This work describes the considerations that led to the approval by the U.S. Food and Drug Administration (FDA), on November 15, 2010, of eribulin mesylate (Halaven; Eisai, Inc.) for the treatment of patients with refractory metastatic breast cancer. The FDA review focused primarily on the results of a single randomized, open-label, multicenter trial of 762 patients with refractory locally advanced or metastatic breast cancer. The patients were randomized to receive eribulin or any single-agent treatment of the physician's choice, selected prior to randomization. The FDA's approval of eribulin mesylate was based on demonstration of a statistically significant prolongation of overall survival (OS) in patients who had been randomized to receive eribulin. The median OS was 13.1 months in the eribulin arm compared with 10.6 months in the control arm [HR 0.81 (95% CI, 0.66–0.99); P = 0.041]. Treatment with eribulin did not show a statistically significant treatment effect [HR 0.87 (95% CI, 0.71–1.05)] on progression-free survival as determined by independent review. This approval highlights the appropriate use of an innovative trial design and shows that improvement in OS is an achievable endpoint in the setting of advanced breast cancer. On the basis of the different conclusions arising from the OS and progression-free survival results, investigators should consider using OS as a primary endpoint in clinical trials for refractory breast cancer. Clin Cancer Res; 18(6); 1496–505. ©2012 AACR.

Previously published works described the results from key clinical trials of eribulin mesylate and investigators' conclusions regarding the safety and activity of this agent in patients with refractory metastatic breast cancer (1–4). This article summarizes the U.S. Food and Drug Administration (FDA)'s independent analyses of the raw clinical data provided in the New Drug Application (NDA) for eribulin mesylate. In addition, it provides a summary of the regulatory issues considered by the FDA during the NDA review process, and the factors that led to the FDA's approval of eribulin mesylate for the treatment of refractory metastatic breast cancer.

On the basis of data provided by the Surveillance, Epidemiology, and End Results (SEER) program, it is estimated that 1 in 8 girls born in the United States today will be diagnosed with breast cancer during their lifetime (5). Despite recent advances in treatment, breast cancer remains the second leading cause of cancer-related death in women (6). Although localized breast cancer is potentially curable, current therapies for metastatic breast cancer generally result in delay of further spread of disease or prolongation of life, not cure.

Although selection of therapy for metastatic breast cancer is largely based on tumor characteristics, such as estrogen receptor (ER), progesterone receptor (PR), and HER2/neu status, optimal treatment requires an approach that is tailored to each patient. A variety of treatment options are available, including tamoxifen, an aromatase inhibitor, fulvestrant for women with ER+ or PR+ disease, or cytotoxic chemotherapy administered as monotherapy or in combination with other cytotoxic drugs or biologics (e.g., trastuzumab for HER2/neu+ tumors).

The choice of chemotherapy following failure of first- and second-line therapy for metastatic breast cancer is not well defined and is based on several factors, including previous chemotherapy exposure, performance status, comorbid conditions, menopausal status, and tumor characteristics. There are many treatment options for patients with metastatic breast cancer that has progressed during or shortly after treatment with an anthracycline or taxane, but few have been approved by the FDA specifically for use in this setting. Capecitabine, as monotherapy or in combination with ixabepilone, is approved to treat metastatic or locally advanced breast cancer after failure of an anthracycline and a taxane, and ixabepilone is approved as monotherapy for advanced breast cancer after failure of an anthracycline, a taxane, and capecitabine; however, a variety of other chemotherapy drugs are also administered to patients with refractory metastatic disease (7).

Eribulin mesylate injection is a synthetic analogue of halichondrin B, a naturally occurring product that is isolated from the marine sponge Halichondria okadai (8). Eribulin binds to tubulin at a site distinct from other microtubule-inhibiting–drug binding sites (9–11). The binding of eribulin to tubulin inhibits the growth phase of microtubules and causes the sequestration of microtubules into nonfunctional aggregates. This interference with microtubular function leads to blockade of the G2–M phase of the cell cycle, disruption of mitotic spindles, and apoptosis of tumor cells (12, 13).

The first-in-humans study of eribulin was conducted by the California Cancer Consortium and the National Cancer Institute in 2002 (14), and Eisai (Andover, MA) commenced phase I studies of eribulin in 2003 (15). Neutropenia, febrile neutropenia, and fatigue were the most common dose-limiting toxicities identified in dose-finding studies (14–16). The first phase II trial of eribulin in patients with metastatic breast cancer evaluated an eribulin dose of 1.4 mg/m2 administered as an i.v. bolus on days 1, 8, and 15 of a 28-day cycle (1). However, the dosing regimen was modified to 1.4 mg/m2 administered on days 1 and 8 of a 21-day cycle to reduce the frequency of dose reductions and delays due to neutropenia (1).

Eisai conducted 2 single-arm studies and one randomized active-controlled trial of eribulin in patients with advanced breast cancer who received prior treatment with anthracycline- and taxane-containing chemotherapy regimens (1–4). Clinical data from these trials provided the primary basis for the NDA submitted by Eisai on March 30, 2010.

A multidisciplinary team consisting of FDA chemists, toxicologists, pharmacologists, physicians, and regulatory specialists reviewed the eribulin NDA. To provide a framework for the discussion of the clinical issues addressed during the review of this application, in this section we briefly summarize the key considerations and findings of the chemistry and manufacturing, nonclinical pharmacology/toxicology, and clinical pharmacology review teams.

