Cancer Treatment Reviews xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Cancer Treatment Reviews journal homepage: www.elsevierhealth.com/journals/ctrv
Anti-Tumour Treatment
A systematic review of bevacizumab efficacy in breast cancer Iben Kümler ⇑, Ole Grummedal Christiansen, Dorte Lisbet Nielsen 1 Department of Oncology, Herlev Hospital, University of Copenhagen, Herlev Ringvej 75, DK-2730 Herlev, Denmark
a r t i c l e
i n f o
Article history: Received 28 February 2014 Received in revised form 8 May 2014 Accepted 12 May 2014 Available online xxxx Keywords: Angiogenesis Bevacizumab Breast cancer VEGF
a b s t r a c t Angiogenesis is a key component of cancer growth, invasion and metastasis. Therefore, inhibition of angiogenesis is an attractive strategy for the treatment of cancer. We systematically describe phase II and III clinical trials of bevacizumab for the treatment of breast cancer. Methods: A computer-based literature search was carried out using PUBMED and conference databases. Original phase II and III studies reporting P15 patients who received bevacizumab were included. Results: 41 phase II trials were identified in the metastatic setting. Most trials found bevacizumab treatment feasible. Response rates (RR) varied from 0% to 76.5%, time to progression (TTP)/progression free survival (PFS) from 2.4 to 25.3 months and overall survival from 11.5 to more than 38 months. 14 phase III trials including more than 4400 patients with MBC unanimously showed increased RR and PFS, however, no trials demonstrated an OS benefit. In the neoadjuvant setting 23 phase II and III trials were identified. All studies found increased pCR/tpCR but no benefit in terms of OS could be demonstrated. The only study conducted in the adjuvant setting failed to show any survival benefit of bevacizumab. Conclusion: Despite increased response rates in both the metastatic and neoadjuvant setting, bevacizumab has failed to show any OS benefit. Future trials should include identification of robust predictive biomarkers in order to improve our understanding of molecular biomarkers and mechanisms. Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
Methods
Neovascularization is one of the hallmarks of tumor invasion and metastasis and is a prerequisite for progression of solid tumors [1]. For this reason, the inhibition of angiogenesis is regarded an attractive therapeutic approach. Bevacizumab is a recombinant humanized monoclonal antibody that binds to the VEGF-A ligand and prevents it from binding to its receptors [2,3]. Few drugs have generated as much discussion in both the medical literature and the popular press as bevacizumab for metastatic breast cancer (MBC). The decision in 2011 of the US Food and Drug Administration (FDA) to withdraw full approval for the drug in first-line MBC after having granted accelerated approval 3 years earlier renewed the debate not only about the approval process, but also about how to measure clinical outcome and how to evaluate risk benefit of a new drug. This review focusses on efficacy of bevacizumab in phase II and III trials in the (neo)adjuvant and the metastatic setting.
We systematically searched PUBMED for phase II and III studies using the following search terms: bevacizumab AND breast cancer with the clinical trial filter activated. We repeated the search without the clinical trial filter, using the same search terms AND (neo)adjuvant/metastatic breast cancer to ensure that all relevant articles were retrieved. Full articles were obtained and references were checked for additional material when appropriate. In addition, abstracts from 2009 to 2013 annual meetings of the American Society of Clinical Oncology (ASCO) and San Antonio Breast Cancer Symposium 2009–2013 were retrieved for relevant abstracts using similar search terms. The reference list was updated in December 2013 (Fig. 1). The following criteria were applied: Studies reporting P15 evaluable patients, studies reporting efficacy parameters of at least response rate (RR) or time to progression (TTP)/progression free survival (PFS) in the metastatic setting, pathological response (pCR) (breast tumor/breast tumor and lymph nodes) in the neoadjuvant or disease free survival (DFS) in the adjuvant setting and papers in English were included. Two authors independently surveyed the literature ((neo)adjuvant: OGC and DLN; MBC: IB and DN). In case of unclarity or
⇑ Corresponding author. Tel.: +45 38689598; fax: +45 44883094. E-mail addresses: [email protected] (I. Kümler), olegrummedal@gmail. com (O.G. Christiansen), [email protected] (D.L. Nielsen). 1 Tel.: +45 38682344; fax: +45 44883094. http://dx.doi.org/10.1016/j.ctrv.2014.05.006 0305-7372/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
2
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
phase III trials evaluating metronomic chemotherapy and bevacizumab in both early and MBC (NCT00925652, NCT01131195, NCT01112826) [32,33].
Search PUBMED (- Dec. 2013): 98 references
48 articles included 50 excluded - Pre-clinical: 1 -Non- BC population: 7 -Non-clinical: 5 -Biomarker: 8 -Phase I: 6 -Non English language: 3 -Safety: 11 -Other: 9
Repeated search + references from other reviews
8 articles
ASCO and SABC (2009-2013)
24 abstracts
Fig. 1. Diagram of included articles.
disagreement, the complete paper was analyzed and a verdict was reached by consensus. Results Our search identified 22 phase II and 2 phase III trials in the neoadjuvant setting and 1 phase III trial in the adjuvant setting. One reference was retrieved by searching in prior reviews. In the metastatic setting 41 phase II and 14 phase III trials were retrieved. Bevacizumab in predominantly human epidermal growth factor receptor 2 (HER2)-negative MBC Phase II studies Several combinations of bevacizumab and chemotherapy have been investigated in the metastatic setting both as first-line treatment as well as in subsequent treatment-lines. Phase II trials are summarized in Table 1. Taxanes are the drugs most often studied in combination with bevacizumab. Two trials of bevacizumab in combination with docetaxel demonstrated RR of approximately 50% with median PFS/TTP of 7.5 and 9.3 months, respectively [4,5]. Addition of capecitabine in two other studies produced similar results with RR of 49% and 61%, respectively, and TTP of 11 months in both studies [6,7]. Furthermore, several studies have reported on combinations of taxanes and various other drugs, all showing RRs and PFS/TTP in the same range [8–24]. Studies not including taxanes have shown modest activity although there are only few trials [25,26]. A trial of pegylated liposomal doxorubicin and bevacizumab as first-line treatment in locally recurrent and MBC was stopped prematurely due to toxicity after the enrolment of 43 patients [27]. A study of bevacizumab in combination with vinorelbine and capecitabine in recurrent inflammatory breast cancer showed RR of 46.4% [28]. Preclinical studies have suggested that metronomic chemotherapy has anti-angiogenic properties [29]. Bevacizumab in combination with metronomic chemotherapy consisting of cyclophosphamide and capecitabine or methotrexate demonstrated clinical benefit (CRB) rates of 64–68% with PFS of 8–9 months in patients who had previously received anthracyclines and taxanes [30,31]. A randomized trial of bevacizumab in combination with metronomic chemotherapy demonstrated a near three-fold increase in RR (10% versus (vs.) 29%) and TTP (2.0 vs. 5.5 months). The results are preliminary and from 2005. No updates or final publication have been identified, nevertheless, these results have prompted the initiation of several
Phase III Results of phase III trials are given in Table 2. The pivotal openlabel E2100 trial randomized 722 patients (90% HER2-negative) to paclitaxel +/ bevacizumab as first-line therapy. An increased RR (21% vs. 37%, P < 0.001) and a doubling in median PFS (5.9 vs. 11.8 months, P < 0.001) were demonstrated. The median OS, however, was similar in the two groups (25.2 vs. 26.7 months, P = 0.16) [34,35]. Unfortunately, data on treatment administered after progression were not collected, precluding an exploratory analysis of the influence of cross over between treatment arms and subsequent therapy on OS. Results from the placebo-controlled AVADO trial including 736 patients randomized to first-line treatment with docetaxel + placebo, docetaxel + bevacizumab (7.5 mg/kg) or docetaxel + bevacizumab (15 mg/kg) revealed a modest yet statistically significant increased RR (46.4% vs. 55.2% vs. 64.1%) and PFS 8.2 vs. 9.0 vs. 10.1 months, though seemingly to a lesser extent than the bevacizumab plus paclitaxel combination. The median OS was similar in the three groups [36]. In the initial registration study AVF2119g the addition of bevacizumab to capecitabine in patients with advanced, pretreated MBC was rather disappointing. Despite a doubling of the RR in the combination arm (19.8% vs. 9.1%) PFS did not improve (4.8 vs. 4.2 months) nor did OS (15.1 vs. 14.5 months) [37]. Because of these results, the FDA required a third arm including capecitabine in the following RIBBON-1 trial. The RIBBON-1 trial evaluated first-line chemotherapy (taxane, capecitabine or anthracycline-based) +/ bevacizumab. Among 1237 patients the addition of bevacizumab resulted in a significant improvement in RR and PFS. A prespecified subgroup analysis showed that the efficacy of bevacizumab was independent of prior therapy and hormone-receptor (HR) status. However, no significant OS advantage was demonstrated [38]. A pooled subgroup analysis of 2447 individual patients included in E2100, AVADO and RIBBON-1 was undertaken. The analysis showed a PFS of 9.2 months with bevacizumab and 6.7 months without (Hazard ratio (HR) 0.64; 95% confidence interval (CI) 0.57–0.71). The one-year median OS rate was increased (71% vs. 65%). However, OS was not different (26.7 vs. 26.4 months) (HR 0.97; 95% CI 0.86–1.08). The magnitude of benefit was similar in all patient subgroup irrespective of baseline characteristics [39]. The RIBBON-2 trial assessed the efficacy of bevacizumab plus non-anthracycline-based, second-line chemotherapy. The addition of bevacizumab significantly improved PFS. Nevertheless, OS was not significantly different [40]. Of note, an exploratory subgroup analysis of 159 (23%) triple-negative breast cancer (TNBC) patients demonstrated significant improvements in RR (41% vs. 18%; P = 0.0078), PFS (6.0 vs. 2.7 months; (HR 0.494; 95% CI 0.33–0.74; P = 0.0006) and median OS (17.9 vs. 12.6 months; HR 0.624, 95% CI 0.39–1.007; P = 0.0534). Despite the small sample size, there was a trend toward improved OS among this subset of patients [41]. The TURANDOT trial evaluated the combination of bevacizumab plus paclitaxel vs. bevacizumab plus capecitabine. Results from an interim OS analysis after inclusion of 564 patients showed that the non-inferiority criterion was not met and OS results were inconclusive, although PFS was better and more patients experienced response in the paclitaxel arm [42]. Final results are expected in 2014. A one year survival of 78% was reported among 63 TNBC patients receiving bevacizumab plus paclitaxel [43]. A randomized study of bevacizumab plus taxane +/ capecitabine was stopped after the first interim analysis showing no PFS benefit but
