A randomized phase III double-blind placebo-controlled trial of first line chemotherapy and trastuzumab with or without bevacizumab for patients with HER2/neu-positive metastatic breast cancer: a trial of the ECOG-ACRIN Cancer Research Group (E1105)

Background In 2008, bevacizumab received accelerated Food and Drug Administration (FDA) approval for use in human epidermal growth factor receptor 2 (HER2)-negative metastatic breast cancer (MBC). Based on the preclinical and preliminary clinical activity of the trastuzumab and bevacizumab combination, ECOG-ACRIN E1105 trial was developed to determine if the addition of bevacizumab to a chemotherapy and trastuzumab combination for first-line therapy would improve progression-free survival (PFS) in patients with HER2-positive MBC. Findings: 96 patients were randomized to receive standard first-line chemotherapy and trastuzumab with or without bevacizumab between November 2007 and October 2009, and 93 began protocol therapy. Induction therapy was given for 24 weeks, followed by maintenance trastuzumab with or without bevacizumab. 60% (56/93) began carboplatin and 74% (69/93) completed 6 cycles of induction therapy. Primary endpoint was PFS. Median PFS was 11.1 and 13.8 months for placebo and bevacizumab arms, respectively (hazard ratio [HR] 95%, Confidence Interval [Cl] for bevacizumab vs. placebo: 0.73 [0.43–1.23], p = 0.24), and at a median follow-up of 70.7 months, median survival was 49.1 and 63 months (HR [95% Cl] for OS: 1.09 [0.61–1.97], p = 0.75). The most common toxicities across both arms were neutropenia and hypertension, with left ventricular systolic dysfunction, fatigue, and sensory neuropathy reported more frequently with bevacizumab. Conclusions In this trial, the addition of bevacizumab did not improve outcomes in patients with metastatic HER2-positive breast cancer. Although the trial was underpowered due to smaller than anticipated sample size, these findings corroborated other clinical trials during this time.


Introduction
In 2008, bevacizumab received accelerated Food and Drug Administration (FDA) approval for use in human epidermal growth factor receptor 2 (HER2)-negative metastatic breast cancer (MBC).This was based off the E2100 study, which showed a combination of paclitaxel plus bevacizumab signi cantly prolonged progression-free survival (PFS) in treatment-naïve patients with MBC (1).By 2011, however, the FDA withdrew its approval for bevacizumab due to the lack of evidence of OS bene t and concern for unacceptable toxicity (2).
During this time period, the interest in the use of bevacizumab in HER2-positive breast cancer was also high due to studies demonstrating association between HER2 ampli cation and increased vascular endothelial growth factor (VEGF) in breast cancer (3)(4)(5).Additionally, a phase II trial at the time showed combining bevacizumab with trastuzumab in the treatment of HER2-positive MBC was both clinically feasible and active in the absence of chemotherapy (6).Based on the preclinical and preliminary clinical activity of the trastuzumab and bevacizumab combination, E1105 was developed to determine if the addition of bevacizumab to rst-line chemotherapy and trastuzumab would improve PFS in patients with HER2-positive MBC.

Participants
Patients ≥ 18 years with histologically con rmed HER2-positive MBC, ECOG performance status of 0 or 1, adequate hematological, neurological, cardiac and end-organ function, and no prior systemic therapies were considered for enrollment (7).Prior taxane and trastuzumab were allowed if given > 12 months prior to recurrence.The study was coordinated by the ECOG-ACRIN Cancer Research Group (ECOG-ACRIN), in collaboration with Radiation Therapy Oncology Group (RTOG), North Central Cancer Treatment Group (NCCTG), Cancer and Leukemia Group B (CALGB), Southwest Oncology Group (SWOG), National Surgical Adjuvant Breast and Bowel Project (NSABP), Cancer Trials Support Unit (CTSU).Written informed consent was obtained from all patients before enrolment.

Treatment
Patients were randomized to receive standard rst-line induction chemotherapy (paclitaxel 90 mg/m2 IV weekly x3 every 4 weeks [6 cycles] or paclitaxel 80 mg/m2 IV weekly x3 every 4 weeks + carboplatin AUC 2 IV weekly x3 every 4 weeks [6 cycles]) with trastuzumab (2 mg/kg IV weekly [after initial loading dose of 4 mg/kg] for 6 cycles) and placebo (PLAC) or bevacizumab (BEV) (10mg/kg IV every 2 weeks for 24 weeks [6 cycles]), followed by maintenance trastuzumab (6 mg/kg IV every 3 weeks) and BEV or PLAC (15 mg/kg IV every 3 weeks) until disease progression, severe adverse event, pregnancy, withdrawal, or death.Participants were allowed to discontinue chemotherapy and proceed to maintenance therapy.If trastuzumab or bevacizumab was discontinued, chemotherapy could continue.A schematic of the trial is provided in Fig. 1.

