First-line chemotherapy with or without biologic agents for metastatic breast cancer

2.1. Goals of MBC treatment 

Despite better diagnostic tools and new therapeutic agents, at the present the main treatment goal in MBC is still palliation. In fact, the median survival of MBC still ranges between 24 and 30 months [8], [9]. Gain in overall survival (OS) and progression-free survival (PFS), together with symptoms control and quality of life improvement, are the main objectives of palliative treatment.

Treatment could represent an urgent need in order to control an aggressive and symptomatic disease: in this case the first choice is represented by the combination of two or more chemotherapeutic drugs or an association with biologic agents [10]. Conversely, in patients with a less aggressive disease the treatment of choice could be a single-agent CT with or without a molecular target drug or, in presence of endocrineresponsive disease, a hormonal therapy [10]. Negative hormone-receptor status, HER2-positive status, disease-free interval shorter than 2 years, presence of visceral metastatic involvement, and extensive tumor burden are some of the factors that have been described to negatively affect prognosis of patients with MBC [11].

In addition, patients presenting with heavy comorbidities or with hepatic or renal failure, should be considered for tailored treatment, taking into account pharmacokinetics and toxicity profile.

2.2. Endpoints in clinical trials 

In the past, OS was considered the most important endpoint for trials in the setting of MBC. OS guaranties an objective parameter and, at the same time, it is also informative on treatment-related toxicities. Notwithstanding, OS could be influenced by other lines of treatment that patients have received after the participation into the trial. Moreover, the vast majority of clinical trials is underpowered to elicit small differences between treatments in terms of survival. This could be a possible explanation for the lack of survival benefit observed in recent clinical trials testing new agents. In Table 1 randomised clinical trials showing a statistically significant OS advantage for CT are reported.

Table 1. Randomized trials of chemotherapy leading to survival advantage.

	Author, year	N	Arm A	Arm B	Median survival (months) Arm A vs Arm B (p value)
	Jassem, 2001	267	Doxorubicin+Paclitaxel	FAC	23.3 vs 18.3 (0.01)
	O'Shaughnessy, 2002	511	Docetaxel+Capecitabine	Docetaxel	14.5 vs 11.5 (0.01)
	Feher, 2005	410	Epirubicin	Gemcitabine	19.1 vs 11.8 (0.0004)
	Bontenbal, 2005	216	Docetaxel+Doxorubicin	FAC	22.6 vs 16.2 (0.02)
	Icli, 2005	201	Oral Etoposide+Cisplatin	Paclitaxel	14.0 vs 9.5 (0.04)
	Stockler, 2007	325	Capecitabine	CMF	22.0 vs 18.0 (0.02)
	Jones, 2005	449	Docetaxel	Paclitaxel	15.4 vs 12.7 (0.03)
	Albain, 2008	529	Paclitaxel+Gemcitabine	Paclitaxel	18.6 vs 15.8 (0.02)

FAC=5-fluorouracil, doxorubicin, cyclophosphamide; CMF=cyclophosphamide, methotrexate, 5-fluorouracil. Modified from Ref. [8].

Alternatively to OS, some Authors propose the objective response rate (RR), PFS or time to progression (TTP) as surrogates endpoints for OS. Each of them presents some limitation. RR measures the therapeutic benefit and it is not applicable whenever lesions are not accurately measurable (i.e. in presence of bone lesions or pleural effusion); moreover, biologic agents could lead a disease control without significant impact on disease volume, leading to a low RR but an effective disease control.

PFS represents the interval between the start of treatment and the progression of disease or the patients death from any cause. Frequently, and wrongly, TTP is considered synonymous of PFS but TTP measures the interval between the start of treatment and the event of progression or progression-related death. Both PFS and TTP are frequently used as OS-surrogates, but a validation of these parameters is still lacking [12]. The Food and Drug Administration (FDA) usually requires OS as the primary endpoint of clinical trials; in particular, in the past, the FDA evaluated only trials on first-line treatment for MBC in which OS was adopted as endpoint. In fact, since it is both a safety and efficacy parameter, OS is still considered the gold standard endpoint in MBC. However, PFS is gaining the scene as an acceptable alternative endpoint if measured properly. In other words, the improvement in PFS has to be of sufficient magnitude (clinically relevant), supported by a robust statistical design and established with confidence (eg. low rate of patient drop out and/or missing data). Blind design, independent review of the results and the presence of other measures supporting the efficacy (i.e. OS, quality of life, RR) are other characteristics that can allow a confident use of PFS in establishing the therapeutic value of new anticancer agents.

A recent and unique example is the accelerated registration of bevacizumab in MBC based on a trial showing a benefit in PFS but not in OS [13].