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Reference
Therapy
Setting*
Number of patients
Primary end point
RR (95% CI)%
Median TTP (95% CI) months
Median OS (95% CI) months
Hurvitz [4]
B + docetaxel (TORI B01)
1st line
76 (67 evaluable)
3
51 (38.7–62.7)
9.3 (8.2–12.4)
Ramaswamy [5]
B + docetaxel
27
RR
52 (32–71)
7.5 (6.2–8.3) (PFS)
Perez [6]
B + docetaxel + capecitabine
45
RR
49 (34–64)
11.1 (PFS) (8.0–14.3)
Bisagni [7]
B (to PD) + docetaxel + capecitabine (maximum 6 cycles) B + paclitaxel
HER2-negative (93%) 1st + 2nd line HER2-negative (73%) Unknown (27%) 1st line HER2-negative (91%) 1st line 1st line
26.3 (20.8 – not reached NR
79
PFS
61
11 (8.2–13.8)
120
PFS
74 (64.5–81.2)
12.9 (11.1–18.2)
Brufsky [9]
B + paclitaxel B + paclitaxel + gemcitabine
HER2-negative (>90%) 1st line
94 93
RR
48.9 (38.5–59.5) 58.7 (47.9–68.9) (P = 0.117)
8.8 (8.1–10.4) 11.3 (9.7–12.7) (PFS) (HR 0.82; P = 0.247)
Conlin [10]
B + nab-paclitaxel q3w B + nab-paclitaxel q2w B + nab-paclitaxel weekly B + paclitaxel + gemcitabine B + paclitaxel B + paclitaxel
HER2-negative 1st line
72 74 76 21 (17 evaluable) 39 43
RR
42 42 41 88 (CBR) 58.8 58
7.7 6.4 7.7 NR 11 NR
154 156 32
PFS PFS
67 50 66.7 (47.2–82.7)
11 8.4 (HR 0.52; P < 0.001) 13.6 (11.2–21.9)
24.1 23.1 (P = 0.44) 26.8 (13.3–41.2)
Aogi [8]
Guardino [11] Inno [12] Kountourakis [13] Lam [14]
1st line 1st line HER2-negative (93%) 1st line 1st line
TTP NR NR
35.8 (26.4 – not estimated) 25.0 (18.8 – NA) 24.3 (20.3 – NA) (HR 0.84; P = 0.475) NR
NR 33 NR
Lo [15]
B + paclitaxel + capecitabine B + paclitaxel B + nab-paclitaxel + carboplatin
Lobo [16]
B + nab-paclitaxel + gemcitabine
1st line
30
PFS
75.9
10.4 (5.6–15.2) (PFS)
Mavroudis [17] Northfelt [18]
B + docetaxel + epirubicin B + nab-paclitaxel + gemcitabine
64 (60 evaluable) 48
RR PFS
64 69 (54–81)
14 (estimated PFS) PFS 11.7 (8.9–13.1)
Polyzos [19]
B + paclitaxel
40
RR
30 (15.8–44.2)
4.8
13
Rugo [20]
B + ixabepilone (weekly) B + ixabepilone (q3w) B + paclitaxel
1st line HER2-negative (96%) 1st line 2nd line (30%) P 3rd line (70%) 1st line
18 months OS 77.2% (51.1–90.5) NR Not reached
46 45 32
RR
48 (32.9–63.1) 71 (55.7–83.6) 63 (43.7–78.9)
9.6 (6.1–11.7) 11.9 (8.7–14.7) 13.5 (10.0–18.2)
Saloustros [21]
B + paclitaxel + carboplatin
45 (38 evaluable)
RR
76
9.2
91% 89% 91% (1-year OS rates) NR
Salvador [22] Seidman [23]
B + paclitaxel + gemcitabine B + nab-paclitaxel (3 dosing schedules)
PFS RR Toxicity
B + docetaxel + cisplatin
RR
72 45 41 46 80
11.5 (9.0–17.6) 21.3 (19.5–26.6) 19.0 (15.0–28.1) 25.3 (18.2–31.0) 6.2 (3.2–9.1 (PFS)
27.4 (21.9-NA) NR
Tai [24]
82 (76 evaluable) 75 54 79 20
Borson [25]
B + gemcitabine
52
PFS
21.4 (10.3–36.8)
4.8 (3.4–7.6)
Gajria [26] Rochlitz [27] Fumagalli [28]
B + capecitabine (7 days–7 days of) B + pegylated liposomal doxorubicin B + vinorelbine + capecitabine (concurrent) B + vinorelbine + capecitabine (sequential) B + capecitabine + cyclophosphamide
41 43 46 (distribution on regimen NR) 46
RR Toxicity PFS RR
20 (9–36) 21 (10–36) 46.4 11.1 (P = 0.025) 48 (33–63)
8 (6–14) 5.7 (4.6–8.1) (PFS) 4.7 4.3 (P = 0.69) 9.3 (5.8–16)
21 (13.7 – not estimable) NR 15.9 (14.0–21.5) NR
24
RR
31.8 (13.9–54.9)
7.5
Dellapasqua [30] Garcia-Sáenz [31]
B + cyclophosphamide + methotrexate + (trastuzumab)
HER2-negative (97%) 1st line
Triple-negative 1st line 1st line HER2-status NR 1st line 1–3 prior regimens (including adjuvant therapy 1st line (prior taxane-containing (neo)adjuvent therapy) 1st (88%) and 2nd line 1st line HER2-status NR Recurrent inflammatory cancer HER2-negative (98%) 1st–3rd line HER2-positive included taxane and anthracycline refractory#
11.5 (8.7–15.1
NR 13.6
3
B, bevacizumab; CBR, clinical benefit rate; CI, confidence interval; NA not assessable; NR, not reported; RR, response rate; PFS, progression free survival; TTP, time to progression. When not otherwise stated all patients had HER2-negative breast cancer. # Only abstract available, number of HER2-positive patients not reported.
*
28.4 (20.2 – not reached) 32 (26.0–37.9)
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Table 1 Phase II trials of bevacizumab in combination with chemotherapy in patients with predominantly HER2-negative metastatic breast cancer. Studies with P15 evaluable patients are included.
4
Reference
Therapy
Setting*
Number of patients
End point
RR (95% CI) %
Median TTP (95% CI) months
Median OS (95% CI) months
Miller [34]
B + paclitaxel
HER2-negative (>95%) 1st line (E2100)
365 (330 evaluable for response) 350 (316 evaluable for response) 248 247 241 232 230
PFS
36.9 21.2 P < 0.001)
11.8 (PFS) 5.9 (P < 0.001)
26.7 25.2 (P = 0.16)
PFS
55.2 (P = 0.07) 64.1 (P < 0.001) 46.4 19.8 (14.7–25.0) 9.1 (5.4–12.9) (P = .001)
9.0 (HR 0.86 P = 0.12) 10.1 (HR 0.77; P = 0.006) 8.2 4.9 (PFS) 4.2 (HR 0.98; P = 0.857)
30.8 (HR 1.05) 30.2 (HR 1.03) 31.9 15.1 14.5
35.4 23.6 (P = 0.0097) 51.3
8.6 (PFS) 5.7 (HR 0.69; P = 0.0011) 9.2
HR 0.85 (P = 0.27) HR 1.03 (P = 0.83)
37.9 (P = 0.0054)
8.0 (HR 0.64; P < 0.001)
39.5 29.6 (P = 0.0193) 44 27
7.2 5.1 (HR 0.78; P = 0.0072) 11 (10.4–12.9) 8.1 (7.1–9.2) (HR 1.36; P = 0.0052) (HR 1.06)
Paclitaxel Miles [36]
Miller [37]
Robert [38]
Brufsky [40] Lang [42]
Lueck [44]
B (7.5 mg/kg q3w) + docetaxel B (15 mg/kg q3w) + docetaxel Placebo + docetaxel B + capecitabine Capecitabine
1st line (AVADO)
HER2-negative (approximately 75–80%) 1–2 prior chemoterapeutic regimens for MBC (previous anthracycline and taxane) (AVF2119g) HER2-negative + undetermined 1st line (RIBBON-1)
409 206 415
PFS
PFS
B + capecitabine Capecitabine B + taxane or anthracycline-based therapy Taxane or anthracycline-based therapy B + chemotherapy Placebo + chemotherapy B + paclitaxel B + capecitabine
2nd line (RIBBON 2) 1st line (TURANDOT)
459 225 268 265
PFS
B + taxane + capecitabine B + taxane
1st line (TABEA)
202
PFS
207
Non-inferior OS
NR
18.0 16.4 (HR 0.90; P = 0.3741) (HR 1.04; P = 0.059)
NR
B, bevacizumab; CI, confidence interval; HR, hazard ratio; NR, not reported; NS, not significant; RR, response rate; PFS, progression free survival; q3w: every third week; TTP, time to progression. When not otherwise stated all patients had HER2-negative breast cancer.
*
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Table 2 Phase III trials of bevacizumab in combination with chemotherapy in patients with predominantly HER2-negative metastatic breast cancer.
5
PFS 61, 24 received maintenance therapy with B + endocrine therapy ER-positive (85%) 1st line B + taxane ? B + endocrine therapy Militello [53]
II
189
AI, aromatase inhibitor; B, bevacizumab; CI, confidence interval; ER, estrogene receptor; HR, hazard ratio; NR, not reported; RR, response rate; PFS, progression free survival; TTP, time to progression.
36 (22–51)
42 (HR 1.18; P = 0.469)
NR PFS 191 HER2-negative 1st line B + endocrine therapy (letrozole/ fulvestrant) Endocrine therapy Martin [52]
III (LEA)
B + fulvestrant or B + anastrozole (5 HER2-positive patients received trastuzumab) Yardley [51]
II
3.8 (HR 0.83; P = 0.14) 13.5 (10.2–18.2)
41
All: 1 year OS 78%, 2 year OS 61%
9 (5–13) 21 (14 – not reached) All: 13.5 (8.8–19.1) (PFS) 18.4 27 47 PFS
9 (3–22) Safety 43 B + letrozole Traina [50]
II
HER2-positive (5%) Prior endocrine therapy (adjuvant + for metastatic disease) (93%) Relapse during or 612 months after adjuvant AI
41 38
NR
26.9 (12.5–36.2)
6.2 (3.6–10.1) (PFS) 6-months PFS rate: 39 (23–58) 17.1 (8.5–26.2) (PFS) 22 (6–22) (18 evaluable) 6-months PFS rate 33 Prior AI; up to 2 prior endocrine therapies II
Median OS (95% CI) months Median TTP (95% CI) months RR (95% CI) % Primary end point Number of patients Setting
B + fulvestrant
Preclinical research has shown significantly crosstalk between HER2 signaling and angiogenesis [54]. Clinical evidence has supported this correlation. In 611 BC tissue specimens, overexpression of HER2 was significantly associated with expression of VEGF [55]. Thus, combination of inhibitors against HER2 and VEGF might be a rational strategy. Results of trials evaluating bevacizumab in HER2positive MBC are given in Table 4. Preliminary data from a phase II trial on trastuzumab + bevacizumab were promising with a 48% RR and a PFS of 9.2 months in the first-line setting. However, no other studies have investigated this chemo-free option, and despite the abstract being published in 2009, no final publication is available [56]. Two small phase II trials combining bevacizumab and trastuzumab with capecitabine or docetaxel as first-line treatment in MBC demonstrated RR of 73% and 81%, respectively [57,58]. A phase II study of the combination of bevacizumab, trastuzumab and vinorelbine however, was deemed unlikely to reach the predefined threshold for declaring success [59].
