Surgery is considered the primary treatment for breast cancer in early-stage patients. The endpoints of breast cancer surgery include the complete resection of the primary tumour, achieving negative margins as this reduces the risk of local recurrence. Moreover, surgery permits the pathologic staging of the tumour and axillary lymph nodes (LN) to make prognostic information available.

Lumpectomy is defined as complete surgical resection of a primary tumour with the objective of achieving widely negative margins (ideally a 1-cm margin). It is applicable in most patients with stage I and stage II invasive carcinomas [63].

Relative contraindications to lumpectomy include small breast size, large tumour size (>5 cm) and collagen vascular disease. Lumpectomy is absolutely contraindicated if there is multifocal disease, a history of previous radiation therapy to the area of treatment, inability to undergo radiation therapy for invasive disease and pregnancy or persistent positive margins following attempts at conservation. Factors that are often considered, but should not be detrimental, include axillary node involvement and tumour location. Although achieving a good cosmetic result is important, it should never be more important than the clinical priority of obtaining negative surgical margins, so lesions involving Paget’s disease of the nipple should be treated with excision of the nipple–areolar complex and reconstruction. Larger lesions in patients who are concerned about the cosmetic result may be better served by standard modified radical mastectomy and concurrent reconstruction [63].

A total mastectomy is defined as complete removal of all breast tissue to the clavicle superiorly, the sternum medially, the inframammary crease inferiorly and the anterior axillary line laterally, with “en bloc” resection of the fascia of the pectoralis major. The nipple–areolar complex is resected along with a skin paddle to achieve a flat chest wall closure when performing a total mastectomy. A total mastectomy does not refer to removal of any axillary nodes but may be performed in conjunction with a sentinel or axillary node dissection [63].

A modified radical mastectomy is defined as a total mastectomy with axillary lymph node dissection. In contrast, a radical mastectomy is defined as a total mastectomy plus “en bloc” resection of the pectoralis major and axillary lymph node dissection [63].

A relative contraindication to modified radical mastectomy is locally advanced cancer requiring neoadjuvant therapy before surgical intervention. Complications after total mastectomy could include the risk of local recurrence (5–10 %), wound infection, seroma, mastectomy skin flap necrosis, haematoma, chronic pain, incisional “dog ears”, lymphoedema and fibrosis.

Many randomised trials support the contention that mastectomy with axillary lymph node dissection is equivalent to breast-conserving therapy with lumpectomy, axillary dissection and whole-breast irradiation, as primary breast treatment for the majority of women with stage I and stage II BCs [64, 65].

The NSABP conducted a landmark randomised study (NSABP-B06) that established breast-conserving surgery with radiation therapy to be equivalent to modified radical mastectomy. At 20-year follow-up, no significant difference was seen in OS, local or regional DFS or distant disease-free survival between the three treatment groups. Patients in the lumpectomy group without radiation had a significantly higher LR rate (39.2 %) than patients undergoing lumpectomy plus radiation therapy (14.3 %). Patients who underwent modified radical mastectomy had a 10.2 % risk of chest wall recurrence [65].

A typical woman with clinical stage I or stage II BC requires pathological assessment of the axillary lymph node status. Performance of SLN mapping and resection in the surgical staging of the axilla is recommended for assessment of the pathologic status of the axillary lymph nodes in patients with stage I or stage II B disease [23, 66]. This recommendation is supported by results of randomised clinical trials showing decreased arm and shoulder morbidity such as pain, lymphoedema and sensory loss, in patients undergoing SLN biopsy [13, 67]. No significant differences in the effectiveness of the SLN procedure or level I and II dissection in determining the presence or absence of metastases in axillary nodes were seen in these studies. However, not all women are candidates for SLN resection. Potential candidates for SLN mapping and excision should have clinically negative axillary lymph nodes, or a negative core or FNA biopsy of any clinically suspicious axillary lymph node(s) [13, 67]. If the SLN cannot be identified or is positive for metastasis, a formal axillary lymph node dissection should be performed or axillary irradiation administered. The optimal technique for axillary radiation is not established in studies, but the axillary nodes can be included in the breast tangential fields. If lymph node mapping identifies sentinel lymph nodes in the internal mammary chain, internal mammary node excision is considered optional.

Traditional level I and level II axillary dissection required that at least ten lymph nodes should be provided for pathologic evaluation to accurately stage the axilla [68]. Axillary dissection should be extended to include level III nodes only if gross disease is apparent in the level I or II nodes [68].

Additionally, in the absence of definitive data demonstrating superior survival with axillary lymph node dissection or SLN resection, these procedures may be considered optional in patients who have particularly favourable tumours, patients for whom the selection of adjuvant systemic therapy is unlikely to be affected by the results of the procedure, elderly patients and patients with serious co-morbid conditions. Women who do not undergo axillary dissection or axillary lymph node irradiation are at increased risk for ipsilateral lymph node recurrence [69]. Women who undergo mastectomy are appropriate candidates for breast reconstruction.