2.3. Biologic features, prognostic-predictive factors and therapeutic management 

Recently, the introduction of genomic evaluation of cancer cells DNA and mRNA lead to a new classification of breast cancer together with the concept that BC groups different diseases and no different subtypes of the same disease. The first study proposing a BC classification by gene-expression patterns using DNA microarrays was published in 2000 [14]. The quantitative analysis of gene-expression patterns provided unique and homogeneous “molecular portraits” leading to a classification into four geno-phenotypes: basal-like, HER2-positive, normal-like and luminal (subsequently divided in subgroup A and B).

Sotiriou and Pusztai [15] summarised the results of gene-expression studies and their potential application in terms of clinical practice and prognostic value, as showed in Table 2.

Table 2. Classification of geno-phenotypes of BC and correlation with morphologic and clinical characteristics.

	Geno-phenotype	IHC and genomic features	CK	Morphologic features and clinical characteristics
	Basal-like	ER−/PgR−/HER2−, high grade	CK 5/6 and CK14 expression; (basal/mioepithelial CK), GF-receptors (EGFR, c-kit) and GF expression; BRCA1 dysfunction	Young age
				Bad prognosis
				Shorter survival
				Marked cellular pleomorphism
				High mitotic index
				Frequent apoptotic cells
	ErbB2+	ER−/PgR−/HER2+, high genomic grade
	Luminal-type	A: ER+, HER2−, low histologic and genomic grade	Luminal epithelial CK	Luminal A: indolent and sensitive to endocrine therapy
		B: ER+ (but lower expression respect of Luminal A), HER2+, high histologic and genomic grade		Luminal B: incomplete sensitivity to endocrine therapy, more aggressive than luminal A

ER=estrogen receptor; PgR=progesterone receptor; GF=growth factor; CK=cytokeratins; pts=patients.

Hormone-receptor status and HER2 status have a prognostic and predictive value. Expression of hormonal receptors could be informative in terms of endocrineresponsiveness. Overexpression or amplification of HER2 usually is associated with more aggressive BC; however, trastuzumab, the monoclonal antibody against HER2, could reverse this poor prognosis [16].

Into the attempt to better tailor treatments, the search for predictive factors deserves a huge effort. An interesting model is represented by HER2-positive disease: if overexpression of HER2 is associated with PTEN-deficiency, less benefit from trastuzumab is expected [17]. Recent studies demonstrated that drugs interfering with phosphatidyl-inositol-3-kinase (PI3K) are effective in overcoming this resistance. For other molecular markers such as topoisomerase II-alpha for anthracycline benefit or protein Tau for taxanes benefit [18], [19], the predictive role is still controversial. Probably, the chemosensitivity is a multifactorial process, consequently the value of a single marker is undetectable or poor informative. Finally, multigene predictive models, looking at different molecular pathways, are actually considered more promising rather than a single gene evaluation.

3.1. Single-agent chemotherapy 

Being one of the most effective drugs in breast cancer, anthracyclines and taxanes are now considered the gold standard for first-line treatment. Nevertheless, as they are often used in adjuvant setting, other drugs (e.g. capecitabine, vinorelbine, gemcitabine or ixabepilone) have to be considered as alternatives for the treatment of metastatic disease.

Due to dose-dependent cardiac dysfunction induced by anthracyclines, it is recommended not to exceed a threshold cumulative dose of 450–550mg/m2 for doxorubicin and 800–1000mg/m2 for epirubicin. To reduce the incidence of cardiac toxicity by anthracyclines, new formulations of anthracyclines have been developed in which the antiblastic drug is incapsulated in liposomes [29].

Pegylated liposomal doxorubicin is as effective as conventional doxorubicin in terms of PFS (6.9 months vs 7.8 months). It induces less hair loss, nausea, vomiting and cardiac adverse effects (HR 3.16; p<0.001), while hand–foot syndrome and stomatitis are more frequent with the pegylated liposomal doxorubicin [30]. TCL-D99 (not pegylated liposomes doxorubicin) is associated with a low rate of cardiac events (13% vs 29% with conventional doxorubicin) [31]. When compared to conventional doxorubicin in a phase III study, a modest advantage in terms of OS was identified for conventional doxorubicin (p=0.009).

The best schedule for taxanes in MBC pre-treated with anthracyclines has longer being studied in many clinical trials.

A recent meta-analysis provided evidence that weekly administration of paclitaxel gives OS advantages compared to three-weekly regimen [32]. On the contrary, the same meta-nalysis failed to show any difference in term of efficacy for the use of docetaxel in weekly or every 3 weeks schedule.

Indeed, a strong dependence between docetaxel dose and anticancer efficacy has been reported: 100mg/m2 of docetaxel given every 3 weeks being superior than 75 or 60mg/m2 three-weekly [33].

