Adjuvant chemotherapy is a treatment for supposed micrometastatic disease. It aims at reducing the subsequent risk of relapse and death after primary breast cancer. In the past 35 years, several chemotherapy regimens have been used in the adjuvant setting. The three chemotherapy regimens that have proved to improve most breast cancer outcome are (1) cyclophosphamide, methotrexate and 5-fluorouracil (CMF) regimen, (2) the anthracycline-based regimens and (3) the taxane-based regimens [1–3]. According to the results of the meta-analysis conducted by the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG), adjuvant chemotherapy decreased the annual relative risk of relapse and mortality by 23% and 17%, respectively. [1]. The assessment of biological features of the disease is crucial to select the best treatment option for patients with early breast cancer. Conventionally, the selection of the most active adjuvant regimen for patients with breast cancer was based on stage rather than on biology. Nevertheless, breast cancer can no longer be considered as a single disease, as several breast cancer subtypes have been defined by genetic array tools [4, 5]. Traditional clinical-pathological features are routinely used for approximations to this classification [6]. Immunohistochemical markers such as estrogen receptor (ER), progesterone receptor (PgR) and human epidermal growth factor receptor 2 (HER2) play a key role in defining the clinicopathological surrogate definitions of subtypes, with – accordingly – different risk factors, natural histories and sensitivity to systemic therapies [7–9]. In the present chapter, we will attempt to address the issue of selecting the most appropriate cytotoxic regimen for adjuvant therapy according to the different subgroups of breast cancer, namely, endocrine-responsive (luminal A and luminal B), HER2-positive and triple-negative breast cancer.

Luminal breast cancers, which represent the most frequent subtypes of breast cancer, include tumours expressing ER. They are defined as a group of tumours with a great heterogeneity in histology, natural history, molecular signatures and response to treatments [10, 11]. According to their different gene expression profiles, two main ER-positive breast cancer subtypes are recognized, specifically luminal A and luminal B [4, 12, 13]. As compared to the luminal A subtype, the luminal B subtype is characterized by lower expression of ER levels, lower or no expression of PR [14] and a higher proliferation index [15]. Also, it is thought that luminal B tumours derive greater benefit from chemotherapy than luminal A tumours [5, 16, 17]. One of the most remarkable and challenging questions is whether (some) luminal breast cancer patients could be spared adjuvant chemotherapy. In this context important and innovative trials have tried to address this question by using genomic tools. In women with early-stage breast cancer, the 70-gene signature test (MammaPrint) has been shown to improve prediction of clinical outcome. The recently published phase 3 MINDACT trial was designed to offer prospective evidence of the clinical utility of using the 70-gene signature in addition to standard clinical-pathological criteria to select patients for adjuvant chemotherapy [18]. This trial randomized 6693 women with early-stage breast cancer and evaluated both the genomic risk (using the 70-gene signature) and the clinical risk (using a modified version of Adjuvant! Online) [18]. Women at low clinical and genomic risk did not receive chemotherapy, while those at high clinical and genomic risk did. In patients with discordant risk results, either the genomic risk or the clinical risk was used to decide the use of chemotherapy. Women at high clinical risk and low genomic risk for recurrence who were spared chemotherapy based on the 70-gene signature had a 5-year rate of survival without distant metastasis of 94.7% (95% confidence interval, 92.5–96.2), and that was 1.5 percentage points lower than the rate with chemotherapy. The subset of patients who had estrogen receptor-positive, human epidermal growth factor receptor 2-negative and either node-negative or node-positive disease had similar rates of survival without distant metastasis. These results tell us that about 46% of women with breast cancer who are at high clinical risk might not need chemotherapy [18].