Tan [49]
Bevacizumab in HER2-positive MBC
Phase
Table 3 summarizes the results from trials on bevacizumab in combination with antihormonal therapy in MBC. Preclinical studies have shown multiple points of overlap between the estrogen (ER) and angiogenetic pathways [48] and it has been hypothesized that adding bevacizumab to hormonal therapy could result in reversal of required endocrine resistance. Trials combining bevacizumab and antihormonal therapy have shown diverse results with RR ranging from 9% in a study of letrozole and bevacizumab to 47% in a first-line study of anastrozole and bevacizumab and TTP of 17.1 and 21 months, respectively. Two studies with fulvestrant reported RRs of 22% and 27%, respectively with PFS as low as 6.2 months. Patients populations have been heterogenous, making direct comparisons difficult [49–51]. The phase III, LEA trial investigated the addition of bevacizumab to endocrine therapy with letrozole (approximately 90%) or fulvestrant in the first-line setting. The trial included 380 patients among whom approximately 50% had received adjuvant endocrine therapy [52]. PFS in the bevacizumab arm was 18.4 months vs. 13.8 in the endocrine monotherapy arm with no significant difference. Finally, Militello et al. investigated continued treatment with bevacizumab plus endocrine therapy in ER-positive patients who had RR or SD P6 months after receiving bevacizumab + taxane. A PFS of 13.5 months and an OS of 36 months was registered among 24 patients, prompting the authors to conclude that maintenance therapy with bevacizumab and endocrine therapy can extent the overall benefit of therapy [53]. No randomized studies evaluating this treatment strategy were identified.
Therapy
Bevacizumab in hormone receptor-positive breast cancer
Reference
significantly more toxicity including six toxic death in the capecitabine arm [44]. Finally, the non-randomized phase IV ATHENA trial further assessed first-line bevacizumab combined with single agent or combination chemotherapy according to physician’s standard of care. Eligible patients had HER2-negative locally recurrent or MBC (or HER2-positive disease progressing after trastuzumabcontaining therapy). A total of 2251 patients were included and efficacy and safety data were similar to that observed in the randomized trials [45,46]. Among 585 TNBC patients the RR was 49% including 10% complete responses (CR), TTP was 7.2 months (95% CI 6.6–7.8 months) and median OS 18.3 months (95% CI 16.4–19.7) [47]. Thus, although these trials demonstrated significant improvements in RR and TTP/PFS, findings have not indicated substantial benefit in terms of survival.
Table 3 Phase II and III trials of bevacizumab in combination with antihormonal therapy in patients with etrogen and/or progesterone receptor-positive metastatic breast cancer. Studies with P15 evaluable patients are included.
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Reference
Therapy
HER2 targeted therapy and/or chemotherapy Hurvitz [56] B + trastuzumab Martin [57] B + trastuzumab + capecitabine
6
Phase
Setting
Number of patients
Primary end point
RR (95% CI) %
Median PFS (95% CI) months
Median OS (95% CI) months
II II
1st line Prior HER2-targeted (neo)adjuvant therapy (25%) 1st line HER2-negative HER2-positive 1st line 1st or 2nd line
50 88
NR RR
48 73 (62–82)
9.2 (5.4–20.5) 14.4 (10.4 – not reached)
NR NR
72 20
PFS
57.7 81.0
8.5 (5.3–10.8) 13.4 (12.1 – NA)
NR
22 7
PFS at 1 year
73 71
9.9 7.8 (PFS at 1 year 36%/ 29%) 16.5 13.7 (HR 0.82; 0.65–1.02, P = 0.0775) 12.2 11.1 (HR 0.65; P = 0.10)
NR
Schwartzberg [58]
B + doxetacel B + trastuzumab + docetaxel
II
Lin [59]
B + trastuzumab + vinorelbine
II
Gianni [60]
B + trastuzumab + docetaxel Trastuzumab + docetaxel
III
1st line
215 206
PFS
76.5 65.9 (P = 0.0265)
Arteaga [61]
B + paclitaxel +/ carboplatin + trastuzumab Paclitaxel +/ carboplatin + trastuzumab
III
1st line
48 48 (terminated due to poor accrual)
PFS
52 52
Lin [62]
B + carboplatin + (trastuzumab)
II
38
Composite RR, brain RECIST
63 45
NR
Rugo [63]
B + lapatinib
II
HER2-positive (79%) P2 lines Brain metastases Prior trastuzumab; Median 3 prior chemotherapeutic regimens
>38 >38 (HR 1.01; o0.74– 1.38; P = 0.9543 14 12 (HR 1.23; P = 0.60) 12 NR
52 (45 evaluable)
PFS at week 12
13.3 (5.1– 26.8)
5.5 (4.5–7.8)
NR
Agents targeting angiogenesis or HER1 Montagna B + erlotinib + capecitabine + cyclophosphamide [64] Dickler [65] B + erlotinib
II
HER2-negative 1st line
26 (24 evaluable)
Safety, RR
62
9.6 (4.7–15.3
II
37
RR
3 (0–8)
2.4 (1.7–4.0)
Mina [66]
B + sorafenib
II
18
PFS
0
2.8 (1.8–6.1)
Robert [67]
B + paclitaxel Sunitinib + paclitaxel
III
HER2-positive (21%) 2nd or 3rd line Triple-negative (50%) 2nd or 3rd line 1st line
243 242
RR
32.1 32.2
Martin [68]
B + paclitaxel Motesanib + paclitaxel Placebo + paclitaxel B + paclitaxel + sunitinib B + paclitaxel
II
HER2-negative 1st line
RR
II, randomized (SABRE-B) III
HER2-negative 1st line
97 91 94 23 (17 evaluable) 23 (18 evaluable)
52 (41–62) 49 (39–60) 41 (31–52) 71 61
9.2 7.4 (HR 1.82; P = 0.996) 11.5 (9.3–15.9) 9.5 (8.4–12.1) 9.0 (7.4–12.4) NR
HER2-negative 1st line
55 57
PFS
29 16
III
HER2-negative 1st line
56 57 58 57
PFS
NR
Mayer [69]
Yardley [70]
B + paclitaxel + everolimus B + paclitaxel + placebo
Dieras [71]
B + paclitaxel + AMG 386 (10 mg/kg) B + paclitaxel + AMG 386 (3 mg/kg) B + paclitaxel + placebo AMG 386
PFS
8.8 (7.4–9.6) 7.1 (5.5–9.1) HR (0.94; P = 0.79) 11.3 9.2 12.2 10.0 (NS)
Mina [71]
NR
NR
AI, aromatase inhibitor; B, bevacizumab; CI, confidence interval; ER, estrogene receptor-positive; HR, hazard ratio; NS, not significant; NR, not reported; RR, response rate; PFS, progression free survival; TTP, time to progression.
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Table 4 Phase II and III trials of bevacizumab in combination with HER2 targeted therapy, agents targeting HER1 or angiogenesis. Studies with P15 evaluable patients are included.
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Likewise, data from the randomized, phase III AVEREL study showed no significant improvement in PFS after adding bevacizumab to trastuzumab and docetaxel in the first-line setting [60]. ECOG 1105 was a randomized, phase III study evaluating trastuzumab + paclitaxel + carboplatin +/ bevacizumab carboplatin in the first-line setting. The study was terminated early due to poor recruitment [61]. Finally, a study of bevacizumab and carboplatin in patients with brain metastases, among which 79% were HER2-positive reported a RR of 45% [62]. Several studies combining bevacizumab and trastuzumab with a chemotherapeuticum are currently ongoing [54]. Conclusively, bevacizumab in the HER2-positive setting have shown minimal, if any clinical benefit. Furthermore, the combination required stringent cardiac surveillance. A phase II study combining bevacizumab and the dual tyrosine kinase inhibitor of HER1 and HER2, lapatinib in 52 patients with HER2-positive MBC (median 5 prior MBC therapies) showed an RR of 13% and a CBR of 31%. The combination was generally well tolerated [63]. Combined therapy of bevacizumab and agents targeting HER1 HER2-negative BC Erlotinib is an inhibitor of the HER1 (EGFR) tyrosine kinase. A phase II study including 26 patients with HER2-negative, advanced BC receiving first-line metronomic chemotherapy (capecitabine + cyclophosphamide) + bevacizumab + erlotinib showed a relatively high RR of 62%, a PFS of 9.6 months, and an OS of 24 months. Hovewer, 55% of the patients were chemotherapy naïve making the result less impressive [64] (Table 4). Mixed population A phase II trial of bevacizumab in combination with erlotinib demonstrated limited activity of the combination [65] (Table 4). Combined therapy of bevacizumab and agents targeting angiogenesis Angiogenesis is a complex process comprised of multiple signaling pathways. Many of these pathways are redundant, with several ligand-receptor combinations resulting in the same eventual down-stream of events. It is unlikely that tumors are entirely dependent on only one abnormally activated signaling pathway; consequently, treatment with an agent that interferes with a single target may be insufficient. TNBC A small phase II study of bevacizumab in combination with sorafenib showed minimal efficacy and substantial toxicity [66]. HER2-negative BC A randomized study comparing paclitaxel plus bevacizumab +/ the tyrosine kinase inhibitor (TKI) sunitinib included 485 patients. A high frequency of grade 3–4 neutropenia was found in the sunitinib arm precluding the delivery of prescribed doses. The sunitinib regimen was found to be clinically inferior to the bevacizumab regimen [67]. In addition, paclitaxel in combination with bevacizumab, placebo or the TKI motesanib, was evaluated in a randomized study including 282 patients. The study showed no difference in PFS of the three regimens. More patients in the motesanib group had serious adverse events especially gastrointestinal toxicity [68]. A randomized phase II study (SABRE-B) of paclitaxel + bevacizumab vs. paclitaxel + bevacizumab + sunitinib as first-line therapy in HER2-negative MBC reported a high rate of dose
7
modifications and/or discontinuations of sunitinib due to toxicity. The study was closed due to non-feasibility [69]. Additionally, a phase III study of first-line bevacizumab and paclitaxel +/ everolimus demonstrated increased RR in the everolimus arm, however, PFS was similar in the two arms [70]. Finally, a randomized study of paclitaxel + bevacizumab +/ different doses of AMG 386, a fusion protein that inhibits angiogenesis, showed no benefit of the combination [71] (Table 4). Bevacizumab in early-stage breast cancer, preoperative therapy Results of phase II and III trials are given in Table 5. TNBC Bevacizumab combined with docetaxel + carboplatin showed a pCR rate of 44% among 45 TNBC patients [72]. Preliminary results from a study of bevacizumab + nab-paclitaxel + carboplatin followed by doxorubicin + cyclophosphamide in this context demonstrated similar results [73]. Furthermore, preliminary results from a phase II evaluating the combination of bevacizumab plus cisplatin in this setting has shown some activity. However, toxicity including hearing loss, hypertension and pulmonary embolism limited the feasibility of this treatment [74]. Preliminary results from a phase II study evaluating single dose bevacizumab followed by bevacizumab + dose-dense doxorubicin + cyclophosphamide followed by paclitaxel showed a pCR rate of 44% in TNBC as compared to 8% in HR-positive BC [75]. The finding of a higher pCR rate in TNBC was confirmed by Sánchez-Rovira et al. investigating the effect of doxorubicin + cyclophosphamide followed by bevacizumab + docetaxel [76]. In addition, a phase II study of bevacizumab, nab-paclitaxel and carboplatin showed 27% tpCR rate (pCR in breast and axilla) in TNBC as compared to 11% in HR-positive BC whereas subsequent dose-dense docetaxel + cyclophosphamide raised the tpCR rate to 81% and 54% in the two subgroups, respectively [77]. Another small study of 49 HER2-negative patients treated with fluorouracil, epirubicin and cyclophosphamide (FEC) followed by bevacizumab and paclitaxel showed cCR of 21% total, but in 15 patients with TNBC the pCR was 47% [78]. HER2-negative BC Two phase II studies in HER2-negative patients investigating bevacizumab + doxorubicin + cyclophosphamide followed by docetaxel + capecitabine and the drug in combination with docetaxel + capecitabine showed somewhat lower pCR rates of 9% and 22%, respectively [79,80]. Preliminary results from two phase II trials investigating bevacizumab in combination with FEC followed by docetaxel or paclitaxel, respectively in patients with HER2-negative stage III or inflammatory BC showed a pCR of 25–27% [81,82]. In addition, preliminary data from a study of bevacizumab in combination with sequential dose-dense chemotherapy in 32 patients with HER2-negative LABC showed a 25 % pCR rate [83]. The GepartQuinto trial evaluated addition of bevacizumab to neoadjuvant treatment with epirubicin + cyclophosphamide followed by docetaxel; the study included 1948 patients with HER2-negative primary BC. A significantly higher tpCR rate was demonstrated in patients who received bevacizumab (18.4% vs. 14.9%; P = 0.04). The effect, however, was restricted primary to the subgroup of patients with TNBC (39.3% vs. 27.9%, P = 0.03) [84], whereas the subgroup of HR-positive BC did not show an improvement (tpCR 7.7% vs. 7.8% P = 1.00). The NSABP B-40, phase III trial evaluated taxane-based chemotherapy +/ capecitabine or gemcitabine +/ bevacizumab followed by doxorubicin plus cyclophosphamide +/ bevacizumab in 1260 patients with HER2-negative primary BC. The study showed a 34.5% pCR rate in the bevacizumab treated population