Eribulin mesylate has a molecular weight of 826.0 (729.9 for the free base) and a complicated chemical structure containing 19 asymmetric centers (17). Manufacture of eribulin mesylate requires a complex series of synthesis and purification steps. Therefore, in its review of the chemistry and manufacturing of the drug, the FDA noted that proper designation of the starting materials was a key factor in ensuring adequate quality and identity of the final drug substance.

The FDA reviewed data from several nonclinical studies submitted in support of the marketing application. Eribulin-related toxicities occurred in hematopoietic organs, testes, liver, and peripheral nerves in rats, and nonreversible testicular effects occurred in both rats and dogs. Additionally, data from genetic toxicology studies provided evidence that eribulin is genotoxic to mammalian cells and thus is potentially carcinogenic (13). An embryo-fetal developmental toxicity study in rats confirmed that eribulin causes adverse effects that are expected with microtubule inhibition (13).

Eisai submitted pharmacokinetic data collected from 393 adult cancer patients who received i.v. injections of eribulin over 2 to 5 minutes at doses ranging from 0.25 mg/m2 to 4.0 mg/m2 (18). These studies showed that the disposition of eribulin followed linear kinetics (19). The mean Cmax (± SD) of a single dose of eribulin at 1.4 mg/m2 was 519 (± 21 ng/mL), and the half-life was ∼40 hours (18). No accumulation occurred following weekly administration of eribulin (19). Eribulin protein binding in the concentration range of 100 ng/mL to 1,000 ng/mL was relatively low, varying from 49% to 65% in human plasma (19).

Although eribulin is primarily excreted unchanged through the fecal route of elimination, exposure appeared to be increased in patients with mild and moderate hepatic impairment and moderate renal impairment (19). The FDA reviewed data from a dedicated hepatic impairment study in 7 patients with Child-Pugh A (mild hepatic impairment) and 5 patients with Child-Pugh B cirrhosis (moderate hepatic impairment). Eribulin systemic exposure increased 1.8-fold and 2.5-fold in patients with mild and moderate hepatic impairment, respectively. On the basis of these data, FDA clinical pharmacologists recommended that eribulin should not be used to treat patients with severe hepatic impairment, and the starting dose of eribulin should be reduced for patients with mild or moderate hepatic impairment (18). Additionally, on the basis of a noncompartmental analysis showing a 2-fold increase in the geometric mean dose-normalized area under curve of eribulin in patients with moderate renal impairment, FDA clinical pharmacologists recommended that the eribulin dose should also be reduced in patients with moderate renal impairment (19).

A dedicated QT study in 26 patients identified a delayed, concentration-independent effect on prolongation of the QTc interval by eribulin; the maximum mean QTcF change from baseline [95% upper confidence interval (CI)] was 11.4 (19.5) ms on day 8 (19). Therefore, FDA reviewers recommended that product labeling include a warning that eribulin causes delayed QT prolongation and should be avoided in patients with congenital long QT syndrome.

Review process

A single, multicenter, open-label, randomized, controlled trial in patients with refractory breast cancer [study E7389-G000-305 (study 305)] provided the primary data to support the eribulin NDA (4). The application also contained supportive data from 2 single-arm phase II studies in similar patient populations [studies E7389-A001-201 (study 201) and E7389-G000-211 (study 211)]. The safety database contained data from 1,222 eribulin-treated patients, including 827 patients with breast cancer who were treated at the proposed eribulin dose and schedule (1.4 mg/m2 i.v. on days 1 and 8 of a 21-day cycle) in studies 305, 201, and 211. On-site inspection of selected study centers in the United States and abroad was performed by the FDA's Division of Scientific Investigations.

Study design

Study 305 was an open-label, randomized, multicenter, international trial of 762 adult women with advanced breast cancer who had received at least 2 prior chemotherapeutic regimens for the treatment of metastatic disease. Additionally, patients were required to have received prior anthracycline- and taxane-based chemotherapy for adjuvant or metastatic disease. Patients with HER2+ or hormone-receptor–positive tumors were permitted, but not required, to have received prior treatment with trastuzumab or hormone therapy, respectively. As a requirement of the protocol, the patients must have experienced disease progression within 6 months of their last chemotherapeutic regimen. The protocol excluded patients with inadequate bone marrow function, brain metastases, meningeal carcinomatosis, pregnancy, severe or unstable intercurrent illness, or baseline peripheral neuropathy of grade 3 severity or greater according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE version 3.0).

This protocol employed a “physician's choice” control arm that allowed investigators to choose any single therapy. The protocol randomized patients (2:1) to receive eribulin mesylate (n = 508) or therapy selected by their physician prior to randomization (control arm, n = 254). Randomization was stratified by geographic region, HER2/neu tumor status, and prior capecitabine exposure. Control-arm therapy could consist of any available single-agent chemotherapy, hormone treatment, or biologic therapy approved for the treatment of cancer, or palliative therapy. Eribulin was administered at a dose of 1.4 mg/m2 i.v. over 2 to 5 minutes on days 1 and 8 of a 21-day cycle, whereas patients randomized to the control arm received the treatment chosen by their physician as directed by the package insert or local practice. The protocol permitted the use of granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage CSF (GM-CSF), and erythropoietin during the study.

Prior to initiation of each cycle, the protocol required improvement of eribulin-related toxicities to grade 2 or less (NCI CTCAE), a platelet count of 75,000/μL, and an absolute neutrophil count of 1,000/μL. Study therapy was permanently discontinued in the case of progressive disease, unacceptable toxicity, withdrawal of consent, a physician's determination that the study therapy was not in the best interest of the patient, development of a comorbid medical condition precluding continued study treatment, pregnancy, lack of study compliance, or delay in treatment by more than 14 days due to toxicities.

