Impact of the number of prior chemotherapy regimens on outcomes for patients with metastatic breast cancer treated with eribulin: A post hoc pooled analysis

Abstract In a pivotal phase 3 study (Study 305), eribulin mesylate improved overall survival (OS) in patients with previously treated metastatic breast cancer (MBC) compared with treatment of physician's choice (TPC). This post hoc, pooled subgroup analysis of two phase 3 studies (Study 305 and Study 301) reports the influence of the number of prior chemotherapy regimens (0‐6) on OS in patients with locally advanced/MBC randomized to eribulin versus TPC/capecitabine. Patients with ≤ 3 prior chemotherapies for locally advanced/MBC had longer median OS with eribulin (15.3 months) versus control (13.2 months; hazard ratio, 0.858; P = .01).

(Study 301) in the intent-to-treat (ITT) populations. 5,6 HRs for EMBRACE only were based on a Cox regression model including HER2/neu status, geographical region, and prior capecitabine treatment as stratification variables. 5 HRs for pooled data were estimated based on the Cox model with stratification factors (HER2 status, region, prior capecitabine use, and study). Median OS was adjusted by study (defined in Twelves et al, 2014 7

) and P-values
were estimated by stratified log-rank test. An exploratory comparative analysis of OS grouped by ≤3 versus >3 prior treatments, and by individual number of prior lines of treatment (ie, 0, 1, 2, 3, 4, 5, and 6), for locally advanced/MBC was completed using data pooled from both studies except as noted (ie, data on ≥5 prior lines of therapy are from EMBRACE only). A pooled analysis of safety data was not possible because the studies used different versions of the Medical Dictionary for Regulatory Activities (version 10.0 for EMBRACE, version 14.1 for Study 301).

| Patients
In EMBRACE, patients were randomized 2:1 to receive eribulin (1.4 mg/m 2 [equivalent to 1.23 mg/m 2 when expressed as a free base] intravenously on days 1 and 8 every 21 days; n = 508) or TPC (n = 254). 5 In Study 301, 554 patients were randomized to receive eribulin and 548 to receive capecitabine. 6 Patient characteristics have been previously reported. 5,6 Almost all (99%) patients had received prior anthracycline and taxane therapy. 5,6 In EMBRACE, the median number of prior chemotherapy regimens for locally advanced/MBC was 3 (with approximately one-quarter having >3 and three-quarters having ≤3). In Study 301, only 1 patient (a protocol deviation) received >3 prior chemotherapy regimens for locally advanced/MBC.
Additional exploratory pooled post hoc analysis for patients receiving 0-6 prior lines of therapy showed a trend for higher OS in patients treated with eribulin compared with control (Table 2), and this trend was emphasized in those patients with 0-3 prior lines of therapy compared with those who had been more heavily pretreated. However, caution is warranted due to the low number of patients, especially in the latter subgroups.

| Safety
The number of prior chemotherapies appeared not to affect the safety of eribulin in EMBRACE. Although neutropenia and asthenia/ fatigue rates were higher with eribulin treatment compared with control, the incidences of both were similar regardless of whether

| D ISCUSS I ON
This exploratory subgroup analysis of EMBRACE 5 and Study 301 6 shows that the OS benefit conferred by eribulin over TPC/capecitabine is predominantly seen in patients who had fewer prior regimens (≤3) for locally advanced/MBC with a median OS benefit of 2.1 months.
This difference in OS was also observed in EMBRACE alone (≤3 prior regimens, 2.6 months; >3 prior regimens, 1.7 months); the number of prior regimens appeared not to affect the safety of eribulin.