The Trial Assigning Individualized Options for Treatment (Rx), or TAILORx, aims at examining whether genes that are frequently associated with risk of recurrence can be used to assign patients with early-stage breast cancer to the most proper and effective treatment [19]. TAILORx seeks to incorporate a molecular profiling test (the 21-gene assay (Oncotype DX Recurrence Score, Genomic Health)) into clinical decision making, thus sparing women needless treatment when chemotherapy is not likely to be of significant benefit. This study involved more than 10,000 women recently diagnosed with hormone receptor-positive, HER2-negative, axillary node-negative breast cancer with tumours of 1.1 to 5.0 cm in the greatest dimension (or 0.6 to 1.0 cm in the greatest dimension but with tumours at intermediate or high grade) who were considered to be candidates to adjuvant chemotherapy according to clinicopathologic features [19]. Patients were categorized into three recurrence risk groups according to their score calculated by the 21-gene assay. Women in the lowest-risk group were allocated to receive hormone therapy, women in the intermediate-risk group were randomized to receive either hormone therapy alone or hormone therapy plus adjuvant chemotherapy and women in the highest-risk group were allocated to receive hormone therapy plus adjuvant chemotherapy. In September 2015, Sparano and colleagues published results from an analysis of the women in the lowest-risk group. 1626 women (15.9%) had a recurrence score of 0–10 and were assigned to receive endocrine therapy alone. Among patients with hormone receptor-positive, HER2-negative, axillary node-negative breast cancer that had a favourable gene expression profile and who would have been considered suitable to adjuvant chemotherapy according to clinicopathologic features, the rates of recurrence at 5 years with endocrine therapy alone were very low [19]. The 5-year rate of invasive disease-free survival was 93.8% (95% confidence interval [CI], 92.4 to 94.9), the rate of freedom from recurrence of breast cancer at a distant site was 99.3% (95% CI, 98.7 to 99.6), the rate of freedom from recurrence of breast cancer at a distant or local-regional site was 98.7% (95% CI, 97.9 to 99.2) and the rate of overall survival was 98.0% (95% CI, 97.1 to 98.6) [19].

The majority of luminal A tumours have an outstanding prognosis with endocrine therapy alone. In this subset the use of chemotherapy is much discussed, particularly in node-negative disease. The International Breast Cancer Study Group (IBCSG) trial IX for postmenopausal women and the IBCSG trial VIII for premenopausal patients [8, 20–22] compared three or six courses of adjuvant cyclophosphamide, methotrexate and fluorouracil (CMF) with or without endocrine therapy versus endocrine therapy alone. In these studies, chemotherapy showed no benefit in ER-positive/HER2- negative breast cancer patients (hazard ratio [HR] 0.90; 95% CI, 0.74–1.11) in the subset of ER-positive, HER2-negative and low-Ki67 tumours, which corresponds to the proxy definition of luminal A disease [22]. Moreover, the results of the retrospective analysis of the NSABP B-20 trial (which compared tamoxifen only versus MF + T versus CMF + T in women with ER+ cancer and found both overall and disease-free survival advantage in both chemotherapy arms versus tam alone) showed that patients with ER-positive/node-negative breast cancer and with a low RS by Oncotype DX treated with tamoxifen did not benefit from the addition of CMF chemotherapy [23]. Nowadays, there are controversies regarding the therapeutic approach to optimally treat luminal A disease with a large tumour burden (e.g. nodal involvement). Usually nodal involvement is supposed to require the use of adjuvant chemotherapy, despite the fact that the good outcome of luminal A disease with endocrine therapy alone and its relative chemoresistance [8, 23] induce consideration that these tumours should be cured according to biology rather than stage. This clinical topic is actually explored in the RxPONDER trial which studies if patients with ER+/ HER 2 negative breast cancer, RS ≤ 25 and with involvement of one to three axillary nodes may be spared adjuvant chemotherapy [19]. Results from this trial should help addressing whether favourable tumour biology is more significant than unfavourable tumour stage when making adjuvant therapy decisions. Up to now, according to the St. Gallen consensus, in patients with luminal A-like disease, chemotherapy could be added to endocrine therapy on the basis of tumour burden and risk assessment. The St. Gallen panellists did not specify a preferred chemotherapy regimen for these patients and expressed the view that any of the standard regimens, including the first- and second-generation regimens (CMF, Adriamycin plus cyclophosphamide (AC), Taxotere plus cyclophosphamide (TC)), could be considered [10].

Luminal B tumours are characterized by higher proliferation rates and an increased risk of relapse when compared to patients with luminal A tumours. Hence, the addition of chemotherapy to endocrine treatment is indicated for the majority of these patients. The benefit of chemoendocrine therapies compared to endocrine therapies alone was clear in several trials [8, 23]. Particularly, data from a large meta-analysis of patients with ER-positive tumours from the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) reported that proportional risk reductions from chemotherapy were slightly affected by age, nodal status, tumour size or differentiation, estrogen receptor status or tamoxifen use [24]. For patients with luminal B subtype cancers, the majority of the St. Gallen panellists considered the use of chemotherapy. Normally, chemotherapy regimens should comprise anthracyclines and taxanes. The optimal adjuvant chemotherapy duration is not established yet, but a duration of 4–6 months is considered to be reasonable [10].