Protocol-specified monitoring included baseline tumor assessments according to Response Evaluation Criteria in Solid Tumors (RECIST) with radiographic imaging and photographs of skin lesions. Repeat imaging or photography of target lesions was performed every 8 weeks (or sooner if disease progression was suspected), at study termination, and every 3 months thereafter for patients who discontinued study therapy prior to disease progression. Laboratory parameters, including serum chemistries, hepatic function parameters, and complete blood counts, were monitored on days 1, 8, and 15 of each 21-day cycle. Symptom-directed physical examinations and queries for constitutional symptoms occurred on days 1, 8, and 15 for the first 2 cycles, and then on days 1 and 8 during subsequent cycles. Adverse events were documented starting from the time the informed consent document was signed until the study termination visit, which occurred 0 to 30 days after the final dose of study medication. Serious adverse events were collected for a minimum of 30 days following study drug discontinuation. Patients were followed every 3 months after the study termination visit for disease progression and survival.

The primary analysis population for efficacy was the intent-to-treat population, which consisted of all patients who were randomized, irrespective of whether they received study therapy. The primary efficacy endpoint was overall survival (OS), defined as the time from randomization until death from any cause. Patients who were lost to follow-up were censored at the last date the patient was known to be alive, and patients who remained alive were censored at the time of data cutoff (May 12, 2009). Secondary endpoints included progression-free survival (PFS), objective response rate, and duration of response based on independent reviewer assessments.

The original statistical analysis plan specified enrollment of 630 patients (420 in the eribulin arm and 210 in the control arm) in order to observe 411 deaths as required for the OS analysis. This sample size, based on an estimated median OS of 12 months and 9 months in the eribulin and control arms, respectively (i.e., HR of 0.75), was calculated to provide an 80% probability of detecting a 3 month difference in OS between the treatment arms, with a 2-sided type I error rate of 5%. After a prespecified interim assessment revealed that the number of patient deaths that occurred within 15 months of enrollment of the first patient was lower than expected, the sample size was increased to allow enrollment of up to a maximum of 1,000 patients (the number of events for the final analysis remained unchanged). The final analysis was done with the use of a 2-sided stratified log-rank test at the nominal significance level of 0.049 (adjusted for one interim analysis). The Kaplan–Meier method was used to estimate the median OS and 2-sided 95% CIs for the treatment arms, and the HR for OS was calculated with the use of a Cox regression model that included HER2/neu status, prior capecitabine exposure, and geographical region as strata.

Efficacy results

The report submitted by Eisai for study 305 contained data obtained from the first patient visit on November 16, 2006, until May 12, 2009, the date of primary data cutoff. A total of 762 patients in 135 centers and 19 countries were randomized to receive eribulin (n = 508) or a single-agent therapy prespecified by their physician (n = 254). As shown in Table 1, the demographic and disease-related characteristics were similar between treatment arms. The majority of patients enrolled in study 305 were white (92%), and all were women. The median time from diagnosis to enrollment was 5 years in both arms. Approximately one fifth (19%) of the patients were accrued from U.S. sites.

Table 1.

Demographic and disease characteristics: study 305

Patient characteristicsEribulin (n = 508)Control (n = 254)Total (N = 762)
Age (years) 
 Median in years (range) 55 (28–85) 56 (27–81) 55 (27–85) 
 <40 7% 7% 7% 
 ≥65 19% 22% 20% 
Geographic region 
 N. America/W. Europe/Australia 64% 64% 64% 
 Eastern Europe 25% 25% 25% 
 Latin America/South Africa 11% 11% 11% 
ECOG performance status 
 0 43% 41% 42% 
 1 48% 50% 49% 
 2 8% 9% 8% 
Not reported 2% 1% 1% 
Reproductive status 
 Postmenopausal 75% 78% 76% 
Tumor receptor status 
 ER+ 66% 67% 67% 
 PR+ 50% 48% 49% 
 HER2/neu overexpression 16% 16% 16% 
 ER, PR, HER2 18% 20% 19% 
Visceral disease at enrollment 
 Present 81% 83% 82% 
Number of organs involved 
 1 17% 14% 16% 
 >1 83% 86% 84% 
Patient characteristicsEribulin (n = 508)Control (n = 254)Total (N = 762)
Age (years) 
 Median in years (range) 55 (28–85) 56 (27–81) 55 (27–85) 
 <40 7% 7% 7% 
 ≥65 19% 22% 20% 
Geographic region 
 N. America/W. Europe/Australia 64% 64% 64% 
 Eastern Europe 25% 25% 25% 
 Latin America/South Africa 11% 11% 11% 
ECOG performance status 
 0 43% 41% 42% 
 1 48% 50% 49% 
 2 8% 9% 8% 
Not reported 2% 1% 1% 
Reproductive status 
 Postmenopausal 75% 78% 76% 
Tumor receptor status 
 ER+ 66% 67% 67% 
 PR+ 50% 48% 49% 
 HER2/neu overexpression 16% 16% 16% 
 ER, PR, HER2 18% 20% 19% 
Visceral disease at enrollment 
 Present 81% 83% 82% 
Number of organs involved 
 1 17% 14% 16% 
 >1 83% 86% 84% 

Abbreviation: ECOG, Eastern Cooperative Oncology Group.

