Effectiveness of Eribulin in Metastatic Breast Cancer Patients With Poor Prognostic Indicators in the United States

J Clin Pathways. 2021;7(5):33-41. doi:10.25270/jcp.2021.06.00001

Survival rates for metastatic breast cancer (MBC) have improved over the past decade due to advances in therapy, but these improvements have not been observed equally across all age groups, races, and other clinical cohorts such as patients with liver metastases.^1,2 An analysis of data from the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) Program found that patients aged 70 years or older experienced significantly shorter overall survival (OS) compared to any other younger age cohorts.³ Comorbidities, poor functional status, and undertreatment are associated with worse outcomes among older patients.⁴ Data from SEER also showed that non-Hispanic African American women have a higher incidence of, and mortality from, MBC compared with other races/ethnicities.⁴ Decreased survival among African American patients has been attributed to socioeconomic factors including access to care, increased prevalence of certain tumor subtypes such as triple negative breast cancer (TNBC), and disparities in treatment.⁵ In addition, liver metastases develop in approximately half of women with MBC and is associated with relatively lower OS compared to those without, resulting in 5-year survival rates of less than 10%.^6-8 As older patients (aged ≥65 years) and African American patient subgroups are underrepresented in MBC clinical trials, the effectiveness of therapies for real-world patients in these poor prognosis subgroups is less well understood.^9,10 Furthermore, understanding the effectiveness of treatments in specific clinical poor risk subgroups (such as those with liver metastases) may lead to further enhancements in delivering patient-centered care.^11,12 

Eribulin was approved in the United States 10 years ago for the treatment of patients with MBC who have previously received an anthracycline and a taxane in either the adjuvant or metastatic setting and at least two chemotherapeutic regimens for the treatment of metastatic disease.¹³ This approval was based on findings from the EMBRACE trial, in which eribulin significantly improved OS (median OS 13.1 months; 95% CI: 11.8-14.3) over treatment with physician’s choice (median OS 10.6 months, 95% CI: 9.3-12.5; hazard ratio 0.81, 95% CI: 0.66-0.99; P=.041) in women who had received 2 to 5 prior lines of therapy.¹⁴ In the EMBRACE trial, 58% of eribulin-treated patients had liver metastases, the median age of the patient population was 55 years, and only 4% of patients were African Americans.¹⁴ Prior published real-world data (RWD) has demonstrated the effectiveness of eribulin in clinical practice, however, RWD on the effectiveness of eribulin in poor prognosis subgroups is lacking.^15-21 Using data collected from a large retrospective study, we examined clinical outcomes for United States patients treated with eribulin therapy (according to the approved indication in clinical practice in the US) in three non-mutually exclusive subgroups that generally have poorer prognoses: older patients (aged ≥65 years), African American patients, and patients with liver metastases. 

A retrospective, multisite patient chart review was conducted.²² Study materials were approved by an independent, central Institutional Review Board, and a waiver was granted for patient consent. Physicians from the Cardinal Health Oncology Provider Extended Network, a group of community oncologists/hematologists from across the United States, abstracted de-identified data from patients’ medical charts who met the study selection criteria. Physicians independently performed the data extraction and were asked to identify and randomly select patients who met the selection criteria. All abstracted data was de-identified in accordance with Health Insurance Portability and Accountability Act (HIPAA). 

The study was limited to patients treated with eribulin in accordance with the approved indication in the United States: female adults with MBC who were aged 18 years or older at the time of initiation of eribulin, had been treated with both an anthracycline and a taxane therapy in either the adjuvant or metastatic setting, and had received at least two chemotherapy regimens in the metastatic setting prior to treatment with eribulin. Patients must have initiated eribulin treatment between January 1, 2011 and December 31, 2017. Patients receiving eribulin treatment for a primary diagnosis other than MBC were excluded.