Induction chemotherapy, also known as primary, pre-operative or neoadjuvant chemotherapy, for BC is indicated in locally advanced breast cancer (LABC) and patients with stage II–III disease in order to downstage the tumour burden and ensure appropriate breast conservation. Neoadjuvant therapy also makes it possible to evaluate drug sensitivity. The achievement of tumour and nodal pathologic complete remission (pCR) or minimal residual disease (MDR <1 cm) favourably affects disease-free survival (DFS) and overall survival (OS) [28].

Several clinical trials have attempted to apply information on drug sensitivity to clinical management on the hypothesis that non-cross-resistant chemotherapy agents increase pCR and ultimately improve survival.

The NSABP B-27 [70] trial was designed to determine the impact of adding docetaxel (DXT) after four cycles of induction AC (doxorubicin [ADR] plus cyclophosphamide [CPM]) on clinical CR, pCR rates, DFS and OS. Compared to pre-operative AC alone, pre-operative AC followed by DXT increased the clinical complete response rate (40.1 % vs. 63.6 %; p < .001), the overall clinical response rate (85.5 % vs. 90.7 %; p < .001), the pathologic complete response rate (13.7 % vs. 26.1 %; p < .001) and the proportion of patients with negative nodes (50.8 % vs. 58.2 %; p < .001) [70]. Comparable results were observed by the EORTC trial 10902 (European Organization for Research and Treatment of Cancer). Patients received four cycles of fluorouracil, EPR and cyclophosphamide (FEC) administered either pre- or post-operatively. No difference was observed between the two groups in terms of OS and progression-free survival (PFS) and time to locoregional recurrence (p = 0.38, p = 0.27, p = 0.61, respectively) [71].

The German Pre-operative ADR and DXT Study (GEPARTRIO) [72] included 2090 patients, with large operable or locally advanced breast cancer (LABC), and the tumour reduction was assessed by ultrasound imaging (response guided) in order to make a decision on further treatment. No evidence of a difference in response to neoadjuvant chemotherapy was found by tumour stage when analysis was adjusted for baseline characteristics. Nevertheless, exploratory analysis suggested that response-guided neoadjuvant chemotherapy might improve survival and is most effective in hormone receptor-positive tumours. If confirmed, the response-guided approach could provide a clinically meaningful advantage for the neoadjuvant over the adjuvant approach in early breast cancer [73].

Bevacizumab is a recombinant humanised monoclonal IgG1 antibody that binds to and inhibits the biological activity of human vascular endothelial growth factor (VEGF) in in vitro and in vivo assay systems [74]. Bear et al. [75] conducted a trial involving 1206 HER2-negative patients. Patients were randomised to receive bevacizumab or not over the first six cycles of chemotherapy. The addition of bevacizumab significantly increased the pCR rates (28.2 % vs. 34.5 %, p = 0.02), particularly in the hormone receptor-positive subcategory of patients (15.1 % vs. 23.2 %, p = 0.007) [75]. The benefit of adding bevacizumab was confirmed in the GeparQuinto trial, carried out by von Minckwitz et al. [76]; patients with HER2-negative BC were randomly assigned to receive induction therapy with or without concomitant bevacizumab treatment. The addition of bevacizumab led to a significant increase of pCR rate (p = 0.04). Meanwhile, in this case, the benefit of bevacizumab treatment on pCR was observed only in the hormone receptor-negative subgroup (27.9 % vs. 39.3 %, p = 0.003).

Induction endocrine treatment in postmenopausal, HR-positive, breast cancer patients has been limited to patients who were not suitable for chemotherapy and surgery. Earlier phase II studies with tamoxifen that focused primarily on elderly and/or frail patients were often unselected for HR status of the tumour and showed response rates ranging from 49 to 68 % [77]. Conflicting results are reported in the literature on the value of factors predictive of response in the neoadjuvant or induction therapy approach. In the subset of patients with ER-positive tumours, pCR rates range from 2 to 10 %, which suggests that other primary endpoints must be considered within this tumour subset [77].

Results from two randomised trials on induction endocrine therapy in HR-positive postmenopausal patients support the theory of a relationship between the probability of response and the degree of HR expression [78, 79]. Higher ER levels are significantly related to a greater probability of response in both studies. In addition, a positive significant correlation between the ER level and the degree of Ki-67 suppression after 2 and 12 weeks of endocrine treatment was reported by Dowsett et al. [80].

The level of expression of ER and progesterone (PgR) might be associated with the probability of response to induction chemotherapy. In a retrospective analysis involving 533 patients, no pCR was observed within the cohort of patients considered as highly endocrine responsive (ER and PgR ≥50 %). This compares with 3.3 % of those with ER or PgR expressed in 0–49 % of the cells and 17.7 % of those patients who are HR negative (p < 0.0001) [81]. Conversely, even with the higher incidence of pCR, a statistically significantly worse DFS and OS were observed for patients with ER and PgR absent tumours vs. patients with tumours expressing high ER and PgR (HR 6.4, 95 % CI 3.5–11.6, for DFS; HR 3.6, 95 % CI 2.4–5.6, for OS); this study conducted by Colleoni et al. shows that response and outcome after induction chemotherapy are correlated with the degree of expression of steroid hormone receptors [81].