Docetaxel (100mg/m2 every 3 weeks) and paclitaxel (175mg/m2 three-weekly) were compared in a randomized study having objective RR as primary endpoint. No difference was detected in primary endpoint, even if docetaxel developed longer TTP and OS but inducing higher haematological and non-haematological toxicity [34].

The ABI-007 is an innovative taxane (paclitaxel nanoparticles bound to albumin, nab-paclitaxel) recently approved for clinical use. Nab-paclitaxel at the dose of 260mg/m2 every 3 weeks has higher RR (33% vs 19%; p=0.001) and TTP (5.7 months vs 4.2 months; p=0.006) compared to three-weekly paclitaxel [35]. The worst toxicity is grade 3 neuropathy; no allergic reaction and lower grade 4 neutropenia were observed during nab-paclitaxel treatment.

Recently, in a phase II study, untreated MBC patients were randomly assigned to receive nab-paclitaxel 300mg/m2 three-weekly vs 100mg/m2 weekly, vs 150mg/m2 weekly vs docetaxel 100mg/m2 every 3 weeks. Although with the limits of a randomized phase II trial, weekly nab-paclitaxel resulted in better RR and PFS than docetaxel after independent radiology review (for nab-paclitaxel at a dose of 150mg/m2/weekly 12.9 months vs 7.5 months; p=0.0065) with less G3-G4 fatigue, neutropenia and febrile neutropenia. Neuropathy was similar among the treatment arms [36].

Conflicting data come from many different studies comparing anthracycline- and taxane-based first-line CT.

The E1193 trial randomly assigned 739 patients with untreated MBC to receive doxorubicin 60mg/m2 vs paclitaxel 175mg/m2 over 24h vs doxorubicin 50mg/m2 plus paclitaxel 150mg/m2 over 24h [22]. Therapy was administered every 3 weeks. In the single-agent arms, crossover was pre-planned at the time of progression. The RR (36% vs 34%; p=0.84) and time to treatment failure (TTF) (5.8 months vs 6 months; p=0.68) were similar for single-agent doxorubicin and paclitaxel. The combination arm showed better RR and TTF but similar OS with single-agent arms [22].

Conversely, in a study run by the EORTC in first-line MBC, single-agent doxorubicin (75mg/m2) demonstrated a clear advantage compared to paclitaxel (200mg/m2) in terms of RR (41% vs 25%; p=0.003) and PFS (7.5 months vs 3.9 months; p<0.001) [37].

A randomized phase III trial randomly assigned 326 MBC patients to receive either docetaxel (100mg/m2) or doxorubicin (75mg/m2) every 3 weeks [38]. The docetaxel arm had a better RR (47.8% vs 33.3%; p=0.008) and a more favourable toxicity profile. There were no differences in TTP (6 months vs 4.8 months) and OS (15 months vs 14 months) between the two arms. Based on the results of this study, docetaxel is now considered to be a valid alternative to anthracyclines in first-line treatment of MBC.

Capecitabine is an orally available pro-drug of 5-fluorouracil (5FU) that needs to be activated into tumour tissue by a multi-step enzymatic process. The last step of this process involves thymidine phosphorylase (TP), an inducible enzyme [39].

There is a growing interest in single-agent capecitabine for first-line MBC progressing after adjuvant anthracycline or taxane-based CT. An Australian phase III study randomly assigned patients with untreated MBC to receive cyclophosphamide – methotrexate – fluorouracil (CMF) or capecitabine 2000mg/m2/day for 14 days every 3 weeks or capecitabine 1300mg/m2/day continuously for 3 weeks [40]. Capecitabine showed a statistically superior OS with an absolute gain of 4 months (22 months vs 18 months; HR 0.72; p=0.02). The toxicity was manageable for capecitabine with less febrile neutropenia, infections and mucositis. No differences in terms of RR, TTP, and OS were evident between the two capecitabine schedules.

Gemcitabine is not recommended as single agent for first-line treatment of MBC. In fact, a phase III study demonstrated the advantage of epirubicin (35mg/m2 on days 1, 8, 15 every 4 weeks) vs gemcitabine (1200mg/m2 on days 1, 8, 15 every 4 weeks) in terms of RR (40.3% vs 16.4%; p=0.0001), TTP (6.1 months vs 3.4 months; p=0.0001) and OS (19.2 months vs 11.8 months; p=0.004). In this study, patients were older than 60 and chemotherapy-naive [41], [42].

Vinorelbine is used both as single agent or in combination with other anticancer drugs. The RR ranges from 35 to 59% when vinorelbine is administered as first-line treatment, and it is about 50% in pre-treated patients [43]. A phase III study demonstrated no difference in PFS (2.9 months vs 2.5 months; HR 1.26; p=0.11) and OS (11 months vs 9 months; HR 1.05; p=0.71) between V and liposomal doxorubicin for patients already treated with anthracyclines or taxanes [44].