Table 2 shows that patients enrolled in study 305 were heavily pretreated. The patients received a median of 4 chemotherapy regimens prior to enrollment. The median duration of the last chemotherapy for metastatic disease prior to study entrance was 3.6 months in the eribulin arm and 3.5 months in the control arm. Aside from a slight imbalance in the percentage of patients who were refractory to anthracyclines, patients in the treatment arms were similar with respect to previous treatment exposure and refractoriness to prior therapy.

Table 2.

Prior anticancer therapy: study 305

ParameterEribulin (n = 508)Control (n = 254)Total (N = 762)
Number of previous chemotherapy regimens 
 ≤3 48% 45% 47% 
 >3 52% 55% 53% 
Number of previous chemotherapy regimens for metastatic disease 
 ≤3 77% 71% 75% 
 >3 23% 29% 25% 
Prior regimen 
 Taxanes 99% 99% 99% 
 Anthracyclines 99% 98% 99% 
 Capecitabine 73% 74% 73% 
 Trastuzumab or lapatinib (if HER2+)a 81% 88% 83% 
 Radiotherapy 83% 77% 81% 
Chemotherapy refractoryb 
 Taxanes 81% 80% 81% 
 Anthracyclines 56% 61% 58% 
 Capecitabine 67% 69% 68% 
ParameterEribulin (n = 508)Control (n = 254)Total (N = 762)
Number of previous chemotherapy regimens 
 ≤3 48% 45% 47% 
 >3 52% 55% 53% 
Number of previous chemotherapy regimens for metastatic disease 
 ≤3 77% 71% 75% 
 >3 23% 29% 25% 
Prior regimen 
 Taxanes 99% 99% 99% 
 Anthracyclines 99% 98% 99% 
 Capecitabine 73% 74% 73% 
 Trastuzumab or lapatinib (if HER2+)a 81% 88% 83% 
 Radiotherapy 83% 77% 81% 
Chemotherapy refractoryb 
 Taxanes 81% 80% 81% 
 Anthracyclines 56% 61% 58% 
 Capecitabine 67% 69% 68% 

aPercentage based on the total number of patients with HER2+ status.

bRefractory defined as progression within 6 months of receipt of chemotherapy.

Among patients randomized to the control arm, 97% of patients received single-agent chemotherapy, and 3% of patients received hormone therapy. No patients received supportive care only. The majority of patients received vinorelbine (26%), gemcitabine (18%), capecitabine (18%), a taxane (16%), or an anthracycline (9%). Patients randomized to receive eribulin received a median of 5 cycles (range 1–23 cycles) of therapy and had a longer median duration of exposure (118 days) compared with patients in the control arm (63 days).

The magnitude of the treatment effect on OS is summarized in Table 3. The final analysis of OS, conducted after 422 events, showed a statistically significant improvement in OS in patients randomized to receive eribulin compared with those randomized to the control arm. The median OS was 75 days longer in the eribulin arm compared with the control arm.

Table 3.

Randomized trial efficacy results: study 305

Efficacy variableEribulin (n = 508)Control arm (n = 254)
OS 
 Number of deaths (%) 274 (53.9) 148 (58.3) 
 Median, months (95% CI) 13.1 (11.8, 14.3) 10.6 (9.3, 12.5) 
 HR (95% CI)a 0.81 (0.66, 0.99)  
 Pb 0.041  
PFSc 
 Number of events (%) 357 (70.3) 164 (64.6) 
 Median, months (95% CI) 3.7 (3.3, 3.8) 2.2 (2.1, 3.4) 
 HR (95% CI)a 0.87 (0.71, 1.05)  
 Pb 0.137  
Objective response rate 
 CR+PR (%) 57 (11.2) 10 (3.9) 
 Nominal Pd 0.0006  
 CR 3 (0.6) 
 PR 54 (10.6) 10 (3.9) 
Efficacy variableEribulin (n = 508)Control arm (n = 254)
OS 
 Number of deaths (%) 274 (53.9) 148 (58.3) 
 Median, months (95% CI) 13.1 (11.8, 14.3) 10.6 (9.3, 12.5) 
 HR (95% CI)a 0.81 (0.66, 0.99)  
 Pb 0.041  
PFSc 
 Number of events (%) 357 (70.3) 164 (64.6) 
 Median, months (95% CI) 3.7 (3.3, 3.8) 2.2 (2.1, 3.4) 
 HR (95% CI)a 0.87 (0.71, 1.05)  
 Pb 0.137  
Objective response rate 
 CR+PR (%) 57 (11.2) 10 (3.9) 
 Nominal Pd 0.0006  
 CR 3 (0.6) 
 PR 54 (10.6) 10 (3.9) 

Abbreviations: CR, complete response; PR, partial response.

aBased on Cox proportional hazards model stratified by geographic region, HER2 status, and prior capecitabine therapy.

bBased on a log-rank test stratified by geographic region, HER2 status, and prior capecitabine therapy.

cBased on independent reviewer assessment.

dBased on Fisher exact test.

The FDA requested the results of an unplanned, updated analysis of OS after ∼75% of patient deaths occurred. The FDA's analysis of these data yielded results that were consistent with the prespecified analysis of OS [HR 0.81 (95% CI, 0.68–0.96), nominal P = 0.014, stratified log-rank test; Fig. 1, confirming the stability of the treatment effect].

Figure 1.

Kaplan–Meier analysis of updated survival data (study 305).

Figure 1.

Kaplan–Meier analysis of updated survival data (study 305).