Aromatase inhibitors (AIs) block the conversion of androgens to oestrogens and reduce oestrogen levels in tissue and plasma [82]. Third-generation AIs include the non-steroidal inhibitors, letrozole and anastrozole, and the steroidal inhibitor, exemestane. The results of large trials conducted in the metastatic and adjuvant setting [80, 83], indicating better outcomes among women given with AIs than those given with tamoxifen stimulated the investigation of these agents in the induction setting in postmenopausal women with HR-positive tumours.

A randomised study (P024 trial) compared 4 months of letrozole or tamoxifen as induction therapy for postmenopausal women with ER- and/or PgR-positive stage II or III breast cancer, which is considered to be unsuitable for breast-conserving surgery (BCS). Letrozole increased the clinical response rate (55 % vs. 36 %, p < 0.001) and the BCS rate (45 % vs. 35 %, p = 0.022) when compared with tamoxifen [84].

In the IMPACT study [79], postmenopausal women with ER-positive tumours received induction tamoxifen, anastrozole or a combination of tamoxifen and anastrozole for 3 months. The response rates were similar for all treatments, 37 % vs. 36 % vs. 39 %, respectively. In the PROACT trial, anastrozole and tamoxifen yielded a similar response rate [85]. Seo et al. conducted a meta-analysis on three studies to show an improved breast-conserving surgery rate in patients receiving AI [86]. In another randomised study in postmenopausal patients with HR-positive breast cancer, exemestane significantly increased the clinical response rate and the rate of BCS (36.8 % vs. 20 %, p = 0.05) [87].

In the ACOSOG Z1031trial, 374 postmenopausal women with clinical stage II or III ER-positive breast cancer were randomised to receive anastrozole, exemestane or letrozole for 16–18 weeks before surgery. The response rates observed were not statistically different [88].

The optimal duration of induction endocrine therapy has not yet been determined. However, an improved response rate has been reported with induction letrozole when the duration of therapy was prolonged beyond 3 months [89]. According to these observations, several authors recommend prolonging endocrine therapy for a minimum of 4–8 months [77].

Crosstalk between the oestrogen receptor (ER) and the phosphoinositide 3-kinase (PI3K)/Akt/mammalian target of rapamycin (mTOR) pathways is a mechanism of resistance to endocrine therapy, and the blockade of both pathways enhances antitumour activity in preclinical models [89, 90]. A new approach to restore endocrine responsiveness in breast cancer might be to use the combination of an AI with a signal transduction inhibitor as a PI3K/mTOR antagonist [91]. Baselga et al. [92] included 270 postmenopausal women with operable ER-positive breast cancer who were randomly assigned to receive 4 months of induction treatment with letrozole (2.5 mg/day) and either everolimus (10 mg/day) or placebo. Clinical response rate in the everolimus arm was higher than that of letrozole alone (i.e. placebo; 68.1 % vs. 59.1 %), which was statistically significant (p = 0.062).

Limited data are available on induction treatments for premenopausal patients. Masuda et al. [93] performed the unique double-blind, multicentre, randomised trial in 197 premenopausal patients with ER-positive, HER2-negative, operable breast cancer were to receive monthly goserelin plus either anastrozole and tamoxifen placebo (n = 98) or tamoxifen and anastrozole placebo (n = 99) over 24 weeks before surgery. More patients in the anastrozole group achieved complete or partial response compared to the tamoxifen group (anastrozole 70.4 % vs. tamoxifen 50.5 % [p = 0.004]).

About 20 % of patients with breast cancer will show overexpressed HER2 gene amplification, which is associated with an aggressive clinical course of the disease. Trastuzumab, a monoclonal antibody (mAb) targeting HER2, has established activity in both the metastatic and adjuvant settings for these patients. Several single-arm phase II trials in the induction therapy setting have been performed suggesting that the addition of trastuzumab to conventional chemotherapy shows promising antitumour activity, with pCR rates ranging between 12 and 76 % [94].

The NOAH (NeOAdjuvant Herceptin) study conducted by Gianni and colleagues [95] involved 235 patients with HER2-positive locally advanced or inflammatory disease. They randomly received either ADR/paclitaxel (TXL)/cyclophosphamide (CPM), methotrexate (MTX) and fluorouracil (5-FU) chemotherapy with or without simultaneous administration of trastuzumab for 30 weeks. This study achieved a significant increase in 3-year event-free survival (EFS) of 71 % with trastuzumab vs. 56 % without (p = 0.013) and an increase in the pCR rate of 38 % vs. 19 % (p = 0.001) [95].

In another study, 120 patients with stage II and III HER2-positive BC, ineligible for BCS, were randomly assigned to receive four cycles of epirubicin (EPR)/CPM (EC), followed by four cycles of DXT with or without trastuzumab concurrently with the DXT. The difference between the groups, in terms of pCR, was not statistically significant (26 % with trastuzumab vs. 19 % without) [96].

The Austrian Breast and Colorectal Cancer Study Group randomised 93 patients with HER2-positive BC to receive six cycles of EPR-DXT or EPR-DXT-capecitabine induction therapy with or without trastuzumab. The difference between the groups, in terms of pCR, was not statistically significant (pCR 38.6 % vs. 26.5 %, p = 0.212) [97].

