pCR after primary chemotherapy. Conversely, patients with well-differentiated estrogen receptor-positive tumors seldom, if ever, achieve a pCR.⁶ Since c-erbB2 positivity in BC strongly correlates with high proliferative activity and estrogen-receptor negativity, an attractive hypothesis is that pCR could be a sensitive and reliable surrogate end point in this subset of patients whereas the heterogeneity of BC could have limited the impact of this end point in the overall population.

On the other hand, it should be noted that in the previously cited phase III trials comparing chemotherapy and trastuzumab versus chemotherapy alone, trastuzumab administration was not interrupted at the end of chemotherapy but was continued for 1 year. The HERA trial, also presented at the recent American Society of Clinical Oncology meeting, confirmed the advantage in overall survival of adding trastuzumab to chemotherapy. However, in this study trastuzumab was administered for 12 months sequentially to adjuvant chemotherapy. On the basis of these data it is impossible to discriminate whether the potentiating effect of adding trastuzumab to chemotherapy may be attributable to an increased pCR or to a maintenance effect: a durable switch off of a metabolic pathway.

In conclusion, primary systemic treatment is increasingly recognized as the best model for the quick development of new treatment strategies in early breast cancer. However, the absence of validated surrogate parameters of treatment efficacy represents a hindrance in the interpretation of the results.

Authors' Disclosures of Potential Conflicts of Interest

The authors indicated no potential conflicts of interest.