Insight or Confusion: Survival After Response-Guided Neoadjuvant Chemotherapy in Breast Cancer+ Author Affiliations
In current clinical practice, the use of neoadjuvant chemotherapy for the treatment of breast cancer is well established. Once reserved for the treatment of locally advanced inoperable tumors, the neoadjuvant treatment strategy has been increasingly employed in operable breast cancer with good reason. This strategy offers a number of advantages over adjuvant therapy, with in vivo assessment of tumor response ranking among the most important. The neoadjuvant strategy has shown great potential as a platform for drug development, and this has led the US Food and Drug Administration to recently strongly consider pathologic response to neoadjuvant therapy as an end point to support accelerated drug approval in high-risk early-stage breast cancer.1 Historical results from a series of important neoadjuvant trials have shaped our current thinking. It has been clearly established that long-term outcomes are similar whether treatment is given preoperatively or postoperatively and that a pathologic complete response (pCR) at the time of surgery is associated with a favorable outcome in all patients who achieve it.2–4 Failure to achieve a pCR is clearly associated with worse long-term outcomes in triple-negative and human epidermal growth factor receptor 2 (HER2) –positive breast cancer, though this negative prognostic association is not observed for the majority of hormone receptor–positive breast cancers.5–7 Rates of pCR following neoadjuvant chemotherapy in hormone responsive breast cancer are uniformly low and because these patients arguably receive their most important therapy, endocrine therapy, after chemotherapy, the focus of current neoadjuvant chemotherapy trials has shifted away from this group of patients in large part. Instead, neoadjuvant endocrine therapy strategies have been increasingly investigated for these patients and the field has moved more generally toward strategies to omit chemotherapy in hormone receptor–positive patients unlikely to derive benefit based on results of multiplex gene expression assays. In the article that accompanies this editorial, von Minckwitz et al8 report survival outcomes after response-guided neoadjuvant chemotherapy in the German Breast Group GeparTrio study. In the German Breast Group tradition, this phase III study represents an enormous effort that tests an interesting strategy of response-guided chemotherapy treatment. More than 2,000 patients with stage II/III breast cancer were treated with an initial two cycles of docetaxel 75 mg/m2, doxorubicin 50 mg/m2, and cyclophosphamide 500 mg/m2 on day 1 every 21 days (TAC). Patients achieving clinical response were then randomly assigned to continuing the same therapy for four or six additional cycles and nonresponders were randomly assigned to continue the same therapy or sequence to an alternate chemotherapy combination of navelbine 25 mg/m2 days 1 and 8 and capecitabine 1,000 mg/m2 orally twice per day on days 1 through 14 of a 21-day cycle (NX) for four cycles. Patients were accrued from 2002 to 2005 and the patient characteristics and therapies are reflective of the times. Notably, HER2 status was unknown in approximately 20% and no patient received neoadjuvant or adjuvant trastuzumab. The primary end points of the study were to compare pCR rates in early responders and clinical response in early nonresponders. The previously reported primary results showed a nonstatistically significant difference in pCR among early responders treated with TAC × 6 compared with TAC × 8 (21% v 23.5%; P = .27)9 and similar clinical responses were observed among early nonresponders with TAC × 6 compared with TAC ? NX (50.5% v 51.2%; P = .008 for noninferiority).10 In the article accompanying this editorial, von Minckwitz et al8 report on the predefined secondary end points of disease-free survival (DFS) and overall survival (OS). At a median follow-up of 62 months, DFS was significantly longer in early responders treated with TAC × 8 compared with TAC × 6 (hazard ratio [HR], 0.78; 95% CI, 0.62 to 0.97; P = .026) and in early nonresponders treated with TAC ? NX compared with TAC × 6 (HR, 0.59; 95% CI, 0.49 to 0.82; P = .001). No significant difference in OS was observed in these two groups, though there was a trend toward improved OS in responders receiving TAC × 8 compared with TAC × 6 (HR, 0.76; 95% CI, 0.57 to 1.01; P = .060). As the sample size was calculated to provide adequate power for the primary end point, the study is underpowered for the end points of DFS and OS. In addition, though the groups were reasonably balanced in terms of baseline characteristics, there were fewer patients with grade 3 tumors in the nonresponder group treated with TAC ? NX (25%) compared with TAC × 6 (36%), and this difference was statistically significant. The unexpected finding, namely, that response-guided therapy influences disease-free survival only in hormone receptor–positive and not hormone receptor–negative breast cancer, rests on a secondary exploratory subgroup analysis and therefore needs to be interpreted with caution. Here, rather than comparing between treatment strategies in the randomized responder versus nonresponder groups, the two investigational response-guided arms (TAC × 8 and TAC ? NX) were grouped and compared with the conventional therapy arms (TAC × 6). Though this grouping is somewhat counterintuitive, as the intent of the response-guided therapy in responders and nonresponders is very different as is the underlying biology, the results suggested that among hormone receptor–positive patients, DFS was comparable in patients