Buzdar and colleagues [98] investigated the effect of the timing of trastuzumab administration with anthracycline and taxane induction chemotherapy in a randomised phase III trial (282 patients). Women with operable HER2-positive invasive breast cancer were randomly assigned into group A (140 pts) (“sequential group”) with 5-fluorouracil-EPR (75 mg m²)-CPM (FEC-75) for four cycles followed by weekly TXL plus weekly trastuzumab for 12 weeks, and group B (“concurrent treatment”) group (142 pts) with weekly TXL and trastuzumab for 12 weeks followed by four cycles of FEC-75 and weekly trastuzumab. The pCR rate was the primary endpoint. The sequential group reached 56.5 % pCR compared to 54.2 % in the concurrent group. Left ventricular ejection fraction dropped in 0.8 % of patients who received sequential treatment and in 2.9 % of patients who received concurrent treatment; by week 24, LVF had dropped to 7.1 % and 4.6 % of patients in the sequential and concurrent group, respectively. Both treatments reached a high response rate, but it should be stressed that concurrent administration of trastuzumab with anthracyclines offers no additional benefit and is not warranted [98].

Lapatinib is an epidermal growth factor receptor inhibitor for use in combination with capecitabine for the treatment of advanced or metastatic HER2-positive breast cancer in women who have received prior therapy and also for use in combination with letrozole to treat HR-positive and HER2-positive ABC in postmenopausal women for whom hormonal therapy is indicated [99]. The GeparQuinto phase III trial compared four cycles of EPR/CPM (EC) followed by DXT administered concurrently with either trastuzumab or lapatinib. This face-to-face comparison of trastuzumab and lapatinib showed that pCR rate with chemotherapy and lapatinib was significantly lower than that with chemotherapy and trastuzumab (p < 0.04) [100]. Baselga et al. [101] conducted the NeoALTTO trial comparing lapatinib plus Taxol vs. trastuzumab plus TXL vs. concomitant lapatinib and trastuzumab plus TXL (152 pts). The pCR rate obtained by the combination, trastuzumab plus lapatinib, was significantly better (p = 0.0001) [101]. An updated analysis performed by Piccart-Gebhart M et al. [102] demonstrated that patients achieving pCR had a better outcome compared with patients not achieving pCR [101].

Pertuzumab is a recombinant humanised MAB that targets the extracellular dimerisation domain (subdomain II) of HER2 and thus blocks ligand-dependent heterodimerisation of HER2 with other HER family members, including EGFR, HER3 and HER4. Pertuzumab inhibits ligand-initiated intracellular signalling through both mitogen-activated protein (MAP) kinase and phosphoinositide 3-kinase (PI3K) signalling pathways, resulting in cell growth arrest and apoptosis. Pertuzumab also mediates antibody-dependent cell-mediated cytotoxicity [103].

Two pivotal clinical studies have been performed with pertuzumab, trastuzumab and DXT in induction treatment for patients with HER2 positive with operable locally advanced or inflammatory breast cancer and a primary tumour size >2 cm. In the first trial, the combination of pertuzumab plus trastuzumab plus DXT, as compared with placebo plus trastuzumab plus DXT, significantly prolonged PFS (p < 0.001), with no increase in cardiac toxic effects [103]. The second study confirmed the previous results, and patients who received pertuzumab plus trastuzumab plus DXT had a significantly improved pCR rate (45.8 %) compared to those given trastuzumab plus DXT, without substantial differences in tolerability [104, 105].

For patients with TNBC, Huober et al. reported a high pCR rate of 39 % in 509 patients treated with TAC (DXT/ADR/CPM) or TAC-NX (DXT/ADR/CPM – vinorelbine/capecitabine) [106]. Di Leo et al. [107] suggested a beneficial outcome of anthracyclines compared to CMF therapy in terms of DFS in 294 patients with TNBC. However, recent studies have cast doubt on the role of anthracyclines in early-stage TNBC [108]. An analysis conducted by von Minckwitz et al. concluded that for patients with TNBC in particular, treatment with higher cumulative doses of anthracyclines (≥300 mg/m²) and taxanes (≥400 mg/m²) provides higher pCR rates compared with lower cumulative dose regimens [109].

Recently, the use of platinum agents has received renewed interest in the treatment of TNBC. The CALGB 40603 study in TNBC patients of carboplatin or bevacizumab combined with chemotherapy resulted in an increased pCR breast rate: 60 % of patients who received carboplatin achieved pCR breast compared with 46 % of those who did not (p = 0.0018). Patients treated with a bevacizumab-containing regimen had a pCR breast rate of 59 % compared with 48 % of those who were not (p = 0.0089). Patients assigned to both agents had the highest pCR breast rate (67 %). A major limitation of this study was that it was not powered to demonstrate any difference on DFS and OS [110].

The addition of carboplatin to a regimen of a taxane, an anthracycline and bevacizumab significantly increases, in TNCB patients, the rate of patients achieving a pCR. In the GeparSixto study [111], including TNBC patients, 53.2 % achieved a pCR with carboplatin, compared with 36.9 % without (p = 0.005).