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
8
Reference
Phase
Patient population
Treatment
End point
Number of patients
pCR (95% CI) %
tpCR (95% CI) %
Triple-negative Kim [72]
II
Triple-negative, stage II/III
B + docetaxel + carboplatin ? surgery
tpCR
45
Snider [73]
II
Triple-negative
pCR
II II
Sánchez-Rovira [76] Sinclair [77]
II
Triple-negative HR + or triple-negative and tumor P 1.5 cm and high grade or N+ or tumor P2.5 cm and intermediate grade Stage II–III, tumor P2 cm (63% HR+)
42 (31 evaluable) 51 84 (HR+) 20 (TN)
42.2 (27.8– 56.6) 52
Ryan [74] Tolaney [75]
B + nabpaclitaxel + carboplatin ? B + doxorubicin + cyclophosphamide B + cisplatin ? surgery ? B + doxorubicin + cyclophosphamide B (single dose) ? B + dose-dense doxorubicin + cyclophosphamide ? B + paclitaxel
44.4 (29.9– 58.9) 55
Doxorubicin + cyclophosphamide ? B + docetaxel
pCR
72
24 (15–36)
II
HR + or triple-negative Stage IIA–IIIC
tpCR
Clavarezza [78]
II
HER2-negative (15 TNBC)
B + nab-paclitaxel + carboplatin B + nab-paclitaxel + carboplatin ? doxorubicin + cyclophosphamide FEC ? B + paclitaxel
pCR
27 28 49
11 54 21 (11.6– 34.4)
HER2-negative Greil [79] Rastogi [80]
II II
Tumor >2 cm, T2–T4 Stage IIIA–B–C
pCR pCR
18 45
22 (6–48) 9 (2.6–22.1)
22 (6–48) 9 (2.6–22.1)
Stage III or inflammatory Inflammatory
tpCR pCR
25 NR
pCR
25
NR
Minckwitz [84]
III, GeparQuinto
B + pegylated liposomal doxorubicin ? B + paclitaxel ? B + cyclophosphamide ? B + hormone therapy Epirubicin + cyclophosphamide (4 cycles) ? docetaxel (4 cycles) ? surgery B (8 cycles) + epirubicin + cyclophosphamide (4 cycles) ? docetaxel (4 cycles) ? surgery
54 101(100 evaluable) 32
NR 27 (18–36)
Carpenter [83]
II II (PACS 09, BEVERLY1) II
B + docetaxel + capecitabine ? surgery B + doxorubicin + cyclophosphamide ? bevacizumab + docetaxel + capecitabine ? surgery ? bevacizumab FEC ? B + paclitaxel B + FEC ? B + docetaxel ? B adjuvant
tpCR
969
Docetaxel ? doxorubicin + cyclophosphamide (B) + docetaxel ? doxorubicin + cyclophosphamide Docetaxel + capecitabine ? doxorubicin + cyclophosphamide B + docetaxel + capecitabine ? doxorubicin + cyclophosphamide Docetaxel + gemcitabine ? doxorubicin + cyclophosphamide B + docetaxel + gemcitabine ? doxorubicin + cyclophosphamide
pCR
tpCR
44
44.4
tpCR
52
tpCR (cohort A)
75
63.5 (49.4– 77.5) 53 (cohort A)
Venturini [81] Viens [82]
Bear [85]
III (NSABP B-40)
LABC Tumor P2 cm, appropriate for adjuvant chemotherapy: T1–T4, N+
Tumor P2 cm, T1c – T3, N0-N2a
NR NR
14.9 (12.7– 17.3) 18.4 (16.0– 21.0) (P = 0.04)
956
(201) 199 (204) 201 (197) 204
HER2-positive Fernandez [88]
II (AVANTHER)
Pierga [89]
II BEVERLY-2
Smith [90]
II
LABC or pathological stage 3 Number in cohorts NR
Yardley [91]
II
T1c-T4, N0-3
B + paclitaxel + trastuzumab ? surgery ? liposomal doxorubicin + cyclophosphamide B + epirubicin + cyclophosphamide + fluorouracil ? bevacizumab + docetaxel + trastuzumab ? surgery ? B + trastuzumab A: EC 4 + B 1 ? docetaxel 4 + B 3 + trastuzumab 4 ? adjuvant B + trastuzumab B: EC 4 ? docetaxel x 4 ? adjuvant B + trastuzumab B + nab-paclitaxel + carboplatin + trastuzumab
Hormone receptor positive Forero-Torres [92]
II
HER2-negative,H+, T2-3, N0-2
B + letrozole
Safety and ORR
25
Forero-Torres [93]
II, randomized
HER2-negative, ER-positive, stage II/III
pCR
Torrisi [94]
II
HR+, T2-4A-C, N0–N2
B + letrozole Letrozole B + capecitabine + vinorelbine + + letrozole
50 25 37 (36
Stage II–III including locally advanced Inflammatory
15 8 44
33.7 31.6 23.5 36.1 27.6 35.8
NR
29
Changes in
25.8 P = 0.75 23.2 P = 0.009 26.9 P = 0.10
54
16 (6–35) 68 (RR) (47–84) 11 (1.9–20.1) 0 0
12 (3–30)
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Table 5 Trials of bevacizumab in early-stage breast cancer, preoperative therapy. Studies with P15 evaluable patients are included.
24 25 Change in biomarker
II
II, randomized
Makhoul [96]
Baar [97]
B: bevacizumab; CR: complete response; H+, hormone receptor-positive; pCR: pathological complete response (breast), TAC: docetaxel, doxorubicin, cyclophosphamide; TN, triple-negative; tpCR, pathological complete response (breast and axilla).
33 41 (90: 27.7– 55.4) 0 8 40 pCR
B + docetaxel + cyclophosphamide ? doxorubicin ? surgery ? B (+trastuzumab as indicated) B + docetaxel Docetaxel
37 30 tpCR II Mixed population Balduzzi [95]
cT2-T4c, cN0-N2 + ER/PgR-negative or cT4d + any ER/PgR and HER2-positive (37%) HER2-positive (20%), H+(55%) Stage IIA–B/IIIAInoperable
Epirubicin + cisplatin + fluorouracil ? B + paclitaxel
RR 86%) Ki67
evaluable)
33 (17–53)
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
9
as compared to 28.4% in patients not receiving the drug (P = 0.02). Of note, in contrast to the Geparquinto study, the effect of bevacizumab was mainly seen in the HR-positive subset of patients (23.2% vs. 15.1%; P = 0.007), with a weaker effect in the HR-negative subset (51.5% vs. 47.1%; P = 0.34) [85]. The discrepancy between the two studies might be related to the inclusion of patients with more advanced disease in the Geparquinto trial and different sequencing of the drug regimens. Furthermore, patients in the Geparquinto trial were withdrawn if they did not respond to the initial four cycles of therapy. In addition, disparities in the classification of patients with hormone-receptor-positive disease vs. patients with TNBC might explain the different findings [86]. Data from Viens et al. supported the findings by von Minckwitz et al. as a significantly greater proportion of pCR was observed in the HR-negative group compared to the HR-positive group (38% vs. 13%; P = 0.007) [87]. HER2-positive BC Four studies investigated the combination of bevacizumab, trastuzumab and chemotherapy in the neoadjuvant setting [88– 91]. All studies reported tpCR rates of approximately 50% comparable to that of trastuzumab–chemotherapy combinations. Two studies found the bevacizumab trastuzumab combination tolerable while Yardley et al. reported an unexpected high rate of discontinuations (20/29) primary due to bevacizumab related toxicity [91]. Hormone receptor positive BC Results from a study evaluating the combination of bevacizumab and letrozole in the preoperative setting demonstrated an 16% pCR rate and a 68% RR among 25 postmenopausal, HR-positive, HER2-negative patients [92], while preliminary results from a randomized phase II study in the same setting showed a 11% pCR rate and a 64.5% RR in the bevacizumab treated subgroup as compared to 0% pCR and a 37.5% RR in the letrozole only arm [93]. In addition, a phase II study of bevacizumab in combination with capecitabine, vinorelbine and letrozole in patients with HR-positive disease showed a RR of 86%, however no pCR was observed [94] . Mixed population Finally, two phase II studies including both HER-positive and negative patients demonstrated pCR/tpCR rates of approximately 30% in the neoadjuvant setting [95,96] whereas a randomized phase II trial including 49 patients with inoperable primary BC treated with docetaxel +/ bevacizumab showed similar RR, PFS and OS in the two groups [97]. Bevacizumab in early-stage breast cancer, adjuvant therapy It has been hypothesized that the most beneficial use of antiangiogenic agents will be in the adjuvant setting. However, concern exists about potential cardiac toxicities including congestive heart failure (CHF) that might develop, particular in patients receiving bevacizumab in combination with anthracyclines. Only two studies have reported data on efficacy of bevacizumab in the adjuvant setting. TNBC BEATRICE was designed to assess the benefit of adding bevacizumab to standard adjuvant chemotherapy, the primary end point being DFS. In total 2591 patients with early TNBC were randomized to standard chemotherapy (anthracycline, taxane or anthracycline + taxane based regimens) alone or in combination with bevacizumab for a total duration of 12 months. The 3-year DFS was nearly identical among women who received bevacizumab and chemotherapy alone (83.7% vs. 82.7%). The related
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
10
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
HR for DFS was 0.87 (95% CI 0.72–1.07, P = 0.18). Further follow-up is needed to assess OS. The side effects included 12% vs. 1% grade 3 or worse hypertension and 1% vs. 0.5% severe cardiac events, all cases occurred in patients receiving anthracycline-based chemotherapy. No increase in fatal adverse events was recorded (
Contents lists available at ScienceDirect
Cancer Treatment Reviews journal homepage: www.elsevierhealth.com/journals/ctrv
Anti-Tumour Treatment
A systematic review of bevacizumab efficacy in breast cancer Iben Kümler ⇑, Ole Grummedal Christiansen, Dorte Lisbet Nielsen 1 Department of Oncology, Herlev Hospital, University of Copenhagen, Herlev Ringvej 75, DK-2730 Herlev, Denmark
a r t i c l e
i n f o
Article history: Received 28 February 2014 Received in revised form 8 May 2014 Accepted 12 May 2014 Available online xxxx Keywords: Angiogenesis Bevacizumab Breast cancer VEGF
a b s t r a c t Angiogenesis is a key component of cancer growth, invasion and metastasis. Therefore, inhibition of angiogenesis is an attractive strategy for the treatment of cancer. We systematically describe phase II and III clinical trials of bevacizumab for the treatment of breast cancer. Methods: A computer-based literature search was carried out using PUBMED and conference databases. Original phase II and III studies reporting P15 patients who received bevacizumab were included. Results: 41 phase II trials were identified in the metastatic setting. Most trials found bevacizumab treatment feasible. Response rates (RR) varied from 0% to 76.5%, time to progression (TTP)/progression free survival (PFS) from 2.4 to 25.3 months and overall survival from 11.5 to more than 38 months. 14 phase III trials including more than 4400 patients with MBC unanimously showed increased RR and PFS, however, no trials demonstrated an OS benefit. In the neoadjuvant setting 23 phase II and III trials were identified. All studies found increased pCR/tpCR but no benefit in terms of OS could be demonstrated. The only study conducted in the adjuvant setting failed to show any survival benefit of bevacizumab. Conclusion: Despite increased response rates in both the metastatic and neoadjuvant setting, bevacizumab has failed to show any OS benefit. Future trials should include identification of robust predictive biomarkers in order to improve our understanding of molecular biomarkers and mechanisms. Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
Methods
Neovascularization is one of the hallmarks of tumor invasion and metastasis and is a prerequisite for progression of solid tumors [1]. For this reason, the inhibition of angiogenesis is regarded an attractive therapeutic approach. Bevacizumab is a recombinant humanized monoclonal antibody that binds to the VEGF-A ligand and prevents it from binding to its receptors [2,3]. Few drugs have generated as much discussion in both the medical literature and the popular press as bevacizumab for metastatic breast cancer (MBC). The decision in 2011 of the US Food and Drug Administration (FDA) to withdraw full approval for the drug in first-line MBC after having granted accelerated approval 3 years earlier renewed the debate not only about the approval process, but also about how to measure clinical outcome and how to evaluate risk benefit of a new drug. This review focusses on efficacy of bevacizumab in phase II and III trials in the (neo)adjuvant and the metastatic setting.