Assessments
Tumor (Computer tomography/ bone scan) and cardiac (echocardiogram or MUGA) assessments were performed at baseline, every 3 months, and 3 months post treatment.Tumor assessments continued until rst progression.Complete blood counts were assessed prior to every cycle.Quality of life (QOL) assessments (FACIT-Fatigue Subscale, FACT/NCCN Breast Symptom Index, FACT/GOG-Ntx, FACT-G item GP5) were completed at baseline, end of cycles 3 and 6 induction, cycle 5 maintenance, 12 months postrandomization, and annually to 60 months post-randomization.

Statistical Considerations and Endpoints
Progression-free survival (PFS) was the primary endpoint and was de ned as time from randomization to rst disease progression via RECIST 1.0, new second breast primaries, or to death from any cause.Blinded treatment assignments were made in permuted blocks in a 1:1 fashion to PLAC or BEV.
Randomization was strati ed by prior adjuvant trastuzumab use, prior taxane use, disease-free interval (≤ 24 months, > 24 months), and planned carboplatin (yes, no).The accrual goal was 416 patients where 301 PFS events provided 86% power to detect a 30% reduction in the failure hazard rate.The trial was monitored by the ECOG-ACRIN DSMC, including a prespeci ed cardiac stopping rule for high rates of clinical CHF in BEV.
Secondary endpoints included overall survival (OS), de ned as time from randomization date to death from any cause.The Kaplan-Meier method was used to estimate time-to-event distributions.Cox proportional hazards models were used to estimate hazard ratios and test for signi cance.Toxicities were graded according to CTCAE version 3.0.Cardiac safety pro les included clinical congestive heart failure (symptomatic decline in LVEF to below the lower limit of normal or symptomatic diastolic dysfunction).Baseline characteristics are reported among 95 of 96 with baseline information available, speci c treatment information is reported among 93 of 96 patients who began protocol therapy, and best response, PFS and OS are analyzed on an intent-to-treat basis.

Baseline Characteristics
Between November 9, 2007, and October 28, 2009, 96 of 416 patients were enrolled.Due to slow accrual, the trial was closed after October 2009.Table 1 provides a summary of baseline characteristics.
* Only patients who began protocol treatment are included in this summary (One patient on Arm A did not begin protocol treatment, and two patients on Arm B did not begin protocol treatment.)AE reporting in the trial was limited to ≥ grade 4 hematologic events, ≥ grade 2 non-hematologic events for a select group of AEs, and ≥ grade 3 non-hematologic events otherwise.
^patient died within 30 days of ending protocol treatment.

Discussion
During the enrollment of the presented study, new drug approvals such as lapatinib (10), decreased metastatic recurrence with trastuzumab, and excitement regarding other HER2 targeted agents (such as pertuzumab and trastuzumab-emtansine) resulted in low patient accrual and early study closure in 2009.Due to low accrual, the study was underpowered, and no signi cant difference in clinical outcomes was observed between treatment arms.The safety pro les for bevacizumab and trastuzumab were consistent with prior phase I, II and III trials.Cardiac toxicity is a known side effect of trastuzumab; however, previous studies found the toxicity was reversible, unlike doxorubicin-induced cardiomyopathy (11,12).We saw few overall cardiac adverse events from bevacizumab in addition to the hallmark hypertension associated with anti-angiogenic drugs (13).The results from this trial corroborated the 2011 FDA decision to remove bevacizumab as a recommended therapeutic option for patients with breast cancer.
Since the closure of E1105, other clinical trials have explored the combination of HER2-targeted and antiangiogenic therapies.A phase II single-arm trial of bevacizumab, trastuzumab, and capecitabine showed clinical activity as rst-line therapy for patients with HER2-positive MBC, with no unexpected toxicities, and a median time to progression of 14.5 months (95% Cl, 10.5 months to NR) (14).The AVEREL study randomized 424 rst-line patients to trastuzumab/docetaxel with or without concomitant bevacizumab (15).Despite a trend favoring bevacizumab PFS (median 13.7 vs. 16.5 months; HR 0.82, log-rank P = 0.078), no difference was observed in overall survival (15).The BETH study randomized 3509 patients with HER2-positive early stage breast cancer to receive standard chemotherapy/trastuzumab with or without bevacizumab for 1 year of adjuvant therapy (16).After 38 months of follow-up, there was no statistically signi cant difference between treatment arms (92% IDFS rates in both groups).Currently, the majority of recent studies using bevacizumab to treat breast cancer are in the HER2-negative patient population or are in subsets of HER2-positive and -negative patients with speci c types of refractory disease (17,18).

Table 2 :
Clinical Response and Outcomes **P-values are two-sided and are based on strati ed test using strati cation factors at randomization.*** Adjusted for Age, ERPR, Primary surgery.

Table 2 :
Clinical Response and Outcomes +last censored observation as upper limit not reached.