Ixabepilone, an analogue of Epotilone B, acts by stabilizing intracellular microtubules and preventing cell division [45]. Recently, it has been approved by FDA for anthracyclines-, taxanes- and capecitabine-resistant MBC on the basis of a phase II trial [46]. Of note, in the first-line setting, ixabepilone at a dose of 40mg/m2 every 3 weeks showed a RR of 41.5% with a median duration of response of 8.2 months and a manageable safety profile (the most common side effect was reversible peripheral neuropathy) [47]. Unlike taxanes, ixabepilone is rarely associated with allergic reactions.

3.2. Combination chemotherapy 

Symptomatic or rapidly progressing disease is more effectively treated with a combination CT, usually anthracycline- or taxane-based.

Patients’ conditions and disease characteristics are of primary importance in the decision-making process. In particular, previous adjuvant CT deeply influences alternative choices for first-line treatment of MBC [48].

Poly-CT with anthracyclines and taxanes leads to a substantial benefit in terms of RR, but superiority in terms of OS was not clearly demonstrated (Table 3) [22], [49], [50], [51], [52], [53], [54], [55].

Table 3. Main studies of poly-chemotherapy with anthracyclines and taxanes.

	Authors, year of publication	N	Treatment	RR (p value)	TTP (p value)	OS (p value)	Outcomes
	Sledge, 2003	739	P 175mg/m2 vs A 60mg/m2 vs AP 50/150mg/m2	34% 36% 47% (0.004, P vs AP) (0.007, A vs AP)	6.0 m 5.8 m 8.0 m (0.009, P vs AP) (0.003, A vs AP)	22.2 m 18.9 m 22.0 m (ns)	AP>P (RR, TTP) AP>A (RR, TTP)
	Biganzoli, 2002	275	AP 60/175mg/m2 vs AC 60/600mg/m2	58% 54% (ns)	6 m 6 m (ns)	20.6 m 20.5 m (ns)	AP=AC
	Jassem, 2001	267	AP 50/220mg/m2 vs FAC 500/50/500mg/m2	68% 55% (0.032)	8.3 m 6.2 m (0.034)	23.3 m 18.3 m (0.013)	AP>FAC (RR, TTP, OS)
	Carmichael, 2001	705	EP 75/200mg/m2 vs EC 75/600mg/m2	67% 56% (<0.05)	6.5 m 6.7 m (ns)	13.7 m 13.8 m (ns)	EP>EC (RR)
	Luck, 2000	429	EP 60/175mg/m2 vs EC 60/600mg/m2	46% 41% (ns)	9 m 7.6 m (ns)	NA	EP=EC
	Nabholtz, 2003	429	AD 50/75mg/m2 vs AC 60/600mg/m2	59% 47% (0.012)	9.3 m 7.9 m (0.014)	22.5 m 21.7 m (ns)	AD>AC (RR, TTP)
	Bontenbal, 2005	216	AD 50/75mg/m2 vs FAC 500/50/500mg/m2	62% 38% (0.001)	8.1 m 6.6 m (0.002)	22.6 m 16.1 m (0.02)	AD>FAC (RR, TTP, OS)
	Mackey, 2002	484	DAC 50/75/500mg/m2 vs FAC 500/50/500mg/m2	55% 44% (0.02)	7.3 m 6.7 m (ns)	21 m 22 m (ns)	DAC>FAC (RR)

D=docetaxel; P=paclitaxel; A=doxorubicin; E=epirubicin; F=5-fluorouracil; G=gemcitabine; ns=not statistically significant; m=months.

In MBC patients pre-treated with anthracyclines, docetaxel associated to capecitabine was proved to be more effective than docetaxel alone in terms of RR (42% vs 30%; p=0.006), TTP (6.1 months vs 4.2 months; HR 0.65; p=0.00019) and OS (14.5 months vs 11.5 months; HR 0.77; p=0.01) [56]. Gastrointestinal side effects and hand–foot syndrome were more common in combination arm, while musculoskeletal disorders and febrile neutropenia were more frequent in the docetaxel arm.

Interestingly, a retrospective analysis of this study demonstrated that capecitabine dose reduction could decrease toxicity without any efficacy loss [57]. These data are consistent with the potential synergistic effect of taxanes and capecitabine: it could be hypothesized that thanks to taxane-induced TP expression, the dose reduction of capecitabine is not associated with a decreased anticancer activity [58], [59].