A meta-analysis has indicated that novel induction schedules in TNBC patients have achieved a significant improvement of pCR rate, particularly among those treated with carboplatin (p = 0.001)- or bevacizumab (p = 0.003)-containing regimens. Moreover, adding carboplatin for TNBC results in an absolute benefit of 13.8 % compared with control arms. However, such treatment had no impact on DFS and OS [112, 113].

Surgical treatment after induction chemotherapy focuses on the management of the tumour itself and the axillary spread. It is important to consider that one of the purposes of induction chemotherapy is breast conservation with accurate disease staging allowing avoidance of a broad axillary dissection. The development of sentinel node biopsy provides a remarkable advantage for breast conservation after induction chemotherapy [13].

If the tumour responds to induction therapy, lumpectomy plus axillary lymph node dissection or sentinel lymph node procedure may be considered. If induction therapy results in non-response or progressive disease or if lumpectomy is not possible, then mastectomy is considered with axillary staging. If a pre-chemotherapy sentinel lymph node procedure was performed and the sentinel lymph node was pathologically negative, then further axillary lymph node staging is not necessary. If a pre-chemotherapy sentinel lymph node procedure was performed and the sentinel lymph node was positive, then a level I/II axillary lymph node dissection should be performed [12, 13].

In the case of an inoperable tumour that fails to respond or where response is insufficient after several courses of induction therapy, an alternative induction must be taken into consideration followed by local treatment using either radiotherapy or surgery, or both [12, 13].

Surgery should be followed by individualised chemotherapy such as taxanes if the full course of planned chemotherapy was not administered pre-operatively, as well as breast and regional lymph node irradiation. There is no role for post-operative chemotherapy if a full course of standard chemotherapy has been completed pre-operatively. Endocrine therapy is carried out for 5 years in women with ER- and/or PR-positive tumours [12, 13].

The purpose of radiation therapy after breast-conserving surgery is to eradicate local subclinical residual disease so reducing local recurrence rates by approximately 75 %. Based on the results from several randomised studies, irradiation of the intact breast is considered the standard of care, even in the lowest-risk disease with the most favourable prognostic features [115].

There are two general approaches used to deliver radiation therapy: conventional external beam radiotherapy (EBRT) and partial breast irradiation (PBI). WBRT consists of EBRT delivered to the breast at a dose of 50–55 Gy over 5–6 weeks. This is often followed by a boost dose specifically directed to the area in the breast where the tumour was removed [115].

Common side effects of radiation therapy include fatigue, breast pain, swelling and skin desquamation. Late toxicity (lasting ≥6 months after treatment) may include persistent breast oedema, pain, fibrosis and skin hyperpigmentation [13].

Partial breast irradiation (PBI) is employed in early-stage breast cancer after breast-conserving surgery as a way of delivering larger fraction sizes whilst maintaining a low risk of late effects. Techniques that can deliver this therapy include interstitial brachytherapy and intracavitary brachytherapy (a balloon catheter inserted into the lumpectomy site [i.e. MammoSite]). Treatment is usually administered twice daily for 5 days. In several nonrandomised studies, these techniques have shown low local recurrence rates comparable to those of EBRT. The American Society of Breast Surgeons recommends several selection criteria when patients are being considered for treatment with accelerated PBI (2): age ≥45 years, invasive ductal carcinoma or ductal carcinoma in situ (DCIS), total tumour size (invasive and DCIS) ≤3 cm, negative microscopic surgical margins of excision and ALN or SLN negative [116].

An observational study with data from the SEER-Medicare Linked Database on 35,947 women aged >66 years-old with invasive breast cancer (79.9 %) or DCIS (20.1 %) established that standard EBRT was associated with a higher 5-year breast preservation rate than either lumpectomy alone or brachytherapy. These results are controversial because the study data did not take into account the use of the newest forms of brachytherapy which might reduce the applicability of these findings [117–120].

According to two major studies, single-dose radiotherapy delivered during, or soon after, surgery for breast cancer is a viable alternative to conventional EBRT in selected patients who are at low risk for local recurrence [121–123].

The guidelines developed by the American Society of Clinical Oncology (ASCO), based on several prospective, randomised clinical trials, recommend that postmastectomy radiation therapy must be performed according to some important criteria: positive postmastectomy margins, primary tumours >5 cm and involvement of ≥4 lymph nodes [124].

Patients with >4 positive lymph nodes should also undergo prophylactic nodal radiation therapy at doses of 45–50 Gy to the axillary and supraclavicular regions. For patients without ALND node involvement, axillary radiation therapy is not recommended.