About 15% of breast cancer presents with HER2 overexpression/amplification. This feature is associated with a poor prognosis [25] and remains the main predictive biomarker for the use of the humanized monoclonal antibody trastuzumab and other anti-HER2 drugs [26, 27]. Since 2005, the adjuvant treatment of this breast cancer subtype has drastically changed with the publications of the findings from the first-generation adjuvant trials combining trastuzumab with chemotherapy, either concomitantly or sequentially [28–31]. Additionally, HER2 amplification is associated with greater sensitivity to chemotherapy, including anthracyclines [32–34] and taxanes [35–37]. Actually the most relevant issue is whether or not to include an anthracycline in the adjuvant treatment of HER2+ breast cancers, particularly in view of the risk of cardiotoxicity which is increased with sequential trastuzumab therapy [38]. An interesting discovery is that HER2 and topoisomerase IIa (TOPO2A) gene coamplification is associated with high sensitivity to anthracycline-based chemotherapy [39]. According to these data, the BCIRG (Breast Cancer International Research Group) [40] retrospectively looked at the predictive value of TOPO2A gene amplification in patients with HER2-overexpressing breast cancer in a randomized trial, which compared anthracycline-taxane-based chemotherapy with taxane only-based chemotherapy. The investigation confirmed a greater benefit for anthracyclines in patients with HER2+/TOPO2A-amplified disease. Nonetheless, the predictive value of TOPO2A gene amplification has not been independently validated and chromosome 17 polysomy may be the more influential predictor [41]. To date, there are insufficient proofs for modifying chemotherapy regimens on the basis of TOPO2A expression, HER2 status or chromosome 17 copy number. So far, for patients with HER2-positive disease, the standard adjuvant treatment is trastuzumab plus chemotherapy, which should include a taxane and an anthracycline according to the St. Gallen guidelines [10].

Triple-negative breast cancers (TNBCs), defined by their lack of immunohistochemical staining for ER and PgR and lack of overexpression or amplification of HER2/neu, are characterized by their aggressive clinical course and poor prognosis [12, 42]. Lacking specific targeted therapy, chemotherapy with standard cytotoxic agents is the only systemic treatment option approved for these patients. There is no robust evidence to advice use, or avoidance, of specific chemotherapy agents in the TNBC subset. Several studies have demonstrated a broad chemosensitivity for these tumours, mainly in the neoadjuvant setting [17, 43–45]. In these trials, TNBCs revealed higher response rates (RR) than other BC subtypes but showed a poor overall survival rate. The TNBC subtype is associated with a paradox: despite a subgroup of patients who are very chemosensitive, the whole subgroup shows poor disease-free and overall survival. This paradox was emphasized by a neoadjuvant analysis in which TNBC patients achieving pCR had an excellent outcome (3-year overall survival 94% vs. 98% for non-TNBC), while TNBC patients not achieving a pCR had a high probability of systemic relapse (63% vs. 76%, respectively) and death (74% vs. 89%, respectively) within 3 years of primary diagnosis [43]. The elucidation of this behaviour could be found in the fact that within the TNBCs exists a subgroup with intrinsic chemoresistance. Thus, the identification of specific biomarkers could be useful to recognize patients at different responsivity to chemotherapy and to develop new therapeutic approaches.

Chemotherapy benefit in TNBC could be related to its high proliferative rate and – to some extent – intersection with BRCA1 mutation-related breast cancers. BRCA1 is involved in homologous recombination, a cellular process of double-strand DNA break repair, and in the cell cycle arrest necessary for DNA damage repair. Loss or inactivation of BRCA1 may induce peculiar susceptibility to DNA damaging agents, such as platinum derivatives, and relative resistance to mitotic spindle poison, like taxanes and vinca alkaloids [46]. Nevertheless, data on the role of platinums in TNBC are still limited and controversial. At the 2013 San Antonio Breast Cancer Symposium, the preliminary results of the CALGB/Alliance 40,603 study were reported. In this trial, patients were randomly assigned in a 2 × 2 schema to receive weekly paclitaxel for 12 courses plus dose-dense anthracycline/cyclophosphamide with or without the addition of biweekly bevacizumab for nine cycles or the addition of carboplatin every 3 weeks for four cycles. The addition of carboplatin to the neoadjuvant regimen significantly augmented the rate of pathologic complete response in patients with triple-negative breast cancer [47]. In contrast, the results from the neoadjuvant trial GEICAM/2006–03 showed that the addition of carboplatin to conventional chemotherapy (epirubicin plus cyclophosphamide followed by docetaxel) in basal-like breast cancer patients did not increase efficacy [48].