We systematically searched PUBMED for phase II and III studies using the following search terms: bevacizumab AND breast cancer with the clinical trial filter activated. We repeated the search without the clinical trial filter, using the same search terms AND (neo)adjuvant/metastatic breast cancer to ensure that all relevant articles were retrieved. Full articles were obtained and references were checked for additional material when appropriate. In addition, abstracts from 2009 to 2013 annual meetings of the American Society of Clinical Oncology (ASCO) and San Antonio Breast Cancer Symposium 2009–2013 were retrieved for relevant abstracts using similar search terms. The reference list was updated in December 2013 (Fig. 1). The following criteria were applied: Studies reporting P15 evaluable patients, studies reporting efficacy parameters of at least response rate (RR) or time to progression (TTP)/progression free survival (PFS) in the metastatic setting, pathological response (pCR) (breast tumor/breast tumor and lymph nodes) in the neoadjuvant or disease free survival (DFS) in the adjuvant setting and papers in English were included. Two authors independently surveyed the literature ((neo)adjuvant: OGC and DLN; MBC: IB and DN). In case of unclarity or
⇑ Corresponding author. Tel.: +45 38689598; fax: +45 44883094. E-mail addresses: [email protected] (I. Kümler), olegrummedal@gmail. com (O.G. Christiansen), [email protected] (D.L. Nielsen). 1 Tel.: +45 38682344; fax: +45 44883094. http://dx.doi.org/10.1016/j.ctrv.2014.05.006 0305-7372/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
2
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
phase III trials evaluating metronomic chemotherapy and bevacizumab in both early and MBC (NCT00925652, NCT01131195, NCT01112826) [32,33].
Search PUBMED (- Dec. 2013): 98 references
48 articles included 50 excluded - Pre-clinical: 1 -Non- BC population: 7 -Non-clinical: 5 -Biomarker: 8 -Phase I: 6 -Non English language: 3 -Safety: 11 -Other: 9
Repeated search + references from other reviews
8 articles
ASCO and SABC (2009-2013)
24 abstracts
Fig. 1. Diagram of included articles.
disagreement, the complete paper was analyzed and a verdict was reached by consensus. Results Our search identified 22 phase II and 2 phase III trials in the neoadjuvant setting and 1 phase III trial in the adjuvant setting. One reference was retrieved by searching in prior reviews. In the metastatic setting 41 phase II and 14 phase III trials were retrieved. Bevacizumab in predominantly human epidermal growth factor receptor 2 (HER2)-negative MBC Phase II studies Several combinations of bevacizumab and chemotherapy have been investigated in the metastatic setting both as first-line treatment as well as in subsequent treatment-lines. Phase II trials are summarized in Table 1. Taxanes are the drugs most often studied in combination with bevacizumab. Two trials of bevacizumab in combination with docetaxel demonstrated RR of approximately 50% with median PFS/TTP of 7.5 and 9.3 months, respectively [4,5]. Addition of capecitabine in two other studies produced similar results with RR of 49% and 61%, respectively, and TTP of 11 months in both studies [6,7]. Furthermore, several studies have reported on combinations of taxanes and various other drugs, all showing RRs and PFS/TTP in the same range [8–24]. Studies not including taxanes have shown modest activity although there are only few trials [25,26]. A trial of pegylated liposomal doxorubicin and bevacizumab as first-line treatment in locally recurrent and MBC was stopped prematurely due to toxicity after the enrolment of 43 patients [27]. A study of bevacizumab in combination with vinorelbine and capecitabine in recurrent inflammatory breast cancer showed RR of 46.4% [28]. Preclinical studies have suggested that metronomic chemotherapy has anti-angiogenic properties [29]. Bevacizumab in combination with metronomic chemotherapy consisting of cyclophosphamide and capecitabine or methotrexate demonstrated clinical benefit (CRB) rates of 64–68% with PFS of 8–9 months in patients who had previously received anthracyclines and taxanes [30,31]. A randomized trial of bevacizumab in combination with metronomic chemotherapy demonstrated a near three-fold increase in RR (10% versus (vs.) 29%) and TTP (2.0 vs. 5.5 months). The results are preliminary and from 2005. No updates or final publication have been identified, nevertheless, these results have prompted the initiation of several
Phase III Results of phase III trials are given in Table 2. The pivotal openlabel E2100 trial randomized 722 patients (90% HER2-negative) to paclitaxel +/ bevacizumab as first-line therapy. An increased RR (21% vs. 37%, P < 0.001) and a doubling in median PFS (5.9 vs. 11.8 months, P < 0.001) were demonstrated. The median OS, however, was similar in the two groups (25.2 vs. 26.7 months, P = 0.16) [34,35]. Unfortunately, data on treatment administered after progression were not collected, precluding an exploratory analysis of the influence of cross over between treatment arms and subsequent therapy on OS. Results from the placebo-controlled AVADO trial including 736 patients randomized to first-line treatment with docetaxel + placebo, docetaxel + bevacizumab (7.5 mg/kg) or docetaxel + bevacizumab (15 mg/kg) revealed a modest yet statistically significant increased RR (46.4% vs. 55.2% vs. 64.1%) and PFS 8.2 vs. 9.0 vs. 10.1 months, though seemingly to a lesser extent than the bevacizumab plus paclitaxel combination. The median OS was similar in the three groups [36]. In the initial registration study AVF2119g the addition of bevacizumab to capecitabine in patients with advanced, pretreated MBC was rather disappointing. Despite a doubling of the RR in the combination arm (19.8% vs. 9.1%) PFS did not improve (4.8 vs. 4.2 months) nor did OS (15.1 vs. 14.5 months) [37]. Because of these results, the FDA required a third arm including capecitabine in the following RIBBON-1 trial. The RIBBON-1 trial evaluated first-line chemotherapy (taxane, capecitabine or anthracycline-based) +/ bevacizumab. Among 1237 patients the addition of bevacizumab resulted in a significant improvement in RR and PFS. A prespecified subgroup analysis showed that the efficacy of bevacizumab was independent of prior therapy and hormone-receptor (HR) status. However, no significant OS advantage was demonstrated [38]. A pooled subgroup analysis of 2447 individual patients included in E2100, AVADO and RIBBON-1 was undertaken. The analysis showed a PFS of 9.2 months with bevacizumab and 6.7 months without (Hazard ratio (HR) 0.64; 95% confidence interval (CI) 0.57–0.71). The one-year median OS rate was increased (71% vs. 65%). However, OS was not different (26.7 vs. 26.4 months) (HR 0.97; 95% CI 0.86–1.08). The magnitude of benefit was similar in all patient subgroup irrespective of baseline characteristics [39]. The RIBBON-2 trial assessed the efficacy of bevacizumab plus non-anthracycline-based, second-line chemotherapy. The addition of bevacizumab significantly improved PFS. Nevertheless, OS was not significantly different [40]. Of note, an exploratory subgroup analysis of 159 (23%) triple-negative breast cancer (TNBC) patients demonstrated significant improvements in RR (41% vs. 18%; P = 0.0078), PFS (6.0 vs. 2.7 months; (HR 0.494; 95% CI 0.33–0.74; P = 0.0006) and median OS (17.9 vs. 12.6 months; HR 0.624, 95% CI 0.39–1.007; P = 0.0534). Despite the small sample size, there was a trend toward improved OS among this subset of patients [41]. The TURANDOT trial evaluated the combination of bevacizumab plus paclitaxel vs. bevacizumab plus capecitabine. Results from an interim OS analysis after inclusion of 564 patients showed that the non-inferiority criterion was not met and OS results were inconclusive, although PFS was better and more patients experienced response in the paclitaxel arm [42]. Final results are expected in 2014. A one year survival of 78% was reported among 63 TNBC patients receiving bevacizumab plus paclitaxel [43]. A randomized study of bevacizumab plus taxane +/ capecitabine was stopped after the first interim analysis showing no PFS benefit but
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Reference
Therapy
Setting*
Number of patients
Primary end point
RR (95% CI)%
Median TTP (95% CI) months
Median OS (95% CI) months
Hurvitz [4]
B + docetaxel (TORI B01)
1st line
76 (67 evaluable)