Close modal

The statistical analysis plan did not specify a method for controlling the overall false-positive rate for secondary endpoints. The HR for PFS as determined by independent reviewer assessment was 0.87 (95% CI, 0.71–1.05), favoring the eribulin arm; however, this effect did not reach statistical significance. The objective response rate by independent review was higher in the eribulin arm compared with the control arm (11.2% vs. 3.9%; nominal P = 0.0006). This response rate was similar to the response rate observed in 2 single-arm trials in patients with relapsed or refractory metastatic breast cancer. In studies 201 and 211, the objective response rates by independent review were 11.5% (95% CI, 5.7%–20.1%) and 9.3% (95% CI, 6.1%–13.4%) with median durations of response of 171 and 126 days, respectively.

Safety results

The primary analysis of safety used data from the single randomized trial (study 305). In addition, pooled safety data from 1,222 eribulin-treated patients with multiple tumor types, including 827 breast cancer patients treated at the proposed dose and schedule of eribulin (1.4 mg/m2 i.v. on days 1 and 8 of a 21-day cycle), were analyzed and compared with the primary safety analysis from study 305. No additional safety signals were identified from this pooled analysis.

Fewer eribulin-treated patients died during the study period (53.9%) compared with patients who received control-arm monotherapy (57.9%). In addition, more patients died within 30 days of the last dose of study drug in the control group (7.7%) compared with the eribulin group (4.0%). The majority of patient deaths were attributed to progressive disease in both treatment groups; investigators attributed ∼3% of deaths in each treatment group to causes other than disease progression. Among eribulin-treated patients, 4 died of complications from febrile neutropenia or infections within 30 days of the last dose of study drug, compared with 3 patients who received another single-agent therapy.

Table 4 shows adverse reactions that occurred in at least 10% of patients in either treatment group. The most common adverse reactions in eribulin-treated patients were neutropenia (82%), anemia (58%), asthenia/fatigue (54%), alopecia (45%), peripheral neuropathy (35%), nausea (35%), and constipation (25%). The most common serious adverse events among eribulin-treated patients were febrile neutropenia (4%) and neutropenia (2%). Two eribulin-treated patients (0.4%) died of complications of febrile neutropenia. The incidences of grade 3 and grade 4 neutropenia were 28% (143/503) and 29% (144/503), respectively, in the eribulin group. The mean time to neutrophil count nadir was 13 days, and the mean time to recovery from severe neutropenia (<500/μL) was 8 days. Dose reduction or discontinuation due to neutropenia was required in 12% (62/503) and <1% of eribulin-treated patients, respectively.

Table 4.

Per-patient incidence of adverse reactions occurring in ≥10% of patients

Eribulin (n = 503)Control (n = 247)
Body system/adverse eventaAll grades≥ Grade 3All grades≥ Grade 3
Blood and lymphatic system disordersb 
 Neutropenia 82% 57% 53% 23% 
 Anemia 58% 2% 55% 4% 
Nervous system disorders 
 Peripheral neuropathyc 35% 8% 16% 2% 
 Headache 19% <1% 12% <1% 
General disorders and administrative site conditions 
 Asthenia/fatigue 54% 10% 40% 11% 
 Mucosal inflammation 9% 1% 10% 2% 
 Pyrexia 21% <1% 13% <1% 
Gastrointestinal disorders 
 Constipation 25% 1% 21% 1% 
 Diarrhea 18% 18% 
 Nausea 35% 1% 28% 3% 
 Vomiting 18% 1% 18% 1% 
Musculoskeletal and connective tissue disorders 
 Arthralgia/myalgia 22% <1% 12% 1% 
 Back pain 16% 1% 7% 2% 
 Bone pain 12% 2% 9% 2% 
 Pain in extremity 11% 1% 10% 1% 
Investigations 
 Weight decrease 21% 1% 14% <1% 
Metabolism and nutrition disorders 
 Anorexia 20% 1% 13% 1% 
Respiratory, thoracic, and mediastinal disorders 
 Cough 14% 9% 
 Dyspnea 16% 4% 13% 4% 
Skin and subcutaneous tissue disorders 
 Alopecia 45% NAd 10% NAd 
Infections and infestations 
 Urinary tract infection 10% 1% 5% 
Eribulin (n = 503)Control (n = 247)
Body system/adverse eventaAll grades≥ Grade 3All grades≥ Grade 3
Blood and lymphatic system disordersb 
 Neutropenia 82% 57% 53% 23% 
 Anemia 58% 2% 55% 4% 
Nervous system disorders 
 Peripheral neuropathyc 35% 8% 16% 2% 
 Headache 19% <1% 12% <1% 
General disorders and administrative site conditions 
 Asthenia/fatigue 54% 10% 40% 11% 
 Mucosal inflammation 9% 1% 10% 2% 
 Pyrexia 21% <1% 13% <1% 
Gastrointestinal disorders 
 Constipation 25% 1% 21% 1% 
 Diarrhea 18% 18% 
 Nausea 35% 1% 28% 3% 
 Vomiting 18% 1% 18% 1% 
Musculoskeletal and connective tissue disorders 
 Arthralgia/myalgia 22% <1% 12% 1% 
 Back pain 16% 1% 7% 2% 
 Bone pain 12% 2% 9% 2% 
 Pain in extremity 11% 1% 10% 1% 
Investigations 
 Weight decrease 21% 1% 14% <1% 
Metabolism and nutrition disorders 
 Anorexia 20% 1% 13% 1% 
Respiratory, thoracic, and mediastinal disorders 
 Cough 14% 9% 
 Dyspnea 16% 4% 13% 4% 
Skin and subcutaneous tissue disorders 
 Alopecia 45% NAd 10% NAd 
Infections and infestations 
 Urinary tract infection 10% 1% 5% 

aAdapted from MedDRA version 10.0.

bBased on laboratory data.

cIncludes peripheral neuropathy, neuropathy, peripheral motor neuropathy, polyneuropathy, peripheral sensory neuropathy, and paresthesia.

dNot applicable (grading system does not specify > grade 2 for alopecia).