A phase III study tested the efficacy and tolerability of the association of paclitaxel and gemcitabine, comparing this regimen to paclitaxel 175mg/m2 every 3 weeks. In this trial, the combination provided a benefit in RR (41.4% vs 26.2%; p=0.0002), TTP (6.1 months vs 3.9 months; p=0.0002) and OS (18.6 months vs 15.8 months; HR 0.78; p=0.018) [60]. Moreover, an improvement in symptoms and in quality of life was observed. However, in this study the single-agent arm is probably suboptimal, provided that weekly paclitaxel has been proven to be superior to 3-weekly paclitaxel.

A similar study was presented at ASCO 2009 [61]. In this study, patients with MBC were randomized to receive either gemcitabine plus docetaxel (75mg/m2 every 3 weeks) vs docetaxel (100mg/m2 every 3 weeks). TTP was slightly better with combination CT but these results are preliminary.

The combination of gemcitabine plus docetaxel was also compared with the combination of capecitabine plus docetaxel for anthracycline pre-treated MBC patients. The majority of patients received study treatment as first-line CT and no statistically significant difference was reported in terms of RR, PFS and OS [62]. However, it is important to note that the primary endpoint of the study was PFS and the trial was set out to determine the superiority of the combination G+docetaxel. Accordingly, although the authors reported that the different combinations were similar with regard to PFS, their conclusion is not supported by the statistical design of the study because the sample size should be larger in order to demonstrate similarity [63].

Even if the association of different drugs seems to be more effective, it is not clear if the sequential use of the same drugs, instead of their concomitant administration could have the same final benefit in terms of survival. As a matter of facts, none of the studies reported above (except for E1193 trial) had built-in crossover at disease progression.

As different trials were run to test the efficacy of different poly-CT regimens in MBC treatment, the Cochrane Collaboration performed a series of meta-analyses to clarify the role of different drugs in breast cancer therapy [8].

Recently, Piccart and coll. published a meta-analysis aiming at evaluating the benefit of taxanes in addition to other drugs in first-line treatment [64]. Eleven trials were analysed: 8 with combination therapies and 3 with single agents. Taxane-based combination CT resulted in higher RR and PFS but no convincing benefit was observed in terms of OS.

At present, more and more patients receive taxanes and anthracyclines as adjuvant CT or first-line treatment for MBC. Moreover, above 55% of breast cancers shows a primary resistance to taxanes-based treatments. What is the best choice for these patients?

A phase III study compared capecitabine alone (2500mg/m2/day continuously for 14 days, every 3 weeks) to capecitabine (2000mg/m2/day on days 1 through 14) plus ixabepilone (40mg/m2 q3w) in anthracycline- and taxane-resistant MBC [65]. PFS was significantly longer in the combination arm (5.7 months vs 4.1 months; HR 0.69; p<0.0001). Furthermore, RR was improved in the combination arm (34.7% vs 14.3%; p<0.0001).

Neuropathy was more common in patients who received capecitabine plus ixabepilone but was completely reversible after treatment end. Twelve patients died within 30 days after last drug intake; five of these were affected by moderate-severe liver failure and febrile neutropenia was the end cause of death. Therefore ixabepilone should be used with caution in case of altered liver enzymes.

On the basis of this study, the FDA approved the association of capecitabine plus ixabepilone in the treatment of MBC resistant to taxanes and anthracyclines [66].

3.3. Bevacizumab 

Angiogenesis plays a central role in tumor growth and metastasis development being the result of the balance between pro-angiogenic factors, such as vascular endothelial growth factors (VEGF), and anti-angiogenic factors, such as thrombospondin-1 and angiostatin [67]. A series of randomized trials analysed the combination of chemotherapeutic agents and biologic agents with anti-angiogenic activity. The humanized monoclonal antibody anti-VEGF, bevacizumab, has been recently approved in combination with paclitaxel for the first-line treatment of HER2-negative MBC. The E2100 study was a randomized trial that assigned patients to receive paclitaxel 90mg/m2/weekly for 3 weeks out of 4 with or without bevacizumab 10mg/kg every 14 days. The combination of paclitaxel and bevacizumab resulted in a PFS benefit (11.3 months vs 5.8 months; HR 0.48; p<0.0001) and in higher RR (48.9% vs 22.2%; p<0.0001) [13]. Recently, an independent and blinded review of radiologic and clinical data from the trial was performed [68]. According to previous report, the addition of bevacizumab to paclitaxel resulted in a statistically significant improvement in PFS using the independent review facility (IRF) assessment. Hazard ratios for PFS (0.48; p<0.0001 for the IRF vs 0.42; p<0.0001 for ECOG investigators) and the improvement in median PFS (11.3 months vs 5.8 months for the IRF vs 11.4 months vs 5.8 months for ECOG investigators) were similar. The consistency between the IRF and ECOG analyses validated the original data previously reported by ECOG in this open-label trial.

