Neoadjuvant chemotherapy for the treatment of breast cancer has undergone significant evolution over time. Historically, neoadjuvant chemotherapy was used only for tumors considered inoperable at presentation. However, this changed after the publication of the well-known National Surgical Adjuvant Breast and Bowel Project study results, which demonstrated that neoadjuvant chemotherapy, in comparison to adjuvant, was not associated with significant differences in disease-free or overall survival and was associated with higher rates of breast-conserving surgery (1,2). Neoadjuvant chemotherapy use was subsequently broadened to include large, operable tumors with the aim of achieving breast conservation.

In addition to breast conservation, neoadjuvant chemotherapy allows for an in vivo assessment of response to therapy, while also providing early treatment of the primary tumor and potential micrometastatic disease (3). Pathological complete response (pCR), often used as a surrogate endpoint to assess the efficacy of neoadjuvant chemotherapy, is also considered to be a strong prognostic measure of long-term clinical outcomes, including disease-free and overall survival (4–6). Hence, identifying and validating factors which predict pCR or improve pCR rates are crucial in breast cancer management.

Furthermore, anthracyclines have been considered traditionally as the most standard and active among breast cancer chemotherapy drugs. In an effort to improve pCR rates, other studies investigated the impact of combining different chemotherapeutic agents in the neoadjuvant setting (7–10). In our study, we not only sought to identify pCR rates and predictors at a single institution, but also to examine the administered chemotherapeutic regimens, and compare and contrast the regimens between patients who did and those who did not achieve pCR.

The results of our unadjusted analyses are in line with previously reported associations between pCR and higher tumor grade and the presence or absence of particular biomarkers. In our multivariate analysis, HER2 amplification, PR status and lack of lymph node involvement were found to be significant predictors of pCR.

Overexpression of HER2, a receptor-like tyrosine kinase, is shared by multiple human carcinomas; however, HER2 amplification in breast cancer is of particular significance in determining therapy and predicting outcome. Present in 20–30% of all breast cancers, HER2 amplification potentiates growth dysregulation, oncogenesis and metastasis, all of which contribute to its association with lower disease-free and overall survival. Tumors that overexpress HER2 are also more likely to be chemoresistant, hence the importance of trastuzumab, a HER2-targeting drug, which was first approved by the Food and Drug Administration in 1998 in the management of these cancers (17). In our study, HER2 overexpression was predictive of pCR in the multivariate model containing individual biomarkers, thus allowing for ER, PR, and HER2 statuses to be directly controlled for in relation to one another. In the model with composite biomarkers, we also observed that the ER+/PR+/HER2- subtype was negatively associated with pCR. The aggregate of these findings suggests that there may be a complex interplay between the tumor biology and treatment with trastuzumab. In reference to pCR rates, we are currently unable to disentangle the potential biological effects of HER2 amplification from the potential benefits gained from trastuzumab treatment. Additional prospective studies are required to elucidate this matter.

Lack of lymph node involvement was also found to be a significant predictor of pCR in patients with HER2 amplification without ER or PR expression, even after adjusting for stage in the regression model. Lymph node involvement at initial presentation is a well-established outcome prognosticator in breast cancer and patients with breast pCR but residual nodal disease have lower rates of overall survival compared with patients who experience breast as well as nodal pCR (1,18).

The assessment of the type of neoadjuvant systemic therapy received, highlights the fact that the majority of the treated patients at our institution received the standard recommended regimen, while also exhibiting improved pCR rates when taxanes are used or added to anthracycline therapy. Multiple trials evaluating a broad range of chemotherapy drugs have also demonstrated that the use of anthracyclines in combination with taxanes is associated with improved pCR rates (7,10). Hence, the consensus statement of the St. Gallen Conference, states that a standard neoadjuvant chemotherapy regimen should include both an anthracycline and a taxane (19).

Platinum agents are DNA-damaging agents, which have been found to be particularly beneficial in improving pCR rates in triple-negative breast cancer (TNBC), particularly BRCA-mutated tumors (which represent up to 50% of all TNBCs). In a meta-analysis of 28 studies, the pooled pCR rate noted following addition of a platinum agent to a standard neoadjuvant chemotherapy regimen for TNBC was 45% (20). Another single-institution study reported pCR rates of 22% following neoadjuvant treatment with single-agent cisplatin therapy in patients with stage II or III TNBC (21). In our study, all 4 patients receiving platinum agents had TNBC and they received single-agent therapy with the platinum agent. Of note, the sole patient exhibiting pCR following administration of the platinum agent had a higher tumor grade (grade 3), lower stage (stage II) and no lymph node involvement. To the best of our knowledge, a review of the literature has retrieved no data to suggest that the improvement in pCR rates following addition of platinum agents to standard neoadjuvant chemotherapy in TNBC translates to improvements in overall survival or disease-free survival. More studies are required to evaluate the potential survival benefits of adding platinum-based therapy to standard neoadjuvant chemotherapy for patients with TNBC.

Our study had certain limitations. First, we were unable to establish causation, but could only show association, due to the fact that retrospective studies are prone to patient and treatment selection bias. Additionally, potential confounding variables may not always be recognized or recorded, due to a lack of knowledge regarding how they interact with the outcome of interest. Second, we were unable to control for administration of trastuzumab therapy with regard to the association of HER2 amplification with pCR. As the provision of HER2-targeted therapy has recently become more standardized, future prospective studies will undoubtedly be better equipped to investigate this issue. Third, all the patients in our study were treated at a high-volume, NCI-designated comprehensive cancer center, although the majority of the patients in this country obtain their chemotherapy from non-academic, community medical oncologists. This may translate to potential differences in patient population, treatment received, treatment duration, or even the definition of pCR; thus, the extent to which our patients exhibit pCR and the reasons why they do so, may not reflect the experience of breast cancer patients receiving care outside a comprehensive cancer center. Finally, details regarding the optimal chemotherapy dose, treatment duration, or concurrent vs. sequential administration for improving pCR rates were not obtained in this study. Other studies are currently underway, however, to address optimal administration times in the neoadjuvant setting to improve pCR rates.

In conclusion, we found in the regression model using the composite HR variable that HR+/HER2- tumors were significantly less likely to undergo pCR compared with HR+/HER2+ tumors, and that tumors with any lymph node involvement were significantly less likely to undergo pCR compared with tumors without lymph node involvement. Additionally, we found that the pCR rates were higher among patients receiving trastuzumab or taxane therapy in addition to anthracycline therapy.