Using Chemotherapy Response to Personalize Choices Regarding Locoregional Therapy: A New Era in Breast Cancer Treatment?

Abstract

The reduction in breast cancer mortality rates in the United States is one of the hallmark successes of modern oncology. This progress is largely the result of a rich tapestry of clinical research that has been conducted over the last three decades and has consistently, yet largely incrementally, improved screening and therapy. However, therapeutic strategies do not necessarily advance in unison, and as a result, evolution of one treatment may prove problematic to appropriate application of another. Such is the case with the growing popularity of neoadjuvant chemotherapy and its impact on decisions regarding locoregional radiation therapy. For decades, radiation therapy indications have been based on initial pathologic analysis of all tumor and nodal tissue that was resected before initiation of systemic therapy. A robust body of both prospective and retrospective literature has emerged to inform radiation treatment decisions on the basis of initial pathologic analysis of the untreated tumor. Such literature is of vital importance, given the recognition that radiation when indicated improves not only locoregional control but also survival. Conversely, radiation does have certain toxicities and may impact reconstructive options; thus, its avoidance when not indicated is also of paramount importance. Unfortunately, this radiation therapy literature does not readily extend to patients who are treated with neoadjuvant chemotherapy. Because neoadjuvant chemotherapy has the potential to sterilize disease in up to 40% of patients (and even a higher percentage of patients with human epidermal growth factor receptor 2/neu positive disease who are receiving trastuzumab), there is often a discordance between what the pathologic stage would have been if the patient had undergone surgery before chemotherapy versus the actual pathologic stage when surgery is performed after chemotherapy. Accordingly, it is likely that the patterns of locoregional failure and the impact of radiation on locoregional control and survival may differ meaningfully for the same nominal pathologic stage, depending on whether the stage is rendered before or after treatment with chemotherapy. Although neoadjuvant chemotherapy results in a loss of information regarding what the pathologic stage would have been before chemotherapy, neoadjuvant chemotherapy results in a gain of information regarding the in vivo chemosensitivity of a patient’s tumor, which can be quantified using the residual cancer burden. Theoretically, this novel information that reflects the biology of the patient’s tumor could be used to personalize subsequent decisions regarding locoregional radiation and/or the addition of further systemic agents. To date, the literature to define radiation indications after neoadjuvant chemotherapy remains limited. A systematic review identified 24 relevant studies, 23 of which were retrospectively conducted at single institutions. The only prospective, multi-institution data that were identified by this review are represented in the article by Mamounas et al that accompanies this editorial. Particular strengths of the study by Mamounas et al include its large sample size, prospective data collection and quality control, and uniform policy prohibiting postmastectomy radiation therapy (PMRT) in patients undergoing mastectomy and prohibiting regional nodal radiation in patients undergoing breast-conserving surgery (BCS). As a result, outcomes reported from this relatively unselected patient group should be more broadly applicable than retrospective, single-institution studies. From a conceptual perspective, the key finding is that response to neoadjuvant chemotherapy can be used to prognosticate risk of locoregional recurrence, implying the hypothesis that chemotherapy response may also predict which patients do and do not require radiation therapy. Specifically, in both the mastectomyand BCStreated cohorts, patients with a pathologic complete response to neoadjuvant chemotherapy in both the primary tumor and the axillary nodes experienced the lowest risk of locoregional recurrence, whereas patients with residual invasive disease in the breast who were pathologically node-negative had an intermediate risk of locoregional recurrence, and patients with residual positive lymph nodes had the highest risk of locoregional recurrence. An important caveat of this finding is that initial clinical nodal assessment was performed using a physical examination without a corroborating ultrasound or needle biopsy. Accuracy of clinical nodal assessment was moderate, given that 86% of patients with clinically node-positive disease who were randomly assigned to initial surgery on the B-18 study were confirmed to have nodal metastasis. The misclassification of clinical nodal status limits somewhat the accuracy of risk estimates that are based on this variable. Furthermore, the current study cannot reliably discriminate whether there may be a difference in locoregional recurrence risk for patients who convert from node-positive to node-negative status with neoadjuvant chemotherapy versus patients who are truly node-negative at initial diagnosis and remain node-negative after chemotherapy. JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 30 NUMBER 32 NOVEMBER 1