3
51 (38.7–62.7)
9.3 (8.2–12.4)
Ramaswamy [5]
B + docetaxel
27
RR
52 (32–71)
7.5 (6.2–8.3) (PFS)
Perez [6]
B + docetaxel + capecitabine
45
RR
49 (34–64)
11.1 (PFS) (8.0–14.3)
Bisagni [7]
B (to PD) + docetaxel + capecitabine (maximum 6 cycles) B + paclitaxel
HER2-negative (93%) 1st + 2nd line HER2-negative (73%) Unknown (27%) 1st line HER2-negative (91%) 1st line 1st line
26.3 (20.8 – not reached NR
79
PFS
61
11 (8.2–13.8)
120
PFS
74 (64.5–81.2)
12.9 (11.1–18.2)
Brufsky [9]
B + paclitaxel B + paclitaxel + gemcitabine
HER2-negative (>90%) 1st line
94 93
RR
48.9 (38.5–59.5) 58.7 (47.9–68.9) (P = 0.117)
8.8 (8.1–10.4) 11.3 (9.7–12.7) (PFS) (HR 0.82; P = 0.247)
Conlin [10]
B + nab-paclitaxel q3w B + nab-paclitaxel q2w B + nab-paclitaxel weekly B + paclitaxel + gemcitabine B + paclitaxel B + paclitaxel
HER2-negative 1st line
72 74 76 21 (17 evaluable) 39 43
RR
42 42 41 88 (CBR) 58.8 58
7.7 6.4 7.7 NR 11 NR
154 156 32
PFS PFS
67 50 66.7 (47.2–82.7)
11 8.4 (HR 0.52; P < 0.001) 13.6 (11.2–21.9)
24.1 23.1 (P = 0.44) 26.8 (13.3–41.2)
Aogi [8]
Guardino [11] Inno [12] Kountourakis [13] Lam [14]
1st line 1st line HER2-negative (93%) 1st line 1st line
TTP NR NR
35.8 (26.4 – not estimated) 25.0 (18.8 – NA) 24.3 (20.3 – NA) (HR 0.84; P = 0.475) NR
NR 33 NR
Lo [15]
B + paclitaxel + capecitabine B + paclitaxel B + nab-paclitaxel + carboplatin
Lobo [16]
B + nab-paclitaxel + gemcitabine
1st line
30
PFS
75.9
10.4 (5.6–15.2) (PFS)
Mavroudis [17] Northfelt [18]
B + docetaxel + epirubicin B + nab-paclitaxel + gemcitabine
64 (60 evaluable) 48
RR PFS
64 69 (54–81)
14 (estimated PFS) PFS 11.7 (8.9–13.1)
Polyzos [19]
B + paclitaxel
40
RR
30 (15.8–44.2)
4.8
13
Rugo [20]
B + ixabepilone (weekly) B + ixabepilone (q3w) B + paclitaxel
1st line HER2-negative (96%) 1st line 2nd line (30%) P 3rd line (70%) 1st line
18 months OS 77.2% (51.1–90.5) NR Not reached
46 45 32
RR
48 (32.9–63.1) 71 (55.7–83.6) 63 (43.7–78.9)
9.6 (6.1–11.7) 11.9 (8.7–14.7) 13.5 (10.0–18.2)
Saloustros [21]
B + paclitaxel + carboplatin
45 (38 evaluable)
RR
76
9.2
91% 89% 91% (1-year OS rates) NR
Salvador [22] Seidman [23]
B + paclitaxel + gemcitabine B + nab-paclitaxel (3 dosing schedules)
PFS RR Toxicity
B + docetaxel + cisplatin
RR
72 45 41 46 80
11.5 (9.0–17.6) 21.3 (19.5–26.6) 19.0 (15.0–28.1) 25.3 (18.2–31.0) 6.2 (3.2–9.1 (PFS)
27.4 (21.9-NA) NR
Tai [24]
82 (76 evaluable) 75 54 79 20
Borson [25]
B + gemcitabine
52
PFS
21.4 (10.3–36.8)
4.8 (3.4–7.6)
Gajria [26] Rochlitz [27] Fumagalli [28]
B + capecitabine (7 days–7 days of) B + pegylated liposomal doxorubicin B + vinorelbine + capecitabine (concurrent) B + vinorelbine + capecitabine (sequential) B + capecitabine + cyclophosphamide
41 43 46 (distribution on regimen NR) 46
RR Toxicity PFS RR
20 (9–36) 21 (10–36) 46.4 11.1 (P = 0.025) 48 (33–63)
8 (6–14) 5.7 (4.6–8.1) (PFS) 4.7 4.3 (P = 0.69) 9.3 (5.8–16)
21 (13.7 – not estimable) NR 15.9 (14.0–21.5) NR
24
RR
31.8 (13.9–54.9)
7.5
Dellapasqua [30] Garcia-Sáenz [31]
B + cyclophosphamide + methotrexate + (trastuzumab)
HER2-negative (97%) 1st line
Triple-negative 1st line 1st line HER2-status NR 1st line 1–3 prior regimens (including adjuvant therapy 1st line (prior taxane-containing (neo)adjuvent therapy) 1st (88%) and 2nd line 1st line HER2-status NR Recurrent inflammatory cancer HER2-negative (98%) 1st–3rd line HER2-positive included taxane and anthracycline refractory#
11.5 (8.7–15.1
NR 13.6
3
B, bevacizumab; CBR, clinical benefit rate; CI, confidence interval; NA not assessable; NR, not reported; RR, response rate; PFS, progression free survival; TTP, time to progression. When not otherwise stated all patients had HER2-negative breast cancer. # Only abstract available, number of HER2-positive patients not reported.
*
28.4 (20.2 – not reached) 32 (26.0–37.9)
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Table 1 Phase II trials of bevacizumab in combination with chemotherapy in patients with predominantly HER2-negative metastatic breast cancer. Studies with P15 evaluable patients are included.
4
Reference
Therapy
Setting*
Number of patients
End point
RR (95% CI) %
Median TTP (95% CI) months
Median OS (95% CI) months
Miller [34]
B + paclitaxel
HER2-negative (>95%) 1st line (E2100)
365 (330 evaluable for response) 350 (316 evaluable for response) 248 247 241 232 230
PFS
36.9 21.2 P < 0.001)
11.8 (PFS) 5.9 (P < 0.001)
26.7 25.2 (P = 0.16)
PFS
55.2 (P = 0.07) 64.1 (P < 0.001) 46.4 19.8 (14.7–25.0) 9.1 (5.4–12.9) (P = .001)
9.0 (HR 0.86 P = 0.12) 10.1 (HR 0.77; P = 0.006) 8.2 4.9 (PFS) 4.2 (HR 0.98; P = 0.857)
30.8 (HR 1.05) 30.2 (HR 1.03) 31.9 15.1 14.5
35.4 23.6 (P = 0.0097) 51.3
8.6 (PFS) 5.7 (HR 0.69; P = 0.0011) 9.2
HR 0.85 (P = 0.27) HR 1.03 (P = 0.83)
37.9 (P = 0.0054)
8.0 (HR 0.64; P < 0.001)
39.5 29.6 (P = 0.0193) 44 27
7.2 5.1 (HR 0.78; P = 0.0072) 11 (10.4–12.9) 8.1 (7.1–9.2) (HR 1.36; P = 0.0052) (HR 1.06)
Paclitaxel Miles [36]
Miller [37]
Robert [38]
Brufsky [40] Lang [42]
Lueck [44]
B (7.5 mg/kg q3w) + docetaxel B (15 mg/kg q3w) + docetaxel Placebo + docetaxel B + capecitabine Capecitabine
1st line (AVADO)
HER2-negative (approximately 75–80%) 1–2 prior chemoterapeutic regimens for MBC (previous anthracycline and taxane) (AVF2119g) HER2-negative + undetermined 1st line (RIBBON-1)
409 206 415
PFS
PFS
B + capecitabine Capecitabine B + taxane or anthracycline-based therapy Taxane or anthracycline-based therapy B + chemotherapy Placebo + chemotherapy B + paclitaxel B + capecitabine
2nd line (RIBBON 2) 1st line (TURANDOT)
459 225 268 265
PFS
B + taxane + capecitabine B + taxane
1st line (TABEA)
202
PFS
207
Non-inferior OS
NR
18.0 16.4 (HR 0.90; P = 0.3741) (HR 1.04; P = 0.059)
NR
B, bevacizumab; CI, confidence interval; HR, hazard ratio; NR, not reported; NS, not significant; RR, response rate; PFS, progression free survival; q3w: every third week; TTP, time to progression. When not otherwise stated all patients had HER2-negative breast cancer.
*
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Table 2 Phase III trials of bevacizumab in combination with chemotherapy in patients with predominantly HER2-negative metastatic breast cancer.
5
PFS 61, 24 received maintenance therapy with B + endocrine therapy ER-positive (85%) 1st line B + taxane ? B + endocrine therapy Militello [53]
II
189
AI, aromatase inhibitor; B, bevacizumab; CI, confidence interval; ER, estrogene receptor; HR, hazard ratio; NR, not reported; RR, response rate; PFS, progression free survival; TTP, time to progression.
36 (22–51)
42 (HR 1.18; P = 0.469)
NR PFS 191 HER2-negative 1st line B + endocrine therapy (letrozole/ fulvestrant) Endocrine therapy Martin [52]
III (LEA)
B + fulvestrant or B + anastrozole (5 HER2-positive patients received trastuzumab) Yardley [51]
II
3.8 (HR 0.83; P = 0.14) 13.5 (10.2–18.2)
41
All: 1 year OS 78%, 2 year OS 61%
9 (5–13) 21 (14 – not reached) All: 13.5 (8.8–19.1) (PFS) 18.4 27 47 PFS
9 (3–22) Safety 43 B + letrozole Traina [50]
II
HER2-positive (5%) Prior endocrine therapy (adjuvant + for metastatic disease) (93%) Relapse during or 612 months after adjuvant AI
41 38
NR
26.9 (12.5–36.2)
6.2 (3.6–10.1) (PFS) 6-months PFS rate: 39 (23–58) 17.1 (8.5–26.2) (PFS) 22 (6–22) (18 evaluable) 6-months PFS rate 33 Prior AI; up to 2 prior endocrine therapies II
Median OS (95% CI) months Median TTP (95% CI) months RR (95% CI) % Primary end point Number of patients Setting
B + fulvestrant
Preclinical research has shown significantly crosstalk between HER2 signaling and angiogenesis [54]. Clinical evidence has supported this correlation. In 611 BC tissue specimens, overexpression of HER2 was significantly associated with expression of VEGF [55]. Thus, combination of inhibitors against HER2 and VEGF might be a rational strategy. Results of trials evaluating bevacizumab in HER2positive MBC are given in Table 4. Preliminary data from a phase II trial on trastuzumab + bevacizumab were promising with a 48% RR and a PFS of 9.2 months in the first-line setting. However, no other studies have investigated this chemo-free option, and despite the abstract being published in 2009, no final publication is available [56]. Two small phase II trials combining bevacizumab and trastuzumab with capecitabine or docetaxel as first-line treatment in MBC demonstrated RR of 73% and 81%, respectively [57,58]. A phase II study of the combination of bevacizumab, trastuzumab and vinorelbine however, was deemed unlikely to reach the predefined threshold for declaring success [59].