with or without a pCR, but response-guided chemotherapy significantly prolonged DFS compared with conventional chemotherapy. For patients with hormone receptor–negative disease achieving a pCR was associated with longer DFS, but response-guided therapy did not prolong DFS. So the challenge now is to put this result in the appropriate context. As we look to the future, does this result provide important insight or leave us more confused? Does response-guided chemotherapy lead to improved outcomes only in patients who are, in principle, the least chemosensitive? Thus far, we have assumed that the obvious path toward optimization of neoadjuvant therapy lies in the identification of highly effective subtype-specific targeted therapies that will increase the pCR rate. As such, we have focused on those subtypes most likely to achieve a pCR. Our best example of the potential of this paradigm is HER2-positive breast cancer. Here we have learned, quite clearly, that the addition of HER2-directed therapy to standard chemotherapy results in a marked increase in the pCR rate compared with chemotherapy alone,11–12 and large adjuvant studies have confirmed long-term gains in DFS and OS with such a strategy.13–15 In the Neoadjuvant Herceptin (NOAH) trial, which compared neoadjuvant chemotherapy with neoadjuvant chemotherapy plus trastuzumab, the roughly 20% gain in pCR rate with the addition of trastuzumab did correspond to improved DFS in that group of patients.16 Though this example demonstrates proof-of-concept, the magnitude of pCR improvement needed to affect long-term outcomes remains unclear,17 and the US Food and Drug Administration is currently wrestling with this issue as they consider using pCR as an end point for accelerated drug approval in high-risk early breast cancer. Though this GeparTrio report focuses on response-guided chemotherapy, in HER2-positive breast cancer we have clearly moved beyond chemotherapy. Though this study's exploratory results suggest that response-guided chemotherapy did not seem to affect long-term outcomes in hormone receptor–negative, HER2-positive breast cancer, it seems unlikely that this would have been the case if patients with a lack of response to two cycles of TAC sequenced onto trastuzumab-containing therapy. In contrast, the negative finding in triple-negative breast cancer is more relevant today. Here, the exploratory results suggest that intensifying the same treatment in responders or switching nonresponders to navelbine and capecitabine did not affect the long-term outcome. Though we still lack effective targeted therapies for this group of patients, we now understand that this group of patients is highly heterogeneous, and emerging data suggest that DNA-damaging chemotherapeutic approaches may be particularly effective in some triple-negative breast cancers with underlying DNA repair defects. We can only speculate as to whether the results would have been different if nonresponders sequenced onto a different chemotherapy combination. Nonetheless, this result strongly suggests that the answers we are seeking in triple-negative breast cancer lie beyond standard chemotherapy. The results suggesting benefit of response-guided neoadjuvant chemotherapy in hormone receptor–positive breast cancer come at a time where we find ourselves moving away from rather than toward more chemotherapy for this group of patients. Though hormone receptor–positive patients were selected based on the presence of high-risk features, the inclusion criteria were such that patients with a low recurrence risk, as assessed by the Oncotype DX assay, could have been enrolled onto this study. It is thus difficult to reconcile this result given that more than half of the hormone receptor–positive patients enrolled were classified as luminal A, defined as estrogen receptor and/or progesterone receptor –positive, HER2-negative with a histologic grade of 1 or 2. Given the exploratory nature of this analysis and because collection of data on postneoadjuvant endocrine therapy was not prospectively captured, my interpretation is that this hypothesis needs to be prospectively tested before response-guided neoadjuvant therapy as investigated in this trial is ready for routine clinical practice in hormone-responsive breast cancer. An important question to consider is whether we have the collective resources and fortitude to move forward with a prospective evaluation of this strategy in hormone receptor–positive breast cancer. If this were pursued, long-term results would not be expected for nearly a decade and we need to consider what the landscape of breast cancer will look like at that time. Because there is no early pCR read-out for such a strategy and it is more challenging to control for postneoadjuvant systemic therapies, my impression is that this neoadjuvant model is not as attractive as our current pCR-based model as a platform for drug development and correlative biology. Several large efforts are evaluating a new model of postneoadjuvant therapy in high-risk triple-negative and HER2-positive breast cancer patients with residual disease after standard neoadjuvant therapy (Clinical Trials No. NCT01772472 and NCT01074970; clinical trial information available at ClinicalTrials.gov). This type of approach is response-guided in principle and synergizes with the current neoadjuvant model based on pCR. As we look to the future, my bias is that advancing neoadjuvant models with the greatest potential to accelerate the pace of progress in early-stage high-risk breast cancer should remain our priority. AUTHOR'S DISCLOSURES OF POTENTIAL CONFLICTS OF INTERESTThe author indicated no potential conflicts of interest. REFERENCES
|
|
|