Among the 503 patients who received eribulin in study 305, a new or worsening peripheral neuropathy occurred in 174 (34.6%), and grade 3 and 4 peripheral neuropathy occurred in 39 (7.8%) and 2 (0.4%), respectively. Peripheral neuropathy was the most common toxicity that led to discontinuation of eribulin (5% of patients; 24/503). Additionally, peripheral motor neuropathy occurred in 20 patients (4%). Peripheral neuropathies were reversible in many, but not all, patients. Neuropathy lasting more than 1 year occurred in 5% of patients (26/503).

Based on the clinically meaningful and statistically significant improvement in OS observed in study 305, the FDA granted full approval to eribulin on November 15, 2010. The FDA considers OS to be the most reliable approval endpoint, because improved OS encompasses both safety and efficacy (deaths are caused by toxicity or progressive disease, or both). Furthermore, OS is an unequivocal, direct measurement of clinical benefit (20).

The improvement in OS observed in patients randomized to receive eribulin in study 305 highlights the continued role of cytotoxic therapy in the treatment of women with advanced breast cancer. In addition, the results of study 305 illustrate that improvements in OS can be shown in clinical trials evaluating treatments for advanced breast cancer.

Eribulin is the first drug to receive initial approval from the FDA for the treatment of refractory metastatic breast cancer based on a demonstrated improvement in OS. The FDA granted regular approval for docetaxel monotherapy in 1998, and for capecitabine plus docetaxel in 2001, based on demonstrated improvements in OS in the second-line metastatic breast cancer setting. However, accelerated approval had been granted to both docetaxel and capecitabine prior to receipt of full approval based on the response rate in single-arm trials (21).

A primary consideration in the FDA's review of this application was whether the results of a single adequate and well-controlled trial were sufficient to support approval. The FDA's Guidance for Industry, entitled “Providing Clinical Evidence of Effectiveness for Human Drug and Biological Products,” states that reliance on a single study should be limited to situations in which a trial “has demonstrated a clinically meaningful effect on mortality, irreversible morbidity, or prevention of a disease with potentially serious outcome and confirmation of the result in a second trial would be practically or ethically impossible” (22).

The Guidance also outlines several characteristics that are considered desirable for a single study to have to support approval. As a large, randomized, multicenter trial that showed consistent results across most patient subsets, study 305 exhibited many of these desirable characteristics. Furthermore, the demonstration of a clinically meaningful and statistically significant improvement in OS in a refractory population with limited treatment options rendered the conduct of a second confirmatory randomized control trial practically unfeasible and potentially unethical.

The FDA carefully considered a number of factors to determine whether study 305 provided sufficient evidence of clinical efficacy to support approval. The review team noted that the results of study 305, with a P-value of 0.041, were not highly persuasive. However, the updated survival analysis reinforced the notion that the improvement in OS was likely to be a true effect. The application was further strengthened by supportive evidence of antitumor activity in study 305, including improvements in the overall response rate. In addition, similar durable objective tumor responses in a comparable patient population were shown in 2 supportive single-arm studies.

The FDA noted that the analysis of PFS as determined by an independent review panel did not reach statistical significance, but it concluded that approval of the NDA for eribulin was justified based on the OS results, supportive objective response rate data, and consistent trend in PFS (favoring the eribulin arm). To explore potential reasons for the discordant strength of the PFS and OS results, the FDA conducted an analysis to determine whether there was a differential approach to therapy between the treatment and control arms that might have contributed to the improvement in survival in patients randomized to receive eribulin. This analysis did not uncover an imbalance in the numbers of patients who went on to receive subsequent cytotoxic, biologic, or supportive care therapy between the arms. Of the patients who discontinued study therapy for reasons other than death, 68% of patients in the eribulin arm and 62% of patients in the control arm subsequently received chemotherapy. In addition, 13% of patients in the eribulin arm and 18% of patients in the control arm subsequently received hormone therapy.

During the review, the FDA also considered the appropriateness of the use of a “physician's choice” design for the control arm. When a placebo or best-supportive-care control arm is considered unethical or unfeasible, studies may use an add-on design, active control design with a single comparator, or a physician's choice design. When there is a well-recognized standard of care (e.g., first-line therapy for Hodgkin lymphoma), the clear choice is to compare the experimental therapy with the standard-of-care therapy (e.g., with an add-on design). Because there is no obvious single standard of care for the treatment of metastatic breast cancer in the third-line setting, the use of a physician's choice design for the control arm in this case was considered reasonable.

A primary advantage of using a physician's choice design for the control arm is that this design reflects the real-world choices made by physicians and their patients. Because investigators were required to choose control therapy prior to randomization, the potential for bias in the choice of therapies was reduced. However, due to the heterogeneity of the therapies used in the control arm, a careful analysis to confirm the acceptability of the control arm was an essential part of the FDA's review of this application. To establish that the results of study 305 supported approval of eribulin, the FDA evaluated the choice of therapy for each patient in the control arm, taking into account prior treatment and response, to determine whether the control therapy reflected standard U.S. medical practice. In addition, the FDA also evaluated whether the choice of control therapies resulted in potential bias by increasing the patients' risk of death due to the use of potentially ineffective or unsafe therapy. Ultimately, the FDA reviewers concluded that the composition of therapies administered to patients in the control arm adequately represented U.S. practice and that the control arm was therefore an acceptable comparator.

