Time trends of neoadjuvant chemotherapy for early breast cancer

Neoadjuvant chemotherapy (NACT) in early breast cancer (EBC) enables in vivo sensitivity testing and less radical surgery as compared to primary surgery and adjuvant chemotherapy (ACT). The aim of our study is to illustrate trends of systemic treatment of EBC. The study analyzed chemotherapy usage and time trends for patients with EBC treated at 104 German breast units between January 2008 and December 2017. The data were obtained through a quality‐controlled benchmarking process. Altogether, 124 084 patients were included, of whom 46 279 (37.3%) received chemotherapy. For 44 765 of these cases, detailed information on systemic treatment and surgery were available. Overall use of chemotherapy declined from 42.0% in 2008 to 32.0% in 2017. During that same time, the proportion of NACT increased from 20.0% to 57.7%, irrespective of tumor subtype. The pathological complete response (pCR) rate (defined as ypT0 ypN0) at surgery after NACT increased from 15.0% to 34.2%. The results from this large cohort from the clinical routine reflect the refined indications for chemotherapy in EBC.

Breast Cancer Trialists' Collaborative Group meta-analysis including 10 randomized NACT trials with a median of 15 years showed equivalent distant disease-free and overall survival in comparison to adjuvant chemotherapy. 5 The decision for systemic therapy in general remains made primarily based on a risk stratification by tumor load and differentiating EBC subtypes defined by immunochemistry. 6 As an additional decision aid, multigene signature testing has become more widely available in initial therapy management, but it is not yet standard in international and national guidelines. Multigene signature testing remains restricted to cases in which a clear therapy decision cannot be made based on all other criteria. 7,8 NACT was used initially to enable or improve the operability of breast tumors that were large, inflammatory, or locally advanced. This approach offers several benefits compared to ACT. Although the improvement in the rate of breast-conserving surgery (BCT) due to NACT is not as important now as it was in the prescreening era, NACT can decrease the need for completing axillary lymph node dissection (cALND), thereby reducing surgical morbidity. 9,10 NACT also provides important prognostic information and enables in vivo testing for drug sensitivity that facilitates postneoadjuvant escalation and de-escalation strategies depending on tumor response. 11 The treatment goal is the achievement of a pathological complete response (pCR), usually defined as no invasive tumor residue in the breast and axilla Patient and tumor characteristics of female cases with primary, nonmetastatic breast cancer who were diagnosed and treated at 104 German Breast Cancer Units between January 1, 2008 and December 31, 2017, and underwent chemotherapy (n = 44 765) Center (WBC) in Düsseldorf, Germany. The WBC is an institution that provides quality control through an annual benchmarking report. 15 The data are also used for the German Cancer Society's periodical recertification process for certified BCUs. 16 Collaborating BCUs col-

| Definitions of tumor histology and stages
Tumor histology was defined according to the World Health Organization criteria, 17 and grading was performed according to the most recent TNM classification. 18 The expression of estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2) and Ki-67 was assessed using formalin-fixed paraffinembedded tumor tissue according to international standards. Positivity for ER and PR was defined as an immunoreactive score (IRS) of  Immunohistochemical information was based on the pretreatment biopsy (if available) for patients with NACT and on the final postsurgery pathological sample for patients with ACT ( Figure 1).

| Response to neoadjuvant chemotherapy
The overall pCR rate (according to the most common definition ypT0 ypN0) rose from 15.0% (2008) to 34.2% (2017; Figure 5). The most prominent effect in the increasing pCR rate was seen in tumors of the (25.3%) and 2017 (42.5%) was seen. On a lower level but with almost a tripled rate, an increase in pCR was also seen in HR+ HER2− tumors with an increase from 5.6% to 16.3% ( Figure 6).

| DISCUSSION
Our results reflect developments in the use of chemotherapy in EBC patients. Since the advent of molecular classification systems, 22  high-risk, genomically low-risk (ie, recurrence score ≤ 11) pN0-1 patients without ACT. In the TAILORx trial (also using Oncotype DX) omission of chemotherapy was associated with similar 9-year invasive DFS (83.3% vs 84.3%). The intention of these trials was to identify patients that can be spared chemotherapy without comprising outcome. Thus, it is reassuring that these developments are an important factor that lead to a substantial reduction of overall chemotherapy use in EBC over time in daily practice in Germany as was already shown for the US. 26 In addition, study results from surgical trials might have had an influence on ACT usage. In recent years, it has been possible to reduce surgical radicality, especially in the axilla, because the need for cALND in cases of affected sentinel lymph nodes (LN) has been put into question. The ACOSOG Z0011 trial showed no difference in outcome in patients with pT1-T2 cN0 breast cancer and involvement of up to two sentinel lymph nodes after breast-conserving surgery and whole-breast irradiation who did not receive cALND. 27 This comparable outcome was achieved despite the fact that a high rate of additional nonsentinel metastases (NSM) can be assumed to be left behind when cALND is omitted. This high prevalence of occult tumor burden was also confirmed in a German study. 28 Although tumor biology is undoubtedly the most important factor in clinical decision making for ACT, pathological LN status still represents a strong prognostic factor, so lack of this information could reduce the usefulness of stratification for ACT. 29 Nonetheless, post hoc studies trying to estimate the impact of information gained from cALND in these situations have shown inconclusive results. 30,31 In EBC, chemotherapy is currently used more often neoadjuvantly Fourth, the dataset lacks information about oncological outcomes, so no analyses could be performed about how the changes in chemotherapy usage over time affected outcomes.

| CONCLUSIONS
The results from this large, nationwide sample from routine clinical practice reflect the refined, more individualized indications for chemotherapy in EBC and its preferred application as NACT.