Measured clinical outcomes included objective response rate (ORR), clinical benefit rate (CBR), progression-free survival (PFS) and OS. The ORR and CBR were calculated based on best response to eribulin therapy, which were evaluated via two methods: (1) recorded in the patient’s electronic health record by their physician as complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD); and (2) calculated retrospectively per Response Evaluation Criteria in Solid Tumors (RECIST) v1.1 using measurements of target lesions provided by the treating oncologist.²³ The ORR was calculated as: (CR+PR) / (all patients). The CBR was calculated as: (CR+PR+SD lasting more than 24 weeks) / (all patients). Physicians indicated if the duration of SD lasted more than 24 weeks after the initiation of eribulin (date of first infusion received) as recorded in patient’s medical chart. Median PFS and OS were calculated from the date of initiation of eribulin using the Kaplan-Meier method. Patients still receiving eribulin or alive at the time of data cut-off were censored on their last recorded clinic visit date. Additionally, landmark PFS estimates (%) at 3-, 6-, 9-, and 12-months and landmark OS estimates (%) at 3-, 6-, 9-, 12-, 24- and 36-months and 95% CIs are reported for each subgroup. Analyses were conducted for each subgroup independently and the subgroups are not mutually exclusive. All analyses were conducted in SAS v9.4 (SAS Institute Inc, Cary, NC, US).

Older Adults (Aged ≥65 Years) Subgroup

Demographics and Clinical Characteristics

A total of 169 patients aged ≥65 years (32.9% of all patients) were included in the analysis. Demographics are shown in Table 1. The average age of patients in this subgroup was 70.3 years at initiation of eribulin, and 20.7% were African American.
Eribulin was third-line therapy for 81.7% of patients while the remainder received eribulin in fourth line or greater. Almost half (45.6%) had impaired physical function with Eastern Cooperative Oncology Group performance status (ECOG-PS) ≥2. Over half of patients (58%) were of the hormone receptor-positive/human epidermal growth factor receptor-negative (HR+/HER2-) subtype and 38.5% had TNBC (Table 2).

Duration of Eribulin Therapy and Time to Next Treatment

At the time of data cut-off, 95.3% of patients had discontinued eribulin. Among patients who discontinued, median duration of eribulin therapy was 5.0 months (interquartile range [IQR]: 5.1). The most common rationale for discontinuation of eribulin as reported by the patient’s physician based on the medical record was disease progression, accounting for 80.1% of patients (Table 2). Discontinuation due to toxicity/intolerability was reported for 1 patient. Dose reduction was observed in 13.6% (total n=23) of patients due to adverse events (n=17) and patient choice (n=6). Another line of therapy was initiated by 20.7% of patients post eribulin with a median time to next treatment of 6.6 months (IQR: 6.3).

Clinical Outcomes

ORRs based on physician report are shown in Figure 1. The ORR among patients aged ≥65 years was 48.5% (95% CI: 40.8%-56.3%) and the CBR was 51.5% (95% CI: 43.7%-59.2%). RECIST-based disease response reported an ORR in 37.9% of patients and CBR in 40.8% of patients. Based on physician report, a total of 5.3% of patients had a CR, while a PR was reported in 43.2% of patients. Median time to best response among these patients was 2.7 months (IQR: 1.6). Duration of tumor response (time from initial response until progression) among these patients who had progressed by the time of data cut-off was a median of 5.2 months (IQR: 4.7).At the time of data cut-off, 156 (92.3%) patients had progressed on eribulin therapy and 143 (84.6%) were deceased. 

Landmark PFS rates were 47.0% (95% CI: 39.2%-54.3%) at 6 months and 17.6% (95% CI: 12.1%-23.8%) at 12 months (Figure 2). Median PFS was 6.0 months (95% CI: 5.0-6.6) as shown in Figure 3. Landmark OS rates were 36.9% (95% CI: 29.6%-44.3%) at 12 months and 18.8% (95% CI: 13.2%-25.2%) at 24 months (Figure 4). Median OS was 9.2 months (95% CI: 8.1-10.5) as shown in Figure 5.