The heterogeneity of the therapies used in the control arm also complicated the safety analysis. To determine whether there was a reasonable likelihood that adverse events were caused by eribulin, the reviewers had to conduct careful analyses comparing the adverse events observed in the eribulin group with those observed in the control group in aggregate, and in patients who received each subtype of chemotherapy. Because of the challenges inherent in analyses of a study using a physician's choice control arm, the FDA recommends that sponsors meet with the FDA to discuss study design prior to conducting studies intended to support a marketing authorization.

The prolongation of OS observed in study 305 reflects both the safety and efficacy of eribulin therapy. In general, the adverse reactions observed in eribulin-treated patients were typical of those that occur with other microtubule-inhibiting drugs, such as vinca alkaloids and taxanes (i.e., myelosuppression and peripheral neuropathy). Other commonly observed toxicities were generally not life threatening and were manageable with supportive care and dose adjustment.

The majority of patients (57%) experienced grade 3 or greater neutropenia during study 305. Patients with impaired hepatic function appeared to have an increased risk of myelosuppression. In an analysis of breast cancer patients treated with eribulin in multiple clinical trials, patients with alanine aminotransferase or aspartate aminotransferase >3 times the upper limit of normal, or with bilirubin >1.5 times the upper limit of normal, experienced a higher incidence of grade 4 neutropenia than patients with normal liver function (19). The study 305 protocol permitted the use of colony-stimulating factors; however, the role of colony-stimulating factors in mitigating the risk of infection with eribulin therapy remains undefined because the use of these drugs was variable during the trial. The low incidence of fatal infections suggests that the use of dose modification combined with appropriate surveillance and prompt institution of antimicrobial therapy in patients with signs of infection mitigated the risk of life-threatening infection conferred by neutropenia.

Peripheral neuropathy occurred frequently after eribulin treatment. An applicant-conducted analysis showed that the incidence of peripheral neuropathy increased from 13% in subjects with a cumulative exposure of ≤5.6 mg/m2 (≤2 cycles) to >40% in subjects with a cumulative exposure >11.2 mg/m2 (>4 cycles). Although in the majority of cases the peripheral neuropathy was mild or moderate in severity and responded to dose delay or reduction, it was of long duration in some patients. Analyses of duration of eribulin-induced neuropathy were complicated by the lack of follow-up of adverse events after 30 days elapsed from discontinuation of study therapy and initiation of other anticancer therapy prior to resolution of neuropathy. Therefore, definitive conclusions regarding the duration of peripheral neuropathies caused by eribulin cannot be made. The label contains a warning to withhold eribulin in patients who experience grade 3 or 4 peripheral neuropathy until resolution to grade 2 or less occurs.

Although the results of an open-label study indicated that there is a risk of delayed QT prolongation with eribulin use, cardiotoxicity and sudden death related to eribulin were not observed in clinical trials of eribulin. Nevertheless, the label recommends avoidance of eribulin in patients with long QT syndrome, and recommends electrocardiogram monitoring if eribulin is administered to patients receiving other medications known to prolong the QT interval, or patients with congestive heart failure or bradyarrhythmia.

The FDA identified one postmarketing requirement during the NDA review. To satisfy this requirement, Eisai will conduct a dedicated clinical trial to assess the safety and pharmacokinetics of eribulin in patients with impaired renal function, which could support specific dosing recommendations in this subpopulation. In addition, Eisai agreed to several postmarketing commitments, including the submission of a final report for study 305 with updated results for OS after 95% of patient deaths have occurred. In addition, Eisai agreed to submit a final report for an ongoing randomized trial comparing eribulin with capecitabine in patients with locally advanced or metastatic breast cancer previously treated with anthracyclines and taxanes.

The approval of eribulin mesylate shows that in the setting of metastatic breast cancer, OS is an achievable endpoint and may also be more reliable and expeditious for showing a meaningful treatment effect. The FDA does not require the use of independent review of radiologic assessments of tumor progression in trials that use OS as the primary endpoint. Additionally, compared with interpretation of OS, it is more challenging to determine whether an improvement in PFS represents clinical benefit, in part because this requires a judgment about whether the magnitude of the treatment effect is meaningful for patients in light of the toxicities of that treatment. Finally, PFS does not appear to be a reliable surrogate for OS in metastatic breast cancer. A recent FDA analysis of data from 12 trials submitted to support approval in metastatic breast cancer failed to show an association between PFS and OS (23). Therefore, investigators should consider the use of OS as a primary endpoint in clinical trials investigating the use of drugs in late-stage diseases such as metastatic breast cancer.

No potential conflicts of interest were disclosed.