Tan [49]
Bevacizumab in HER2-positive MBC
Phase
Table 3 summarizes the results from trials on bevacizumab in combination with antihormonal therapy in MBC. Preclinical studies have shown multiple points of overlap between the estrogen (ER) and angiogenetic pathways [48] and it has been hypothesized that adding bevacizumab to hormonal therapy could result in reversal of required endocrine resistance. Trials combining bevacizumab and antihormonal therapy have shown diverse results with RR ranging from 9% in a study of letrozole and bevacizumab to 47% in a first-line study of anastrozole and bevacizumab and TTP of 17.1 and 21 months, respectively. Two studies with fulvestrant reported RRs of 22% and 27%, respectively with PFS as low as 6.2 months. Patients populations have been heterogenous, making direct comparisons difficult [49–51]. The phase III, LEA trial investigated the addition of bevacizumab to endocrine therapy with letrozole (approximately 90%) or fulvestrant in the first-line setting. The trial included 380 patients among whom approximately 50% had received adjuvant endocrine therapy [52]. PFS in the bevacizumab arm was 18.4 months vs. 13.8 in the endocrine monotherapy arm with no significant difference. Finally, Militello et al. investigated continued treatment with bevacizumab plus endocrine therapy in ER-positive patients who had RR or SD P6 months after receiving bevacizumab + taxane. A PFS of 13.5 months and an OS of 36 months was registered among 24 patients, prompting the authors to conclude that maintenance therapy with bevacizumab and endocrine therapy can extent the overall benefit of therapy [53]. No randomized studies evaluating this treatment strategy were identified.
Therapy
Bevacizumab in hormone receptor-positive breast cancer
Reference
significantly more toxicity including six toxic death in the capecitabine arm [44]. Finally, the non-randomized phase IV ATHENA trial further assessed first-line bevacizumab combined with single agent or combination chemotherapy according to physician’s standard of care. Eligible patients had HER2-negative locally recurrent or MBC (or HER2-positive disease progressing after trastuzumabcontaining therapy). A total of 2251 patients were included and efficacy and safety data were similar to that observed in the randomized trials [45,46]. Among 585 TNBC patients the RR was 49% including 10% complete responses (CR), TTP was 7.2 months (95% CI 6.6–7.8 months) and median OS 18.3 months (95% CI 16.4–19.7) [47]. Thus, although these trials demonstrated significant improvements in RR and TTP/PFS, findings have not indicated substantial benefit in terms of survival.
Table 3 Phase II and III trials of bevacizumab in combination with antihormonal therapy in patients with etrogen and/or progesterone receptor-positive metastatic breast cancer. Studies with P15 evaluable patients are included.
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Reference
Therapy
HER2 targeted therapy and/or chemotherapy Hurvitz [56] B + trastuzumab Martin [57] B + trastuzumab + capecitabine
6
Phase
Setting
Number of patients
Primary end point
RR (95% CI) %
Median PFS (95% CI) months
Median OS (95% CI) months
II II
1st line Prior HER2-targeted (neo)adjuvant therapy (25%) 1st line HER2-negative HER2-positive 1st line 1st or 2nd line
50 88
NR RR
48 73 (62–82)
9.2 (5.4–20.5) 14.4 (10.4 – not reached)
NR NR
72 20
PFS
57.7 81.0
8.5 (5.3–10.8) 13.4 (12.1 – NA)
NR
22 7
PFS at 1 year
73 71
9.9 7.8 (PFS at 1 year 36%/ 29%) 16.5 13.7 (HR 0.82; 0.65–1.02, P = 0.0775) 12.2 11.1 (HR 0.65; P = 0.10)
NR
Schwartzberg [58]
B + doxetacel B + trastuzumab + docetaxel
II
Lin [59]
B + trastuzumab + vinorelbine
II
Gianni [60]
B + trastuzumab + docetaxel Trastuzumab + docetaxel
III
1st line
215 206
PFS
76.5 65.9 (P = 0.0265)
Arteaga [61]
B + paclitaxel +/ carboplatin + trastuzumab Paclitaxel +/ carboplatin + trastuzumab
III
1st line
48 48 (terminated due to poor accrual)
PFS
52 52
Lin [62]
B + carboplatin + (trastuzumab)
II
38
Composite RR, brain RECIST
63 45
NR
Rugo [63]
B + lapatinib
II
HER2-positive (79%) P2 lines Brain metastases Prior trastuzumab; Median 3 prior chemotherapeutic regimens
>38 >38 (HR 1.01; o0.74– 1.38; P = 0.9543 14 12 (HR 1.23; P = 0.60) 12 NR
52 (45 evaluable)
PFS at week 12
13.3 (5.1– 26.8)
5.5 (4.5–7.8)
NR
Agents targeting angiogenesis or HER1 Montagna B + erlotinib + capecitabine + cyclophosphamide [64] Dickler [65] B + erlotinib
II
HER2-negative 1st line
26 (24 evaluable)
Safety, RR
62
9.6 (4.7–15.3
II
37
RR
3 (0–8)
2.4 (1.7–4.0)
Mina [66]
B + sorafenib
II
18
PFS
0
2.8 (1.8–6.1)
Robert [67]
B + paclitaxel Sunitinib + paclitaxel
III
HER2-positive (21%) 2nd or 3rd line Triple-negative (50%) 2nd or 3rd line 1st line
243 242
RR
32.1 32.2
Martin [68]
B + paclitaxel Motesanib + paclitaxel Placebo + paclitaxel B + paclitaxel + sunitinib B + paclitaxel
II
HER2-negative 1st line
RR
II, randomized (SABRE-B) III
HER2-negative 1st line
97 91 94 23 (17 evaluable) 23 (18 evaluable)
52 (41–62) 49 (39–60) 41 (31–52) 71 61
9.2 7.4 (HR 1.82; P = 0.996) 11.5 (9.3–15.9) 9.5 (8.4–12.1) 9.0 (7.4–12.4) NR
HER2-negative 1st line
55 57
PFS
29 16
III
HER2-negative 1st line
56 57 58 57
PFS
NR
Mayer [69]
Yardley [70]
B + paclitaxel + everolimus B + paclitaxel + placebo
Dieras [71]
B + paclitaxel + AMG 386 (10 mg/kg) B + paclitaxel + AMG 386 (3 mg/kg) B + paclitaxel + placebo AMG 386
PFS
8.8 (7.4–9.6) 7.1 (5.5–9.1) HR (0.94; P = 0.79) 11.3 9.2 12.2 10.0 (NS)
Mina [71]
NR
NR
AI, aromatase inhibitor; B, bevacizumab; CI, confidence interval; ER, estrogene receptor-positive; HR, hazard ratio; NS, not significant; NR, not reported; RR, response rate; PFS, progression free survival; TTP, time to progression.
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Table 4 Phase II and III trials of bevacizumab in combination with HER2 targeted therapy, agents targeting HER1 or angiogenesis. Studies with P15 evaluable patients are included.
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Likewise, data from the randomized, phase III AVEREL study showed no significant improvement in PFS after adding bevacizumab to trastuzumab and docetaxel in the first-line setting [60]. ECOG 1105 was a randomized, phase III study evaluating trastuzumab + paclitaxel + carboplatin +/ bevacizumab carboplatin in the first-line setting. The study was terminated early due to poor recruitment [61]. Finally, a study of bevacizumab and carboplatin in patients with brain metastases, among which 79% were HER2-positive reported a RR of 45% [62]. Several studies combining bevacizumab and trastuzumab with a chemotherapeuticum are currently ongoing [54]. Conclusively, bevacizumab in the HER2-positive setting have shown minimal, if any clinical benefit. Furthermore, the combination required stringent cardiac surveillance. A phase II study combining bevacizumab and the dual tyrosine kinase inhibitor of HER1 and HER2, lapatinib in 52 patients with HER2-positive MBC (median 5 prior MBC therapies) showed an RR of 13% and a CBR of 31%. The combination was generally well tolerated [63]. Combined therapy of bevacizumab and agents targeting HER1 HER2-negative BC Erlotinib is an inhibitor of the HER1 (EGFR) tyrosine kinase. A phase II study including 26 patients with HER2-negative, advanced BC receiving first-line metronomic chemotherapy (capecitabine + cyclophosphamide) + bevacizumab + erlotinib showed a relatively high RR of 62%, a PFS of 9.6 months, and an OS of 24 months. Hovewer, 55% of the patients were chemotherapy naïve making the result less impressive [64] (Table 4). Mixed population A phase II trial of bevacizumab in combination with erlotinib demonstrated limited activity of the combination [65] (Table 4). Combined therapy of bevacizumab and agents targeting angiogenesis Angiogenesis is a complex process comprised of multiple signaling pathways. Many of these pathways are redundant, with several ligand-receptor combinations resulting in the same eventual down-stream of events. It is unlikely that tumors are entirely dependent on only one abnormally activated signaling pathway; consequently, treatment with an agent that interferes with a single target may be insufficient. TNBC A small phase II study of bevacizumab in combination with sorafenib showed minimal efficacy and substantial toxicity [66]. HER2-negative BC A randomized study comparing paclitaxel plus bevacizumab +/ the tyrosine kinase inhibitor (TKI) sunitinib included 485 patients. A high frequency of grade 3–4 neutropenia was found in the sunitinib arm precluding the delivery of prescribed doses. The sunitinib regimen was found to be clinically inferior to the bevacizumab regimen [67]. In addition, paclitaxel in combination with bevacizumab, placebo or the TKI motesanib, was evaluated in a randomized study including 282 patients. The study showed no difference in PFS of the three regimens. More patients in the motesanib group had serious adverse events especially gastrointestinal toxicity [68]. A randomized phase II study (SABRE-B) of paclitaxel + bevacizumab vs. paclitaxel + bevacizumab + sunitinib as first-line therapy in HER2-negative MBC reported a high rate of dose
7