In terms of toxicity, patients receiving the combination paclitaxel plus bevacizumab experienced about 20% more grade ≥3 events with respect to patients treated with paclitaxel alone. In particular, a greater incidence of neuropathy, hypertension, fatigue, infections, neutropenia, gastrointestinal symptoms, headache, proteinuria and cerebrovascular ischemia was described. Fatal adverse events were reported in six patients (1.7%) receiving bevacizumab and deaths were related to gastrointestinal perforation (two pts), cardiac heart failure (two pts), diarrhea, abdominal pain and hypotension (two pts).

Recently, the updated results of a double-blind, placebo-controlled, randomized trial, also known as AVADO trial, were presented [69]. This study evaluated the association of bevacizumab and docetaxel in first-line treatment of HER2-negative MBC. The trial compared two different doses of bevacizumab (7.5 and 15mg/kg every 3 weeks) in combination with docetaxel 100mg/m2 every 3 weeks vs docetaxel single agent (placebo arm). The study rapidly randomized 736 patients. At a median follow-up of 25 months, the addition of 15mg/kg bevacizumab to paclitaxel lead to a statistically significant benefit in PFS (primary endpoint) with respect to placebo (median PFS: 10.0 months; HR 0.67; p=0.0002). A less pronounced treatment benefit was also observed for the 7.5mg/kg bevacizumab arm (median PFS: 9.0 months; HR 0.80; p=0.0450). Median PFS in the placebo arm was 8.1 months. The abovementioned data are based on a stratified analysis conducted excluding events before progression (i.e. patients were censored if they had started non-protocol therapy before disease progression). In addition, in patients with measurable disease at baseline, bevacizumab resulted in a significant better RR (64.1% vs 46.4% at the dose of 15mg/kg and 55.2% vs 46.4% at the dose of 7.5mg/kg). There was no difference in median OS between the study arms (range 30–32 months), although a superior 1-year survival rate was observed between 15mg/kg arm and placebo arm (84% vs 76%; p=0.02). Concerning the safety profile, grade 3 or 4 adverse events in the bevacizumab arm were mainly due to hypertension, neutropenia and febrile neutropenia. A trend toward lower incidence of venous thromboembolism was observed with bevacizumab treatment. Notably, bleeding events were rare and similar between treatment arms.

Even if the comparison between different studies is not formally correct, the two trials (E2100 and AVADO) have been analysed face-to-face in order to explain the different magnitude of gain in PFS gain observed with the addition of bevacizumab to chemotherapy (5.5 months in E2100 vs 1.9–0.9 months in AVADO trial) as reported in Table 4. Probably, difference in the control arm and in type and duration of the concomitant therapy with bevacizumab are reasonable explanations. The control arm of E2100 showed a smaller RR with respect to the control arm of AVADO (22% vs 46.4%) and a shorter PFS (5.8 months vs 8.1 months). Moreover, therapy with taxane were continued until progression in 80% of patients in E2100 trial while only in 50% of AVADO patients.

Table 4. Characteristics of the three randomized studies that compared chemotherapy alone vs chemotherapy plus bevacizumab in the first-line setting.

		E2100	AVADO	RIBBON 1
	Design	Randomized, open, two-arm trial with chemotherapy±bevacizumab	Randomized, double-blind, placebo-controlled trial, with 3 arms of treatment with chemotherapy±bevacizumab at two dose schedule	Randomized, double-blind, placebo-controlled trial, with chemotherapy±bevacizumab
	Treatment	First-line	First-line	First-line
	Chemotherapy	Paclitaxel (90mg/m2, weekly)	Docetaxel (100mg/m2, q21)	Capecitabine (2000mg/m2 days 1–14 q21), nab-paclitaxel (260mg/m2 q21), docetaxel (75–100mg/m2 q21), anthracycline-based chemotherapy (q21)
	Bevacizumab	10mg/kg q14	7.5 or 15mg/kg q21	15mg/kg q21
	No. patients	722	736	1237 (two cohorts)
	Primary endpoint	PFS	PFS	PFS
	Crossover	Not planned	After disease progression	After disease progression

Recently, RIBBON-1 randomised trial confirmed the significant benefit in terms of PFS when bevacizumab is combined with first-line CT. The combination of capecitabine+bevacizumab is associated with an improvement in PFS (9.7 months vs 6.2 months in the independent review evaluation; HR 0.68, 95 CI 0.54–0.86) as well as the combination of bevacizumab with taxanes or anthracyclines (10.7 months vs 8.3 months; HR 0.76; 95% CI 0.59–0.97). Safety was consistent with previous bevacizumab trials. Mature OS data are still not available [70].