Duration of Eribulin Therapy and Time to Next Treatment

At the time of data cut-off, almost all (99.0%) patients with liver metastases had discontinued eribulin, with the median duration of eribulin therapy of 4.8 months (IQR: 4.8). Disease progression was the most common rationale for discontinuation (81.4%), while discontinuation due to toxicity/intolerability was observed in 3 patients (Table 2). Dose reduction was needed in 13.3% (total n=39) of patients due to adverse events (n=33), patient choice (n=5), or drug holiday (n=1). A total of 33.7% of patients with liver metastases initiated another line of therapy post eribulin, and median time to next treatment among these patients was 6.9 months (IQR: 5.4).

Clinical Outcomes

ORRs based on physician report are shown in Figure 1. The ORR was 45.9% (95% CI: 40.1%-51.8%) in patients with liver metastases including 4.4% with a CR and 41.5% with a PR; CBR was 48.6% (95% CI: 42.8%-54.5%). RECIST-based disease response reported an ORR in 37.8% of patients and CBR in 40.5% of patients. Physician-reported time to best response among patients with a CR or PR was median 2.9 months (IQR: 2.0), and duration of tumor response was 4.1 months (IQR: 3.8).

At the time of data cut-off, 277 (94.2%) and 263 (89.5%) patients with liver metastases had progressed on eribulin therapy or were deceased, respectively. Landmark PFS rates were estimated to be 41.5% (95% CI: 35.8%-47.1%) and 12.8% (95% CI: 9.2%-17.1%) at 6 months and 12 months, respectively (Figure 2). Median PFS was 5.3 months (95% CI: 4.6-5.8) as shown in Figure 3. Landmark OS rates were estimated to be 34.1% (95% CI: 28.7%-39.6%) and 15.4% (95% CI: 11.5%-19.9%) at 12 months and 24 months,
respectively (Figure 4). Median OS was 9.1 months (95% CI: 8.2-9.9) as shown in Figure 5.

African American Subgroup

Demographics and Clinical Characteristics

The African American subgroup was comprised of 135 patients, representing 26.3% of all patients. Demographics are shown in Table 1. The mean age at eribulin initiation was 56.6 years and the majority of patients (74.1%) were younger than 65 years. Eribulin was third-line therapy for 80.7% of patients while the remainder received eribulin in fourth line or greater. A total of 39.3% of patients had an ECOG-PS ≥2 and a high proportion (61.5%) of patients had TNBC (Table 2).

Duration of Eribulin Therapy and Time to Next Treatment

A total of 131 (97.0%) African American patients had discontinued eribulin at data collection, with disease progression as the most common rationale for discontinuation (75.6%) (Table 2). Two patients discontinued due to toxicity/intolerability. Among patients who discontinued, median duration of eribulin therapy was 6.0 months (IQR: 5.6). A total of 11.9% (total n=16) of patients required dose reduction due to adverse events (n=11), patient choice (n=4), or drug holiday (n=1). Among the 36.3% of patients who initiated another line of therapy post eribulin, median time to next treatment was 8.7 months (IQR: 6.5).

Clinical Outcomes

Physician-reported ORRs are shown in Figure 1. The ORR among African American patients was 52.6% (95% CI: 43.9%-61.2%) and CBR was 54.8% (95% CI: 46.0%-63.3%). RECIST-based disease response reported an ORR in 47.4% of patients and CBR in 49.6% of patients. Based on physician report, a CR was observed in 7.4% of patients, while a PR was seen in 45.2%. Time to best response among patients with a CR or PR was a median of 3.3 months (IQR: 2.5), and the median duration of tumor response was 5.6 months (IQR: 4.3).

