ADJUVANT CHEMOTHERAPY FOR ER+ BREAST CANCER

Adjuvant breast cancer treatments—chemotherapy, human epidermal growth factor receptor 2 (HER2)–targeted therapies, and endocrine therapy—prevent recurrence and extend survival. Unfortunately, because risk assessment is imprecise and treatments are not uniformly effective, many women are treated to benefit a small number. If these therapies were entirely harmless, we would have few qualms about overtreatment. Chemotherapy has the most onerous short-term side effects and is the treatment that patients most wish to avoid. Moreover, long-term toxicities include secondary leukemia, heart failure, neuropathy, premature menopause, and infertility. Some women who receive adjuvant chemotherapy do not return to work or face difficulty in role functioning.^[1,2] The recognition that benefits are limited and that toxicity can be formidable has led to gradual adjustments in most oncologists' approach. There has also been a steady improvement in prognosis over the past decades, partially attributed to better outcomes associated with screen-detected cancers.^[3,4]

Over the past 15 years, multiple genomic assays have been developed that more precisely characterize the risk of developing recurrent breast cancer.^[5] The 70-gene assay (Mammaprint) was evaluated in a complex randomized trial, which demonstrated that women with clinically high-risk but genomically low-risk cancers derive little benefit from adjuvant chemotherapy.^[6] In the United States, the most widely used genomic assay for estrogen receptor–positive (ER+) breast cancer is the 21-gene recurrence score (RS). In 2004, Paik et al. demonstrated that the RS was prognostic in a group of node-negative patients.^[7] Subsequently, these investigators found that the RS was predictive of chemotherapy benefit and that patients with low and intermediate scores do not appear to achieve substantial risk reduction.^[7,8] The results of randomized trials are pending, but US oncologists have embraced the RS in decision-making. Its use, initially in node-negative patients, has expanded to include many women with node-positive disease, particularly those with limited lymph node involvement.^[9]

Kurian and colleagues demonstrate that use of adjuvant chemotherapy for patients with stage I–II ER+/HER2- disease declined from 26.6% to 14.1% and 81.1% to 64.2% among node-negative and node-positive patients, respectively, between 2013 and 2015.^[10] The frequency of RS use remained unchanged among node-negative patients. RS use increased among node-positive patients from 26.1% to 42.7%, but the authors determined that the test accounted for only one-third of the decline in chemotherapy administration. The increased use of RS among node-negative patients extended for years beyond the initial publications.^[11] Similarly, uptake in node-positive patients appears to be evolving gradually since the 2010 publication.^[9]Data from the prospective randomized trials with RS will likely lead to increased and more informed use in the future, particularly if the trials substantiate the retrospective studies.^[12]

There is little doubt that genomic assays that predict risk and possible chemotherapy benefit have played a major role in reduction of chemotherapy utilization. But genomic assays are only part of the story. With the advent of targeted therapy for breast cancer, both oncologists and their patients appear less willing than in the past to take a toxic treatment for a very small benefit. In the 1990s, the assumption that guided oncologists was that a benefit as small as 1% would be sufficient for patients to accept adjuvant chemotherapy.^[13–15] While there is a high degree of variability in patient preferences, our strong sense is that such small benefits are no longer acceptable. Oncologists have also found that they can identify patients at very low risk of recurrence based on clinical characteristics (tumor size, nodal status, hormone receptors, grade). Some oncologists may even skip ordering a genomic assay if they believe the score is likely to be favorable, particularly if other patient characteristics suggest that risks of treatment are likely to outweigh benefit.

The decline in chemotherapy use from 2013 to 2015 also coincided with changes in Medicare Part B reimbursement mandated by budget sequestration, which went into effect in March of 2013. This legislation cut payments by 2%, yielding a reduction in the margins that oncology practices generate from the "buy and bill" model of chemotherapy delivery. Whether this change in reimbursement directly altered provider behavior is not clear, but there has been a shift in chemotherapy administration from smaller practices to hospitals, where physicians may derive less financial incentive to prescribe chemotherapy.^[16]

Although there were no major changes in adjuvant chemotherapy guidelines during the study period, a National Comprehensive Care Network database study demonstrated excellent outcomes for patients with small ER+/HER2- tumors not treated with chemotherapy with distant recurrence-free survival rates of 98% (T1a) and 96% (T1b).^[17] These data suggest that most patients with small ER+ tumors should not receive chemotherapy and may have also led to a decline in chemotherapy use.

Genomic expression assays may ultimately have an even greater impact on the management of young women with ER+ breast cancer, a population that faces the long-term consequences of treatment toxicities. Due to the small proportion of young women in NSABP B14 and B20 and the increased risk of recurrence in very young women,^[18] many oncologists have been uneasy trusting a favorable genomic assay in a young patient with node-negative disease. Predictive evidence of the RS in node-positive patients remains limited to postmenopausal women.^[9] Nonetheless, it is likely that the RS assay and other genomics predictors reflect disease biology that is independent of age. It is known that young women with ER+ disease do indeed have outcomes inferior to their older counterparts,^[19] but there is reason to believe that this disparity may relate to the adequacy of endocrine therapy in young women. Recent data from the SOFT trial clearly indicate that many young women can be effectively managed with endocrine therapy alone.^[20–22]

In spite of our increasing ability to identify patients with more indolent tumors and excellent outcomes, de-escalation of treatment remains a challenge. While the Kurian study^[10] demonstrated that the proportion of oncologists who recommended chemotherapy in the "less favorable" scenario dropped from 91.3% to 56.4% after the low RS was revealed, limiting chemotherapy use in real time is far more vexing than when one sits in front of the computer monitor. It is both understandable and easier for most oncologists to err on the side of overtreatment than undertreatment. As clinicians, we fear omitting a therapy that could be life saving, but benefits and toxicities must be balanced. Clinical trials such as TAILORx and RxPONDER will likely identify patient populations that can safely forego chemotherapy. For those who do still receive chemotherapy, recent data support limiting anthracyclines to women with higher-risk features, thus reducing long-term complications.^[23]

De-escalation of therapy is not a concept that is confined to ER-positive and HER2-negative disease. In the setting of HER2-positive and triple-negative disease, several agents—capecitabine, platinum, pertuzumab, and neratinib—have been used to escalate the intensity of treatment.^[24–27] Although these approaches represent advances, the benefits for an individual are often marginal. Whether we are focusing on traditional chemotherapy or targeted approaches, we must strive to integrate prognostic biomarkers such as pathologic complete response and develop new prognostic and predictive biomarkers. Our goal is to provide the adjuvant treatment that allows each patient to remain cancer free and, at the same time, avoid unnecessary toxicity. Kurian and colleagues^[10] demonstrate substantial progress in the de-escalation of treatment for patients with stage I–II ER+ disease over a span of only two years. We hope and anticipate that such trends will continue and expand to other subgroups of patients.