Postmastectomy radiotherapy was shown in previous meta-analyses to reduce the risks of both recurrence and breast cancer mortality in all women with node-positive disease considered together. However, the benefit in women with only 1–3 positive LNs is uncertain. A meta-analysis of individual data for 8135 patients [125] randomly assigned to treatment groups during 1964–86 in 22 trials of radiotherapy to the chest wall and regional lymph nodes after mastectomy and axillary surgery vs. the same surgery but no radiotherapy has examined the question. Follow-up lasted 10 years for recurrence and for mortality. In this analysis 3786 women had axillary dissection to at least level II and had zero, 1–3 or ≥4 positive nodes. All were in trials in which radiotherapy included the chest wall, supraclavicular or axillary fossa (or both) and internal mammary chain. For 700 women with axillary dissection and no positive nodes, radiotherapy had no significant effect on locoregional recurrence (2P >0 · 1). For 1314 women with axillary dissection and 1–3 positive nodes, radiotherapy reduced locoregional recurrence (2P <0 · 00001), overall recurrence (2P = 0 · 00006) and BC mortality (2P = 0 · 01); 1133 of these women were in trials in which systemic therapy with CMF or tamoxifen was given in both trial groups, and for them, radiotherapy once more reduced locoregional recurrence (2P <0 · 00001), overall recurrence (2P = 0 · 00009) and BC mortality (2P = 0 · 01). For 1772 patients with axillary dissection and ≥4 positive nodes, radiotherapy reduced locoregional recurrence, overall recurrence and BC mortality (2P = 0 · 04, 2P = 0 · 0003 and 2P < 0 · 00001, respectively). The clear conclusion is that after mastectomy and axillary dissection, radiotherapy reduced both recurrence and BC mortality in the women with one to three positive lymph nodes in these trials, even when systemic therapy was given [125].

Tumours less than 0.5 cm in greatest diameter that do not involve the lymph nodes are considered as favourable, and as adjuvant systemic therapy does not add any benefit, it is not recommended as treatment of the invasive BC. Endocrine therapy may be considered to reduce the risk of a second contralateral BC, especially in those with ER-positive disease. The NSABP demonstrated a correlation between the ER status of a new contralateral breast tumour and the original primary tumour, reinforcing the notion that endocrine therapy is unlikely to be an effective strategy to reduce the risk of contralateral BC in patients diagnosed with ER-negative tumours [141].

Patients with invasive ductal or lobular tumours >5 cm in diameter and no lymph node involvement may be separated into those patients with a low risk of recurrence and those with unfavourable prognostic features that allow consideration of adjuvant therapy. Unfavourable prognostic features include intramammary angiolymphatic invasion, high nuclear grade, high histological grade, HER2-positive status or HR-negative status. The decision to undertake an endocrine therapy and chemotherapy in these relatively lower-risk subgroups of women must be based on balancing the expected absolute risk reduction with the patient’s general condition in the context of likely toxicity to achieve that incremental risk reduction [12, 13]. The incremental benefit of combination chemotherapy in patients with lymph node-negative, HR-positive BC may be relatively small [142]. Therefore, the tumour HR status must be included as one of the factors considered when making decisions about chemotherapy for patients with node-negative, HR-positive BC. Patients for whom this evaluation may be especially important are those with tumours characterised as 0.6–1.0 cm and HR positive that are grade 2 or 3 or have unfavourable features or greater than 1 cm and HR positive and HER2 negative. However, chemotherapy should not be withdrawn from these patients solely on the basis of ER-positive tumour status [13, 142, 143]. Assessment of the recurrence score is indicated in these cases taking into account the context of other elements of risk.

Patients with LN-positive disease are candidates for chemotherapy and if the tumour is HR positive, for the addition of endocrine therapy [13]. In postmenopausal women, with HR-positive disease, an AI should be utilised unless a contraindication exists. In premenopausal women, adjuvant tamoxifen is recommended. If both chemotherapy and tamoxifen are administered, data from the Intergroup trial 0100 suggest that delaying initiation of tamoxifen until after completion of chemotherapy improves DFS compared with concomitant administration [144]. Consequently, chemotherapy followed by endocrine therapy should be the preferred therapy sequence.

It is also important to consider subsets of patients with early BC of the usual histologies based upon responsiveness to endocrine therapy and trastuzumab (i.e. hormone receptor status, HER2 status). Patients are then further stratified based upon risk for recurrence of disease assessed by anatomic and pathologic characteristics (i.e. tumour grade, tumour size, axillary lymph node status, angiolymphatic invasion [13, 23].

Patients with invasive BCs that are ER or PR positive should be considered for adjuvant endocrine therapy regardless of patient age, lymph node status or whether or not adjuvant chemotherapy is to be administered; HER2 amplification is a marker of relative endocrine resistance independent of type of endocrine therapy [145, 146]. However, the use of adjuvant endocrine therapy in the majority of women with HR-positive BC regardless of menopausal status, age or HER2 status of the tumour is recommended.

The most definitely recognised adjuvant endocrine therapy is tamoxifen for both premenopausal and postmenopausal patients [143]. In ER-positive BC patients, adjuvant tamoxifen decreases the annual probabilities of recurrence by 39 % and the annual probabilities of death by 31 % irrespective of the use of chemotherapy, patient age, menopausal status or axillary lymph node (LN) status. Prospective, randomised trials demonstrate that the optimal duration of tamoxifen appears to be 5 years. In patients receiving both tamoxifen and chemotherapy, chemotherapy should be given first, followed by tamoxifen [144].

