The rationale supporting the adjuvant use of trastuzumab is derived from an improved understanding of the molecular diversity of human breast cancer2,3,21,33
and recognition that this disease should not be treated with the one-size-fits-all approach that has characterized most approaches to adjuvant chemotherapy for the past four decades. The significant efficacy shown by trastuzumab in treating first-line metastatic HER2-positive breast cancer21
prompted its evaluation in early-stage disease. This study is one of four large, randomized adjuvant trials undertaken as part of this evaluation. Like the previously reported studies,22–24
our study showed improvements in efficacy with the adjuvant use of trastuzumab. The major safety issue that we observed was increased cardiac dysfunction when trastuzumab was used in combination with anthracycline-based chemotherapy, a finding that is also concordant with the results of previous studies.22,23,25,26
Despite these similarities, there are three distinct differences between our study and the other trials of adjuvant trastuzumab. First, we performed the intention-to-treat analyses of efficacy and safety as originally and prospectively defined in the study protocol; no statistical modifications or study-group exclusions were permitted for combined or early analyses of data. Second, after the initial efficacy results for adjuvant trastuzumab were reported in April 2005, only 23 patients (2.1%) in the control group (i.e., those receiving AC-T) crossed over to receive trastuzumab before reaching prespecified study end points. This permitted unbiased long-term comparisons of both efficacy and safety among the three study groups. Third, and perhaps most important, our study was the only one to include a nonanthracycline chemotherapy regimen.
Almost all current adjuvant therapies incorporate the use of anthracyclines on the basis of the meta-analysis by the Early Breast Cancer Trialists' Group, which showed an improved efficacy of anthracycline-containing chemotherapy regimens over regimens without an anthracycline.34
However, analyses of HER2 status among approximately 5200 patients who participated in eight large, randomized adjuvant trials comparing anthracycline with nonanthracycline chemotherapy regimens showed that only women with HER2-positive breast cancer derived an incremental benefit from anthracycline use.35–39
No incremental benefit was apparent for the 75 to 80% of breast cancers that were HER2-negative. A meta-analysis of these and additional data has been reported.40
Detailed data derived from an analysis of the TOP2A
gene in our study showed that TOP2A
coamplification occurred in 35% of HER2-positive patients and was directly associated with the incremental benefit that anthracyclines provide over nonanthracycline regimens in HER2-positive disease.30
The use of a non-anthracycline (TCH) regimen in our study allowed us to ask whether preclinical synergies observed between platinum-based chemotherapy and trastuzumab, as well as between docetaxel and trastuzumab,15,27
might compensate for the incremental efficacy advantage of anthracyclines in HER2-positive breast cancers. Although our study was not powered to detect equivalence between the two trastuzumab-containing groups, outcome data showed that TCH had similar efficacy (with respect to disease-free and overall survival) despite the omission of anthracyclines and the use of two rather than three cytotoxic drugs. In our study, this phenomenon was shown even in patients who are thought to be at highest risk for recurrence (i.e., those with ≥4 positive lymph nodes or a primary tumor measuring >2 cm at diagnosis). Moreover, although the overall numerical advantage (29 disease-free survival events) in favor of AC-T plus trastuzumab over TCH was not significant, this difference in efficacy came at the cost of a highly significant increase in congestive heart failure (21 cases vs. 4 cases). These differences in rates of congestive heart failure were not taken into account in any of the protocol-specified efficacy comparisons. In addition, our findings show that the use of AC-T plus trastuzumab resulted in a significantly increased risk of sustained subclinical loss in LVEF. Finally, the use of TCH had the pragmatic advantage of allowing patients to complete all adjuvant chemotherapy in 12 rather than 16 weeks, as well as allowing them to complete 12 months of trastuzumab therapy 12 weeks sooner and requiring fewer infusion visits.
The emerging further understanding of long-term and life-altering toxic effects associated with adjuvant anthracyclines may provide the most compelling support for nonanthracycline regimens. The well-known, long-term side effects of anthracyclines include significantly increased risks of congestive heart failure,41,42
myelodysplasia, and acute leukemia.43
However, recent analyses of the National Cancer Institute Surveillance, Epidemiology, and End Results (SEER) database indicate that we may be underestimating the full effect of anthracycline use on long-term cardiac and hematologic health in patients treated for breast cancer with these agents,41–43
possibly because the majority of adjuvant studies have been designed primarily to evaluate differences in efficacy and have used follow-up periods of 7 to 10 years for both efficacy and safety analyses. Attempts to obtain longer-term safety data, including cardiac outcomes, are systematically biased by loss of later follow-up.
In addition, the SEER database is derived from information on women 65 years of age or older. The mean age of women in the HER2-positive adjuvant studies was 51 to 52 years, whereas the mean age of the overall breast-cancer population is approximately 62 years. Given that data from our study and other trials22,23,25,26
show that trastuzumab augments the incidence of anthracycline-associated congestive heart failure and subclinical loss of LVEF, the full effect of any nonsymptomatic damage induced by the combination of trastuzumab with anthracycline-based regimens may not be apparent until much later as these younger women have additional, age-related cardiac insults. Our findings on long-term, subclinical LVEF losses, as well as the published results from the SEER database analyses,41–43
involving some 42,000 women, support this concern.
Furthermore, although myeloid growth factors permit delivery of full-dose, accelerated-schedule adjuvant anthracycline-cyclophosphamide treatments, registry data suggest a doubling of the incidences of acute leukemia and myelodysplasia associated with these regimens, though the absolute risk remains low.43
Of note, all cases of acute leukemia in our study occurred in patients who had previously been exposed to an anthracycline. Some observers have argued that these major toxic effects are rare and are offset by significant efficacy gains obtained with anthracyclines. However, our data do not fully support this argument. We did not find any significant incremental therapeutic benefits of combined trastuzumab-anthracycline treatment as compared with TCH, yet we did see significant increases in both acute and chronic toxic effects with the trastuzumab-anthracycline regimen. Consequently, we believe that TCH offers an effective alternative to the anthracycline-based regimens and their associated risks.
An improved understanding of the molecular basis of malignant disease is allowing the development of rational treatment strategies that are more effective and less toxic than traditional empiric regimens. The identification and characterization of the HER2 alteration in a subset of human breast cancers and the subsequent development of trastuzumab represent the practical realization of this translational ideal. Our findings show that we can further exploit this new translational knowledge to optimize efficacy while simultaneously minimizing acute and chronic toxic effects in the adjuvant treatment of HER2-positive breast cancer.