1.
Vahdat
LT
,
Pruitt
B
,
Fabian
CJ
,
Rivera
RR
,
Smith
DA
,
Tan-Chiu
E
, et al
Phase II study of eribulin mesylate, a halichondrin B analog, in patients with metastatic breast cancer previously treated with an anthracycline and a taxane
.
J Clin Oncol
2009
;
27
:
2954
61
.
2.
Cortes
J
,
Vahdat
L
,
Blum
JL
,
Twelves
C
,
Campone
M
,
Roché
H
, et al
Phase II study of the halichondrin B analog eribulin mesylate in patients with locally advanced or metastatic breast cancer previously treated with an anthracycline, a taxane, and capecitabine
.
J Clin Oncol
2010
;
28
:
3922
8
.
3.
Twelves
C
,
Cortes
J
,
Vahdat
LT
,
Wanders
J
,
Akerele
C
,
Kaufman
PA
. 
Phase III trials of eribulin mesylate (E7389) in extensively pretreated patients with locally recurrent or metastatic breast cancer
.
Clin Breast Cancer
2010
;
10
:
160
3
.
4.
Cortes
J
,
O'Shaughnessy
J
,
Loesch
D
,
Blum
JL
,
Vahdat
LT
,
Petrakova
K
, et al
EMBRACE (Eisai Metastatic Breast Cancer Study Assessing Physician's Choice Versus E7389) investigators. Eribulin monotherapy versus treatment of physician's choice in patients with metastatic breast cancer (EMBRACE): a phase 3 open-label randomised study
.
Lancet
2011
;
377
:
914
23
.
5.
National Cancer Institute [homepage on the Internet]
.
Altekruse
SF
,
Kosary
CL
,
Krapcho
M
,
Neyman
N
,
Aminour
R
,
Waldron
W
, et al
editors
.
SEER Cancer Statistics Review, 1975–2007
. 
2010
.
[cited 2011 Dec 30] Available from
: http://seer.cancer.gov.
6.
Centers for Disease Control and Prevention [homepage on the Internet]. U.S. Cancer Statistics Working Group
. 
United States Cancer Statistics: 1999–2007 incidence and mortality web-based report
. 
2010
.
[cited 2012 Jan 11] Available from: http://www.cdc.gov/uscs
.
7.
Moreno-Aspitia
A
,
Perez
EA
. 
Anthracycline- and/or taxane-resistant breast cancer: results of a literature review to determine the clinical challenges and current treatment trends
.
Clin Ther
2009
;
31
:
1619
40
.
8.
Hirata
Y
,
Uemura
D
. 
Halichondrins—antitumor polyether macrolides from a marine sponge
.
Pure Appl Chem
1986
;
58
:
701
10
.
9.
Lobert
S
,
Vulevic
B
,
Correia
JJ
. 
Interaction of vinca alkaloids with tubulin: a comparison of vinblastine, vincristine, and vinorelbine
.
Biochemistry
1996
;
35
:
6806
14
.
10.
Kingston
DG
. 
Tubulin-interactive natural products as anticancer agents
.
J Nat Prod
2009
;
72
:
507
15
.
11.
Smith
JA
,
Wilson
L
,
Azarenko
O
,
Zhu
X
,
Lewis
BM
,
Littlefield
BA
, et al
Eribulin binds at microtubule ends to a single site on tubulin to suppress dynamic instability
.
Biochemistry
2010
;
49
:
1331
7
.
12.
Jordan
MA
,
Kamath
K
,
Manna
T
,
Okouneva
T
,
Miller
HP
,
Davis
C
, et al
The primary antimitotic mechanism of action of the synthetic halichondrin E7389 is suppression of microtubule growth
.
Mol Cancer Ther
2005
;
4
:
1086
95
.
13.
Kotch
L
,
Pilaro
A
. 
Pharmacology review application number: 201532
. 
2010
.
14.
Synold
TW
,
Morgan
RJ
,
Newman
EM
,
Lenz
HJ
,
Gandara
DR
,
Colevas
AD
, et al
A phase I pharmacokinetic and target validation study of the novel anti-tubulin agent E7389: a California Cancer Consortium trial
.
J Clin Oncol
23
:
16S
, 
2005
(
suppl; abstr 3036
).
15.
Tan
AR
,
Rubin
EH
,
Walton
DC
,
Shuster
DE
,
Wong
YN
,
Fang
F
, et al
Phase I study of eribulin mesylate administered once every 21 days in patients with advanced solid tumors
.
Clin Cancer Res
2009
;
15
:
4213
9
.
16.
Goel
S
,
Mita
AC
,
Mita
M
,
Rowinsky
EK
,
Chu
QS
,
Wong
N
, et al
A phase I study of eribulin mesylate (E7389), a mechanistically novel inhibitor of microtubule dynamics, in patients with advanced solid malignancies
.
Clin Cancer Res
2009
;
15
:
4207
12
.
17.
Ledford
H
. 
Complex synthesis yields breast-cancer therapy
.
Nature
2010
;
468
:
608
9
.
18.
Shord
S
,
Marathe
A
,
Zhao
H
. 
Clinical pharmacology and biopharmaceutics reviews NDA 210532
. 
2010
.
19.
Halaven (eribulin mesylate) package insert
.
Andover (MA)
:
Esai
; 
2011
.
20.
U.S. Department of Health and Human Services
. 
Guidance for industry. Clinical trial endpoints for the approval of cancer drugs and biologics
.
Washington, DC
:
U.S. Dept. of Health and Human Services
; 
2011
.
21.
Johnson
JR
,
Williams
G
,
Pazdur
R
. 
End points and United States Food and Drug Administration approval of oncology drugs
.
J Clin Oncol
2003
;
21
:
1404
11
.
22.
U.S. Department of Health and Human Services
. 
Guidance for industry: providing clinical evidence of effectiveness for human drug and biological products
. 
2011
.
23.
Cortazar
P
,
Zhang
JJ
,
Sridhara
R
,
Justice
RL
,
Pazdur
R
. 
Relationship between OS and PFS in metastatic breast cancer (MBC): review of FDA submission data
.
J Clin Oncol
29
: 
2011
(
suppl; abstr 1035
).