The ten-year duration of tamoxifen treatment in women with stage I–III hormone receptor (HR)-positive breast cancer is based on the data collected mainly from two randomised trials, the aTTom and ATLAS in which more than 13,000 randomised patients were analysed. Adjuvant endocrine therapy with tamoxifen for 5 years should be offered to pre-/perimenopausal women. After this, if the patient remains premenopausal, they should be offered continued tamoxifen for an additional 5 years. If postmenopausal, patients should be treated with continued tamoxifen or an aromatase inhibitor (AI) for an additional 5 years, totalling 10 years of adjuvant endocrine therapy [13, 62, 115, 147].

In the ATLAS trial, in the second decade after diagnosis, women who continued on tamoxifen had a 25 % lower recurrence rate and 29 % (p = 0.002) lower breast cancer mortality rate compared with women who stopped after 5 years of tamoxifen. The risk of dying due to breast cancer at 5–14 years after diagnosis was 12.2 % for continuing tamoxifen vs. 15 % for those who only had 5 years of treatment, representing an absolute gain of 2.8 %, favouring continuing treatment with tamoxifen for 10 years (p = 0.01) [148].

In the aTTom study, women with breast cancer who had been taking tamoxifen for 5 years were randomised to continue receiving tamoxifen for another 5 years or stop taking the drug. Allocation to continue tamoxifen reduced breast cancer recurrence (p = 0.003). Longer treatment also reduced breast cancer mortality (392 vs. 443 deaths after recurrence, p = 0.05) [149].

Combining the similar results of aTTom and its international counterpart ATLAS enhances statistical significance of long-term effects of continuing adjuvant tamoxifen in terms of recurrence (p < 0.0001), breast cancer mortality (p = 0.002) and overall survival (p = 0.005) benefits [149].

The disadvantages of taking tamoxifen for 10 years may include continuing menopausal symptoms, such as night sweats and hot flashes. Longer duration of treatment increases the risk of endometrial cancer: in the aTTom trial, there were 102 cases and 37 (1.1 %) deaths attributed to endometrial cancer in the 10-year tamoxifen arm, compared with 45 cases and 20 (0.6 %) deaths in the 5-year group. However, the benefits demonstrated by a longer duration of treatment with tamoxifen outweighed the possible risks [149].

A number of studies have evaluated AIs in the treatment of postmenopausal women with early-stage BC. These studies have employed the AIs as initial adjuvant therapy, as sequential therapy following 2–3 years of tamoxifen or as extended therapy following 4.5–6 years of tamoxifen. The AIs are not active in the treatment of women with functioning ovaries. Tamoxifen should not be used in women whose ovarian function cannot be reliably assessed [13]. The results from two prospective, randomised clinical trials have provided evidence of an OS benefit for patients with early-stage BC receiving initial endocrine therapy with tamoxifen followed sequentially by anastrozole (p = 0.045) or exemestane (p = 0.05) when compared with tamoxifen as the only endocrine therapy [150, 151]. In addition, the National Cancer Institute Canada Clinical Trials Group (NCIC CTG) MA-17 study results in women who had completed 4.5–6 years of adjuvant tamoxifen demonstrated that extended therapy with letrozole provides benefit in postmenopausal women with HR-positive, early BC. The results showed fewer recurrences or new contralateral BCs with extended therapy (p < 0.001). No difference in OS was demonstrated, although there was a survival advantage in the subset of patients with axillary lymph node-positive disease (p = 0.04) [152].

However, no survival differences have been reported for patients receiving initial adjuvant therapy with an AI vs. first-line tamoxifen [153, 154].

Tamoxifen and AIs have different side effect profiles. Both contribute to hot flashes and night sweats and may cause vaginal dryness. Aromatase inhibitors are more commonly associated with musculoskeletal symptoms, osteoporosis and increased rate of bone fracture, whilst tamoxifen is associated with an increased risk of uterine cancer and deep venous thrombosis.

The ATAC study demonstrated that anastrozole is superior to tamoxifen or the combination of tamoxifen and anastrozole in the adjuvant endocrine therapy of postmenopausal women with HR-positive BC (p = 0.003) [146]. Nevertheless, no difference in OS has been observed (p = 0.2).

Breast International Group (BIG) BIG1-98 tested the use of tamoxifen alone for 5 years, vs. letrozole alone for 5 years, or tamoxifen for 2 years followed sequentially by letrozole for 3 years or letrozole for 2 years followed sequentially by tamoxifen for 3 years. DFS was superior in the letrozole-treated women (p = 0.003), but no difference in OS has been observed. The overall incidence of cardiac adverse events was similar (letrozole, 4.8 %; tamoxifen, 4.7 %). However, the incidence of grade 3–5 cardiac adverse events was significantly higher in the letrozole arm, and both the overall incidence and incidence of grade 3–5 thromboembolic events were significantly higher in the tamoxifen arm. A higher incidence of bone fracture was observed for women in the letrozole arm compared with those in the tamoxifen arm (9.5 % vs. 6.5 %) [155].