modifications and/or discontinuations of sunitinib due to toxicity. The study was closed due to non-feasibility [69]. Additionally, a phase III study of first-line bevacizumab and paclitaxel +/ everolimus demonstrated increased RR in the everolimus arm, however, PFS was similar in the two arms [70]. Finally, a randomized study of paclitaxel + bevacizumab +/ different doses of AMG 386, a fusion protein that inhibits angiogenesis, showed no benefit of the combination [71] (Table 4). Bevacizumab in early-stage breast cancer, preoperative therapy Results of phase II and III trials are given in Table 5. TNBC Bevacizumab combined with docetaxel + carboplatin showed a pCR rate of 44% among 45 TNBC patients [72]. Preliminary results from a study of bevacizumab + nab-paclitaxel + carboplatin followed by doxorubicin + cyclophosphamide in this context demonstrated similar results [73]. Furthermore, preliminary results from a phase II evaluating the combination of bevacizumab plus cisplatin in this setting has shown some activity. However, toxicity including hearing loss, hypertension and pulmonary embolism limited the feasibility of this treatment [74]. Preliminary results from a phase II study evaluating single dose bevacizumab followed by bevacizumab + dose-dense doxorubicin + cyclophosphamide followed by paclitaxel showed a pCR rate of 44% in TNBC as compared to 8% in HR-positive BC [75]. The finding of a higher pCR rate in TNBC was confirmed by Sánchez-Rovira et al. investigating the effect of doxorubicin + cyclophosphamide followed by bevacizumab + docetaxel [76]. In addition, a phase II study of bevacizumab, nab-paclitaxel and carboplatin showed 27% tpCR rate (pCR in breast and axilla) in TNBC as compared to 11% in HR-positive BC whereas subsequent dose-dense docetaxel + cyclophosphamide raised the tpCR rate to 81% and 54% in the two subgroups, respectively [77]. Another small study of 49 HER2-negative patients treated with fluorouracil, epirubicin and cyclophosphamide (FEC) followed by bevacizumab and paclitaxel showed cCR of 21% total, but in 15 patients with TNBC the pCR was 47% [78]. HER2-negative BC Two phase II studies in HER2-negative patients investigating bevacizumab + doxorubicin + cyclophosphamide followed by docetaxel + capecitabine and the drug in combination with docetaxel + capecitabine showed somewhat lower pCR rates of 9% and 22%, respectively [79,80]. Preliminary results from two phase II trials investigating bevacizumab in combination with FEC followed by docetaxel or paclitaxel, respectively in patients with HER2-negative stage III or inflammatory BC showed a pCR of 25–27% [81,82]. In addition, preliminary data from a study of bevacizumab in combination with sequential dose-dense chemotherapy in 32 patients with HER2-negative LABC showed a 25 % pCR rate [83]. The GepartQuinto trial evaluated addition of bevacizumab to neoadjuvant treatment with epirubicin + cyclophosphamide followed by docetaxel; the study included 1948 patients with HER2-negative primary BC. A significantly higher tpCR rate was demonstrated in patients who received bevacizumab (18.4% vs. 14.9%; P = 0.04). The effect, however, was restricted primary to the subgroup of patients with TNBC (39.3% vs. 27.9%, P = 0.03) [84], whereas the subgroup of HR-positive BC did not show an improvement (tpCR 7.7% vs. 7.8% P = 1.00). The NSABP B-40, phase III trial evaluated taxane-based chemotherapy +/ capecitabine or gemcitabine +/ bevacizumab followed by doxorubicin plus cyclophosphamide +/ bevacizumab in 1260 patients with HER2-negative primary BC. The study showed a 34.5% pCR rate in the bevacizumab treated population
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
8
Reference
Phase
Patient population
Treatment
End point
Number of patients
pCR (95% CI) %
tpCR (95% CI) %
Triple-negative Kim [72]
II
Triple-negative, stage II/III
B + docetaxel + carboplatin ? surgery
tpCR
45
Snider [73]
II
Triple-negative
pCR
II II
Sánchez-Rovira [76] Sinclair [77]
II
Triple-negative HR + or triple-negative and tumor P 1.5 cm and high grade or N+ or tumor P2.5 cm and intermediate grade Stage II–III, tumor P2 cm (63% HR+)
42 (31 evaluable) 51 84 (HR+) 20 (TN)
42.2 (27.8– 56.6) 52
Ryan [74] Tolaney [75]
B + nabpaclitaxel + carboplatin ? B + doxorubicin + cyclophosphamide B + cisplatin ? surgery ? B + doxorubicin + cyclophosphamide B (single dose) ? B + dose-dense doxorubicin + cyclophosphamide ? B + paclitaxel
44.4 (29.9– 58.9) 55
Doxorubicin + cyclophosphamide ? B + docetaxel
pCR
72
24 (15–36)
II
HR + or triple-negative Stage IIA–IIIC
tpCR
Clavarezza [78]
II
HER2-negative (15 TNBC)
B + nab-paclitaxel + carboplatin B + nab-paclitaxel + carboplatin ? doxorubicin + cyclophosphamide FEC ? B + paclitaxel
pCR
27 28 49
11 54 21 (11.6– 34.4)
HER2-negative Greil [79] Rastogi [80]
II II
Tumor >2 cm, T2–T4 Stage IIIA–B–C
pCR pCR
18 45
22 (6–48) 9 (2.6–22.1)
22 (6–48) 9 (2.6–22.1)
Stage III or inflammatory Inflammatory
tpCR pCR
25 NR
pCR
25
NR
Minckwitz [84]
III, GeparQuinto
B + pegylated liposomal doxorubicin ? B + paclitaxel ? B + cyclophosphamide ? B + hormone therapy Epirubicin + cyclophosphamide (4 cycles) ? docetaxel (4 cycles) ? surgery B (8 cycles) + epirubicin + cyclophosphamide (4 cycles) ? docetaxel (4 cycles) ? surgery
54 101(100 evaluable) 32
NR 27 (18–36)
Carpenter [83]
II II (PACS 09, BEVERLY1) II
B + docetaxel + capecitabine ? surgery B + doxorubicin + cyclophosphamide ? bevacizumab + docetaxel + capecitabine ? surgery ? bevacizumab FEC ? B + paclitaxel B + FEC ? B + docetaxel ? B adjuvant
tpCR
969
Docetaxel ? doxorubicin + cyclophosphamide (B) + docetaxel ? doxorubicin + cyclophosphamide Docetaxel + capecitabine ? doxorubicin + cyclophosphamide B + docetaxel + capecitabine ? doxorubicin + cyclophosphamide Docetaxel + gemcitabine ? doxorubicin + cyclophosphamide B + docetaxel + gemcitabine ? doxorubicin + cyclophosphamide
pCR
tpCR
44
44.4
tpCR
52
tpCR (cohort A)
75
63.5 (49.4– 77.5) 53 (cohort A)
Venturini [81] Viens [82]
Bear [85]
III (NSABP B-40)
LABC Tumor P2 cm, appropriate for adjuvant chemotherapy: T1–T4, N+
Tumor P2 cm, T1c – T3, N0-N2a
NR NR
14.9 (12.7– 17.3) 18.4 (16.0– 21.0) (P = 0.04)
956
(201) 199 (204) 201 (197) 204
HER2-positive Fernandez [88]
II (AVANTHER)
Pierga [89]
II BEVERLY-2
Smith [90]
II
LABC or pathological stage 3 Number in cohorts NR
Yardley [91]
II
T1c-T4, N0-3
B + paclitaxel + trastuzumab ? surgery ? liposomal doxorubicin + cyclophosphamide B + epirubicin + cyclophosphamide + fluorouracil ? bevacizumab + docetaxel + trastuzumab ? surgery ? B + trastuzumab A: EC 4 + B 1 ? docetaxel 4 + B 3 + trastuzumab 4 ? adjuvant B + trastuzumab B: EC 4 ? docetaxel x 4 ? adjuvant B + trastuzumab B + nab-paclitaxel + carboplatin + trastuzumab
Hormone receptor positive Forero-Torres [92]
II
HER2-negative,H+, T2-3, N0-2
B + letrozole
Safety and ORR
25
Forero-Torres [93]
II, randomized
HER2-negative, ER-positive, stage II/III
pCR
Torrisi [94]
II
HR+, T2-4A-C, N0–N2
B + letrozole Letrozole B + capecitabine + vinorelbine + + letrozole
50 25 37 (36
Stage II–III including locally advanced Inflammatory
15 8 44
33.7 31.6 23.5 36.1 27.6 35.8
NR
29
Changes in
25.8 P = 0.75 23.2 P = 0.009 26.9 P = 0.10
54
16 (6–35) 68 (RR) (47–84) 11 (1.9–20.1) 0 0
12 (3–30)
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
Table 5 Trials of bevacizumab in early-stage breast cancer, preoperative therapy. Studies with P15 evaluable patients are included.
24 25 Change in biomarker
II
II, randomized
Makhoul [96]
Baar [97]
B: bevacizumab; CR: complete response; H+, hormone receptor-positive; pCR: pathological complete response (breast), TAC: docetaxel, doxorubicin, cyclophosphamide; TN, triple-negative; tpCR, pathological complete response (breast and axilla).
33 41 (90: 27.7– 55.4) 0 8 40 pCR
B + docetaxel + cyclophosphamide ? doxorubicin ? surgery ? B (+trastuzumab as indicated) B + docetaxel Docetaxel
37 30 tpCR II Mixed population Balduzzi [95]
cT2-T4c, cN0-N2 + ER/PgR-negative or cT4d + any ER/PgR and HER2-positive (37%) HER2-positive (20%), H+(55%) Stage IIA–B/IIIAInoperable
Epirubicin + cisplatin + fluorouracil ? B + paclitaxel
RR 86%) Ki67
evaluable)
33 (17–53)
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
9
as compared to 28.4% in patients not receiving the drug (P = 0.02). Of note, in contrast to the Geparquinto study, the effect of bevacizumab was mainly seen in the HR-positive subset of patients (23.2% vs. 15.1%; P = 0.007), with a weaker effect in the HR-negative subset (51.5% vs. 47.1%; P = 0.34) [85]. The discrepancy between the two studies might be related to the inclusion of patients with more advanced disease in the Geparquinto trial and different sequencing of the drug regimens. Furthermore, patients in the Geparquinto trial were withdrawn if they did not respond to the initial four cycles of therapy. In addition, disparities in the classification of patients with hormone-receptor-positive disease vs. patients with TNBC might explain the different findings [86]. Data from Viens et al. supported the findings by von Minckwitz et al. as a significantly greater proportion of pCR was observed in the HR-negative group compared to the HR-positive group (38% vs. 13%; P = 0.007) [87]. HER2-positive BC Four studies investigated the combination of bevacizumab, trastuzumab and chemotherapy in the neoadjuvant setting [88– 91]. All studies reported tpCR rates of approximately 50% comparable to that of trastuzumab–chemotherapy combinations. Two studies found the bevacizumab trastuzumab combination tolerable while Yardley et al. reported an unexpected high rate of discontinuations (20/29) primary due to bevacizumab related toxicity [91]. Hormone receptor positive BC Results from a study evaluating the combination of bevacizumab and letrozole in the preoperative setting demonstrated an 16% pCR rate and a 68% RR among 25 postmenopausal, HR-positive, HER2-negative patients [92], while preliminary results from a randomized phase II study in the same setting showed a 11% pCR rate and a 64.5% RR in the bevacizumab treated subgroup as compared to 0% pCR and a 37.5% RR in the letrozole only arm [93]. In addition, a phase II study of bevacizumab in combination with capecitabine, vinorelbine and letrozole in patients with HR-positive disease showed a RR of 86%, however no pCR was observed [94] . Mixed population Finally, two phase II studies including both HER-positive and negative patients demonstrated pCR/tpCR rates of approximately 30% in the neoadjuvant setting [95,96] whereas a randomized phase II trial including 49 patients with inoperable primary BC treated with docetaxel +/ bevacizumab showed similar RR, PFS and OS in the two groups [97]. Bevacizumab in early-stage breast cancer, adjuvant therapy It has been hypothesized that the most beneficial use of antiangiogenic agents will be in the adjuvant setting. However, concern exists about potential cardiac toxicities including congestive heart failure (CHF) that might develop, particular in patients receiving bevacizumab in combination with anthracyclines. Only two studies have reported data on efficacy of bevacizumab in the adjuvant setting. TNBC BEATRICE was designed to assess the benefit of adding bevacizumab to standard adjuvant chemotherapy, the primary end point being DFS. In total 2591 patients with early TNBC were randomized to standard chemotherapy (anthracycline, taxane or anthracycline + taxane based regimens) alone or in combination with bevacizumab for a total duration of 12 months. The 3-year DFS was nearly identical among women who received bevacizumab and chemotherapy alone (83.7% vs. 82.7%). The related
Please cite this article in press as: Kümler I et al. A systematic review of bevacizumab efficacy in breast cancer. Cancer Treat Rev (2014), http://dx.doi.org/ 10.1016/j.ctrv.2014.05.006
10
I. Kümler et al. / Cancer Treatment Reviews xxx (2014) xxx–xxx
HR for DFS was 0.87 (95% CI 0.72–1.07, P = 0.18). Further follow-up is needed to assess OS. The side effects included 12% vs. 1% grade 3 or worse hypertension and 1% vs. 0.5% severe cardiac events, all cases occurred in patients receiving anthracycline-based chemotherapy. No increase in fatal adverse events was recorded (