In conclusion, treatment combining bevacizumab and CT represents an important therapeutic option in metastatic HER2-negative patients not previously treated for MBC. In addition, a phase III trial, known as RIBBON-2, showed improved PFS with combination bevacizumab plus standard chemotherapy versus chemotherapy alone for second-line treatment of MBC [71].

3.4. PARP inhibitors 

DNA-repair mechanisms are critically involved in surveillance and maintenance of genome integrity either in normal and neoplastic cells, among them poly(ADP-ribose) polymerase 1 (PARP1) functions as a key molecule in the repair of DNA single-strand breaks. Inhibitors of PARP1 could enhance anticancer activity in BC cells, especially those harboring defects in DNA homologous recombination such is the case of BRCA-1 and BRCA-2 mutated BC. Triple-negative BC is a particular subtype of BC associated with aggressive behavior and often, with BRCA mutation. Recently, the association of gemcitabine and carboplatin with BSI-201, a PARP1 inhibitor, has been shown to be superior to chemotherapy alone in triple-negative BC. Although this is a randomised phase II trial, the preliminary analysis demonstrated that the addition of PARP1 inhibitor to CT was superior in terms of anticancer activity (RR 48% vs 16%), PFS (median 6.9 months vs 3.3 months) and OS (median 9.2 months vs 5.7 months) in this particular phenotype without increasing CT-related toxicity [72].

Olaparib, another oral PARP inhibitor showing efficacy in BRCA-1 and -2 mutation carriers with breast, prostate and ovarian cancer [73], proved to be active even in CT-refractory BRCA mutated MBC in a phase II trial (RR 41%) [74]. This is a further proof of concept study demonstrating the anticancer activity of these agents against a genetically defined target.

4.1. Trastuzumab 

HER2 (human epithelial growth factor receptor 2) is a cell-surface receptor of HER2/ErbB2/Neu family which includes HER1/EGFR, HER3 and HER4. The activation of these proteins depends on their omo- or hetero-dimerization. This process leads to the phosphorylation of different intracellular substrates which activate cell growth signalling. In particular, HER2 triggers phosphatidyl-inositol-3-kinase (PI3K)-AKT and mitogen-activated protein kinase (MAPK) pathways which down regulate cycline D1 and p27 nuclear transcription, inducing cell proliferation and prolonged cell survival as final result.

HER2 protein is overexpressed in 25% of all human BC. This subtype of BC has a more aggressive behaviour with high cell proliferation and propensity to early metastatization [75]. The biological aggressiveness results in bad prognosis, with a reduced estimated life for patients [76].

The standardized methods to evaluate HER2 status include the immunohistochemical analysis of protein expression on cell surface and HER2 gene amplification detection through fluorescent in situ hybridization (FISH) [77].

During the last years, HER2 has become the target for cancer research, having such a strong impact on prognosis. Trastuzumab is a humanized monoclonal antibody against HER2.

It is the first therapy targeting HER2 approved by the FDA [78]. Its employment in clinical practice has drastically changed the prognosis of HER2 amplified BC. A recent review of breast cancer patients survival, conducted at the MD Anderson Cancer Centre of Houston, showed that women with HER2/neu-positive MBC who received trastuzumab had a 44% reduction in the risk of death compared with women with HER2/neu-negative disease (HR 0.56; 95% CI 0.45–0.69; p<0.0001) [16].

As single agent, trastuzumab showed a low anticancer activity, however two different randomized studies demonstrated an advantage in terms of RR, TTP and OS for patients treated with trastuzumab plus chemotherapy versus chemotherapy alone (Table 5) [79], [80].

Table 5. Randomized studies with trastuzumab associated to chemotherapy.

	Slamon (NEJM 2001) (N=469)	AC or P	CT+T	p value
	RR	32%	50%	<0.001
	TTP*	4.6 m	7.4 m	<0.001
	OS	20 m	25 m	0.046

	Marty (JCO 2005) (N=186)	D	CT+T	p value
	RR*	34%	61%	0.001
	TTP	6.1 m	10.7 m	0.001
	OS	23 m	31 m	0.032

(() Primary endpoint; AC=adriamycin, cyclophosphamide; P=paclitaxel; T=trastuzumab; D=docetaxel; m=months.

Several chemotherapeutic agents have a synergic activity with trastuzumab, as demonstrated in pre-clinical assays (taxanes, vinorelbine, capecitabine, platinum compounds, doxorubicin) [81], [82], [83], [84], [85], [86], [87], [88]. Hence, they are studied in association with the anti-HER2 antibody.

However, some HER2-positive BC show an inborn or acquired resistance to anti-HER2 trastuzumab therapy [89]. A common strategy to overcome the resistance to trastuzumab+antiblastic drug is going on with trastuzumab while changing the chemotherapeutic agent [90].