Four trials have studied the use of tamoxifen for 2–3 years followed sequentially by a third-generation AI vs. continued tamoxifen. The Italian Tamoxifen Anastrozole (ITA) [155, 156] and the Intergroup Exemestane Study (IES) trials both randomised postmenopausal women with BC who had completed a total of 2–3 years of tamoxifen to AI [150]. An improvement in terms of DFS was seen in all trials but not in OS except with the exemestane study in which a significant difference was observed only in patients with ER-positive tumours (log-rank test p = 0.05) compared to unknown HR status. The Austrian Breast and Colorectal Cancer Study Group (ABCSG) trial 8 [157] and the Arimidex–Nolvadex (ARNO 95) trial [158] demonstrated the superiority of sequential AIs in disease-free survival but gave conflicting results in terms of survival. A meta-analysis gathering updated data of the above-mentioned studies ABCSG 8, ARNO 95 and ITA showed significant improvement in overall survival (p = 0.04) with a switch to anastrozole [159].

The differences in design and patient populations among the studies of the AIs do not allow for the direct comparison of the results of these studies. Thus, it is not known whether initial, sequential or extended use of adjuvant aromatase inhibitors is the optimal strategy. The optimal duration of aromatase inhibitor treatment is also not known, nor is the optimal use vis-à-vis chemotherapy established [13]. The various studies are consistent in demonstrating that the use of a third-generation AIs in postmenopausal women with HR-positive BC lowers the risk of recurrence, including ipsilateral breast tumour relapse, contralateral BC and distant metastatic disease, compared to tamoxifen alone whether the AI is used as initial adjuvant therapy, sequential therapy or extended therapy. Thus, it is recommended that postmenopausal women with early BC should receive an aromatase inhibitor as initial adjuvant therapy, as sequential with tamoxifen or as extended therapy in those situations where endocrine therapy is to be utilised [13, 62].

It is very important to emphasise that the aromatase inhibitors do not adequately suppress ovarian oestrogen synthesis in women with functioning ovaries. Moreover, some patients with chemotherapy-induced amenorrhoea may regain ovarian function whilst receiving AI therapy and even become pregnant [160]. Premenopausal women should not be given with therapy with an AI outside the limits of a clinical trial. Women who are premenopausal at the time of diagnosis and who become amenorrheic with chemotherapy may have continued oestrogen production from the ovaries in the absence of menses. Serial assessment of circulating LH, FSH and oestradiol to assure a true postmenopausal status is mandatory if this subset of women is to be considered for therapy with an AI [83].

The historical studies of CMF chemotherapy vs. no chemotherapy have shown disease-free and overall survival advantages with CMF chemotherapy [161]. Trials adding an anthracycline as in CAF/FAC (CPM, ADR, 5-FU) have shown that the use of ADR and the use of full-dose chemotherapy regimens are important [162, 163]. In the Early BC Trialists’ overview of polychemotherapy, comparison of anthracycline-containing regimens with CMF showed a further 12 % reduction in the annual probabilities of recurrence (p = 0.006) and a further 11 % reduction in the annual odds of death (p = 0.02) with anthracycline-containing regimens [161].

The results of two randomised trials comparing AC with or without sequential TXL therapy in women with axillary node-positive BC suggest improvement in DFS rates and OS with the addition of TXL [163, 164]. On retrospective analysis, the apparent advantage of the TXL-containing regimen appears greater in women with ER-negative BCs.

One trial conducted by Levine et al. [165] in premenopausal women with node-positive BC randomly assigned them to receive CMF therapy or CEF (CPM-EPR-5-FU) chemotherapy using high-dose EPR. Both 10-year relapse-free survival (53 % vs. 63 %; p = 0.009) and OS (70 % vs. 77 %; p = 0.03) favoured the CEF arm. A second trial carried out by the French Adjuvant Study Group compared CEF given at two dose levels of EPR (50 mg/m² q 3 weeks vs. 100 mg/m² q 3 weeks) in pre- and postmenopausal women with node-positive BC. Five-year DFS (55 % vs. 66 %; p = 0.03) and OS (65 % vs. 77 %; p = 0.007) both favoured the EPR 100 mg/m² arm [166].

Another trial compared two dose chemotherapy levels of EC (EPR plus CPM) with CMF in women with node-positive BC. This study showed that higher-dose EC was equivalent to CMF and superior to moderate-dose EC in DFS and OS [167]. An additional randomised trial in patients with axillary LN-positive BC compared six cycles of FEC vs. three cycles of FEC followed by three cycles of DXT [167, 168]. Five-year disease-free survival (78.4 % vs. 73.2 %, adjusted p = 0.012) and overall survival (90.7 % vs. 86.7 %, p = 0.017) were superior with sequential FEC followed by DXT [168].

Final results from a randomised trial comparing DXT, ADR and CPM (TAC) vs. FAC in axillary LN-positive BC demonstrated that TAC is superior to FAC. Estimated 5-year DFS with TAC was 75 %, 68 % for FAC (p = 0.001) and 87 % for survival with TAC and 81 % with FAC (p = 0.008) [169].

At a median follow-up of 73 months, results from the 3-arm randomised NSABP B-30 trial comparing TAC vs. AT vs. AC followed by DXT (ACT) demonstrated that ACT had a significant advantage in DFS (p = 0.006) but not in OS (p = 0.086) when compared with TAC. In addition, both DFS (p = 0.001) and OS (p = 0.034) were significantly increased when ACT was compared with AT, demonstrating non-inferiority compared with TAC [170].