The number of patients that had progressed on eribulin therapy by data cut-off was 118 (87.4%) and 106 (78.5%) deaths were recorded. Landmark PFS rates were 55.8% (95% CI: 47.0%-63.8%) at 6 months and 18.7% (95% CI: 12.5%-26.0%) at 12 months (Figure 2), with a median PFS of 7.0 months (95% CI: 5.5-8.3; Figure 3). Landmark OS rates were 43.7% (95% CI: 35.1%-52.0%) at 12 months and 23.7% (95% CI: 16.8%-31.3%) at 24 months (Figure 4). Median OS was 10.3 months (95% CI: 9.4-13.4) as shown in Figure 5.

Liver Metastases Subgroup

Demographics and Clinical Characteristics

Demographics for the 294 patients with liver metastases (57.3% of all patients) are shown in Table 1. Patients were on average aged 58.8 years at initiation of eribulin, and eribulin was third line-therapy for 78.2% of patients, while the remainder received eribulin in fourth line or greater. In this subgroup, 41.8% of patients had an ECOG-PS ≥2 and 54.1% had TNBC (Table 2).

Duration of Eribulin Therapy and Time to Next Treatment

At the time of data cut-off, almost all (99.0%) patients with liver metastases had discontinued eribulin, with the median duration of eribulin therapy of 4.8 months (IQR: 4.8). Disease progression was the most common rationale for discontinuation (81.4%), while discontinuation due to toxicity/intolerability was observed in 3 patients (Table 2). Dose reduction was needed in 13.3% (total n=39) of patients due to adverse events (n=33), patient choice (n=5), or drug holiday (n=1). A total of 33.7% of patients with liver metastases initiated another line of therapy post eribulin, and median time to next treatment among these patients was 6.9 months (IQR: 5.4).

Clinical Outcomes

ORRs based on physician report are shown in Figure 1. The ORR was 45.9% (95% CI: 40.1%-51.8%) in patients with liver metastases including 4.4% with a CR and 41.5% with a PR; CBR was 48.6% (95% CI: 42.8%-54.5%). RECIST-based disease response reported an ORR in 37.8% of patients and CBR in 40.5% of patients. Physician-reported time to best response among patients with a CR or PR was median 2.9 months (IQR: 2.0), and duration of tumor response was 4.1 months (IQR: 3.8).

At the time of data cut-off, 277 (94.2%) and 263 (89.5%) patients with liver metastases had progressed on eribulin therapy or were deceased, respectively. Landmark PFS rates were estimated to be 41.5% (95% CI: 35.8%-47.1%) and 12.8% (95% CI: 9.2%-17.1%) at 6 months and 12 months, respectively (Figure 2). Median PFS was 5.3 months (95% CI: 4.6-5.8) as shown in Figure 3. Landmark OS rates were estimated to be 34.1% (95% CI: 28.7%-39.6%) and 15.4% (95% CI: 11.5%-19.9%) at 12 months and 24 months,
respectively (Figure 4). Median OS was 9.1 months (95% CI: 8.2-9.9) as shown in Figure 5.

Many clinical trials enroll patients who are younger, healthier, and less diverse than the real-world disease population.²⁴ To our knowledge, the current study represents the largest retrospective analyses to date of real-world outcomes in poor prognosis MBC subgroups (older adults aged ≥65 years, African Americans, and patients with liver metastases) treated with eribulin in contemporary US clinical practice consistent with the approved US indication. Eribulin was third-line therapy for nearly 80% of patients in all subgroups, with the remainder receiving the drug in fourth line or later. The prognostic factors (aged ≥65 years, African American, and the presence of liver metastases) selected for analyses are generally associated with worse outcomes,^3,4,8 however, clinical benefit was still observed in the current study. ORR was 48.5%, 52.6%, and 45.9% respectively across older adults (aged ≥65 years), African American, and patients with liver metastases subgroups. Median PFS of 6.0, 7.0, and 5.3 months were observed as well as median OS of 9.2, 10.3, and 9.1 months in the three groups, respectively. Moreover, outcomes in the subgroups were consistent with those in the overall eribulin-treated population from our study, which showed an ORR of 54.4%, PFS of 6.1 months, and OS of 10.6 months.²² We observed overlap between the patient cohorts: 21% of older patients (aged ≥65 years) and 26% of patients with liver metastases were African Americans, and 60% of older patients (aged ≥65 years) and 56% of African American patients had liver metastases at the initiation of eribulin.