A phase III randomized trial has been conducted on 156 HER2-positive patients progressed to trastuzumab plus CT treatment; patients were randomly assigned to receive capecitabine alone or in association to trastuzumab [91]. The arm in which treatment with monoclonal antibody was maintained showed better outcomes as TTP (8.2 months vs 5.6 months; HR 0.69; p=0.03) and RR (48.1% vs 27%; p=0.001). The short follow-up (15.6 months) makes impossible a final analysis on survival data. Unfortunately, this trial was prematurely closed due to low accrual, mainly because of the approval of lapatinib in combination with capecitabine after trastuzumab progression, and this could be considered a limitation in the final data interpretation. Safety data showed a slight increase in cardiac toxicity in the trastuzumab arm (13% vs 5% all grades), although not statistically significant with the caveat of the small sample size. Indeed, from the very beginning trials, the most clinically relevant adverse event of trastuzumab has been the cardiotoxicity, appearing as left ventricular ejection fraction (LVEF) decrease or symptomatic congestive heart failure. The incidence of heart dysfunction varies from 4 to 7% with trastuzumab alone, to 27% when trastuzumab and anthracyclines are concomitantly administered. Unlike anthracyclines, trastuzumab cardiotoxicity seems at least partially reversible and not associated with ultrastructural changes. The increased cardiac toxicity induced by the association of trastuzumab with anthracyclines may be a limit for treatment [92]. Recently, many phase II studies have analysed the safety of the combination of liposomal doxorubicin and trastuzumab [93]. Although there is growing evidence to potentially reconsidering the taboo about the concomitant use of anthracyclines and trastuzumab [94], this approach is, at the moment, not recommended out of clinical studies.

4.2. Lapatinib 

Lapatinib is an orally available dual kinase inhibitor of EGFR and ERBB2 [95]. Binding the intracellular ATP domain, it stops the signalling of activated EGF and HER2 receptors. Furthermore, lapatinib induces cell apoptosis, inhibits in vitro the intracellular signalling mediated by the insulin growth factor 1 (IGF1) receptor in trastuzumab resistant cancer cells and inhibits the constitutive phosphorylation of truncated HER2 receptor p95erbB2 [95].

Lapatinib employed in clinical practise is approved by the FDA for patients affected by HER2-positive BC pre-treated with trastuzumab, anthracyclines and taxanes. The approval is based on the results of a randomised phase III study [96].

In this study, patients with HER2-positive MBC were randomized to receive either capecitabine (2500mg/m2 every day for 2 weeks out of 3) alone or in association with lapatinib (1250mg daily). The study was early closed after 324 patients because the first ad interim analysis showed a significant reduction in risk of progression (HR 0.49; p<0.001) and a significant increase in TTP (8.4 months vs 4.4 months) for the lapatinib plus capecitabine arm. The final analysis after a longer follow-up, confirmed preliminary results (TTP 6.2 months vs 4.3 months; HR 0.57; p=0.0001) [97].

Moreover, less patients treated with the combination developed brain metastases (4 pts vs 13 pts; p=0.045). This is an interesting preliminary finding that should be confirmed in larger studies. The treatment with lapatinib showed a manageable toxicity profile, even though the association with capecitabine led to greater incidence of diarrhoea and skin rush. No major cardiotoxicity occurred during the combination treatment, even if all the patients received previously trastuzumab and anthracyclines.

As first-line treatment of HER2 unselected patients with MBC, the association of lapatinib with 3-weekly paclitaxel was not associated to any benefit over paclitaxel alone, however in a post-hoc analysis the subgroup of HER2-positive patients experienced a benefit in terms of TTP, overall response rate and clinical benefit. This observation needs to be confirmed in pre-planned targeted trials [98], [99].

Having a different target on HER2 receptor, a synergistic activity for lapatinib plus trastuzumab has been hypothesized. Pre-clinical studies on breast cancer cells overexpressing HER2 demonstrated in vitro the synergy of the two drugs [100]. These data were confirmed in a phase III study conducted on patients pre-treated with different lines of trastuzumab-based chemotherapy (median 4–5 different regimens). Trastuzumab and lapatinib association turned out to be better than lapatinib alone both in clinical benefit (24.7% vs 12.4%; p=0.01), PFS (12 weeks vs 8.1 weeks; p=0.0008) and OS (14.0 months vs 9.5 months; HR 0.74; p=0.026) [101], [102]. Notably, the significant survival benefit was obtained despite 52% of patients randomized to lapatinib monotherapy crossed over to lapatinib plus trastuzumab.

The role of lapatinib and the combination of trastuzumab plus lapatinib deserve to be further investigated as first-line treatment of HER2-positive MBC.