Potential biological bases for taxane sensitivity in TNBC are the high tumour proliferative rate and the presence of aberrant p53 in about 50% of TNBC [49]. Data about the role of taxanes in the TNBC population come from retrospective subgroup analyses of randomized trials. Some trials reported a favourable impact of taxanes on the prognosis of TNBC [35, 50–52], while others did not show any additional benefit from taxane therapy in this population [53, 54]. Preclinical and clinical data regarding the efficacy of anthracyclines in TNBC are limited and contradictory. A meta-analysis of randomized trials comparing anthracycline-based regimens versus CMF reported that TNBC patients have a 23% reduction in the risk of recurrence with the use of anthracyclines [55]. Conversely, a retrospective analysis from the MA.5, which compared CMF versus CEF, revealed an almost significantly (P = 0.06) higher disease-free survival (DFS) for basal-like breast cancer receiving CMF (5y–DFS: CMF = 0.71; CEF = 0.51) [56].

The activity of alkylating agents in TNBC is hard to measure. Retrospective studies suggest that TNBC may have particular sensitivity to alkylating agents. Among these, a study by Colleoni and colleagues reported that classical CMF had greatest benefit in patients with triple-negative to node-negative breast cancer [22]. Regarding the choice of the adjuvant chemotherapy regimen in this subtype, the St. Gallen panel strongly endorsed both anthracyclines and taxanes and did not believe that platinum or regimens emphasizing alkylating agents were specifically required [10].

Thus far, tools to personalize chemotherapy use in breast cancer are lacking and biomarker studies for chemotherapy benefit have been disappointing. Because of breast cancer heterogeneity, individual biomarkers are unlikely to be useful in aiding in the current clinical management decision-making process. Multiple parameters can influence response to chemotherapy but none has been endorsed into clinical use because of their inability to predict response to treatment.

A biomarker frequently investigated is p53, which is a transcription factor that mediates antiproliferative mechanisms in response to various forms of cellular stresses, in particular DNA damage [57]. As a predictive biomarker for treatment response, the role of p53 remains unclear. Preclinical studies suggested p53-dependent anthracycline-induced apoptosis and p53-independent taxane activity [58, 59].

Still, dedicated clinical research has not defined a predictive role for p53 mutations. In the past, many reports have retrospectively evaluated the role of p53 in subgroups from biologically unselected breast cancer trials. However, the clinical data were contradictory and inconclusive, and no robust predictive correlation was demonstrated [60–64]. The only study designed to prospectively evaluate the predictive role of p53 was the neoadjuvant phase III EORTC 10994/BIG 00–01 trial [65]. In this study, although p53 status was prognostic for overall survival, it was found to be not predictive of sensitivity to taxanes. In the retrospective exploratory analysis performed on samples collected in the context of the prospective BIG 02–98 randomized phase III clinical trial, it was found that p53 mutations had no value in predicting response to docetaxel therapy in node-positive breast cancer patients [66]. These findings exemplify how a single biomarker may be inadequate to predict treatment response to chemotherapy. Actually p53 should be considered as a surrogate measure for cellular capacity for apoptosis and cell proliferation control rather than as a drug target.

TOPO2A, the molecular target of anthracyclines, has also been extensively investigated. The association of TOPO2A status with anthracycline efficacy, in patients with HER2-amplified tumours, has been addressed by many studies, but the reported results are conflicting, and the use of TOPO2A as a predictive factor is actually an issue of dispute [67–74]. Actually, while some studies showed a lack of association between TOPO2A expression and responsiveness to anthracyclines [67], others demonstrated TOPO2A overexpression as being associated with higher response rates in patients treated with anthracycline combinations [71–73] but also with shortened survival [72]. The definitive answer after years of disagreement seems to come by the meta-analysis by Di Leo and colleagues showing that patients with HER2 non-amplified or TOPO2A normal tumours gain some additional benefit from treatment with anthracyclines. Therefore, there is no evidence for the use of anthracyclines only in patients with HER2-amplified or TOPO2A-aberrated tumours [55]. The prediction of treatment efficacy according to one gene only might denote an incomplete approach, and other biological factors, such as stromal activation and markers of immune response, might have a role in predicting the responsiveness to anthracyclines [69]. In the era of personalized medicine, a multifactorial approach should be taken and should be the challenge of the future.