 Two previous studies have evaluated the efficacy of eribulin in older patients with MBC. One study assessed patients aged 70 years and older in a post hoc analysis of two single-arm phase 2 studies and one randomized phase 3 study of eribulin.²⁵ The second was a pooled analysis incorporating four real-world studies and clinical trial data.²⁶ These studies found that age had no statistically significant impact on patients’ response to eribulin treatment or on PFS or OS.^25,26 Similarly, we found that in older patients aged ≥65 years the ORR was 48.5%, which was consistent with that observed in the overall eribulin-treated population in our study (54.4%).²² To the best of our knowledge, no previous studies have assessed the effectiveness of eribulin for MBC within the African American subgroup. In our study, African American patients experienced comparable clinical benefits to the overall study cohort, which included 64.9% Caucasian patients. Although patients with liver metastases typically have a worse prognosis than patients with non-liver metastases,⁸ a pooled analysis of three clinical trials showed that eribulin was efficacious in patients with MBC and liver metastases with a median OS of 13.4 months.²⁷ In our study, patients with liver metastases also demonstrated a satisfactory benefit from eribulin treatment. The results of this real-world study demonstrate that, in subgroups with historically poor prognosis, eribulin therapy generated a strong clinical response and provided a durable benefit when use was consistent with the approved US indication. It is important to note that, as is the case with all real-world studies, differences in patient profiles and assessment frequency between clinical practice and clinical trials may lead to differences in outcomes, precluding direct comparison of study results. 

The primary limitations of this study include the potential bias in selecting providers and patients who may not be representative of providers and patients outside of the participating practices. Moreover, while the participating providers were asked to randomly select from all eligible patients, the total number of eligible patients from each provider was unknown, thus the differences between patients included vs those not included may bias the results. We believe this bias to be limited, however, as we strictly required that all selected patients received eribulin per the US prescribing information. In addition, differences in the timepoints of assessment may lead to differences in disease response rates and time to progression estimates. 

The strengths of this study are its inclusion of patients with comorbidities or poor prognostic factors, providing updated findings on eribulin utilization and outcomes in US clinical practice within poor prognostic subgroups that may be understudied. In addition, all patients were required to have initiated eribulin more than 2 years prior to data collection, thus allowing precision in estimates of PFS and OS.

In this retrospective, real-world study of patients with MBC who had demographic and clinical characteristics typically associated with poor prognosis, outcomes were consistent with those of the broader patient population. This research demonstrates that when used in accordance with the US-approved indication, eribulin provides an effective treatment option for MBC patients with poor prognosis.

Authors: Sarah S Mougalian, MD¹; Jingchuan Zhang, PhD²; Talia Miller, MSW, MPH³; Marjorie E Zettler, PhD, MPH³; Bruce A Feinberg, DO³

Affiliations: ¹Yale Cancer Center, Yale School of Medicine, New Haven, CT 
²Eisai Inc, Woodcliff Lake, NJ 
³Cardinal Health Specialty Solutions, Dublin, OH

Correspondence: Sarah S Mougalian, MD
Yale Cancer Center
Yale School of Medicine
New Haven, CT 06510
Phone: (203) 785-4022
Email: sarah.mougalian@yale.edu

Disclosures: SSM: Eisai Inc. (consulting fees), Celgene Corporation (consulting fees), Pfizer (research funding paid directly to her institution), Genentech (research funding directly to her institution). JZ: Eisai Inc. (employee). TM: Cardinal Health Specialty Solutions (employee). MEZ: Cardinal Health Specialty Solutions (employee). BAF: Cardinal Health Specialty Solutions (employee).