In the present analysis of the efficacy of eribulin in an off-trial setting, PFS and OS in patients with heavily pretreated HER2-negative advanced breast cancer were similar to those observed in the phase III EMBRACE trial (4.1 vs 3.7 months and 11.1 vs 13.1 months, respectively) [6–8]. Patient characteristics in our study more closely resemble the patient characteristics in the EMBRACE study population than the 301 study population, as most patients had been previously treated with anthracyclines, taxanes and capecitabine [6]. In contrast, the response rate and clinical benefit in our study were higher than in the EMBRACE trial and more similar to the 301 trial results [7]. There was an increased risk of mortality in patients with TNBC compared to HR-positive disease, given the aggressiveness of this tumor subtype. In the pooled analyses, longer median OS was observed in patients with more than two organs involved and in those with visceral metastases [8]; however, in our study, more benefit and longer OS were seen in patients with lower burden of disease and fewer than two organs involved.The more favourable results obtained in these subgroups of patients may reflect their independent better prognosis [8,12,20].
In previous reports, patient age had not been described as an independent factor for PFS or OS. A pooled analysis of 827 patients treated in either the EMBRACE trial or the preceding phase II studies reported outcomes according to age. A total of 10% of patients who were included in these studies were aged ≥70 years. No significant differences were observed in ORR, PFS, or OS by age. Toxicity was similar across all age groups, although the incidence of grade 3 and 4 fatigue and peripheral neuropathy was highest in patients aged ≥70 years [14,21]. Elderly patients with good ECOG-Performance Status treated with eribulin have also shown similar OS, PFS, response rate, clinical benefit and tolerability in smaller studies when compared to younger patients [21,22]. In our cohort of patients, we observed significant differences in OS according to age, with significantly better OS and PFS in patients ≥70 years of age compared to those younger. Similar ORR but with a distinct pattern regarding PR or SD and tolerability were observed in patients in both age groups.
The role of BMI and comorbidities in the outcomes of patients treated with eribulin is still a matter of debate. In some historical series, patients who are overweight and obese have shown worse outcomes compared to the normal weight group [20,23,24]. In an Italian observational multicentric study, there was a significantly better PFS in the lowest category of BMI, but no differences in OS were observed [25]. In our cohort, no differences in OS were found amongst patients with normal weight compared to overweight and obese patients or those with and without comorbidities.
In our study, 21% of patients had CNS involvement, which is often an exclusion criterion in most clinical trials, unless patients are previously treated with local therapy. Of those patients, 25% achieved SD and 20% achieved PR, which supports the efficacy of eribulin in patients with CNS involvement. The efficacy in patients with CNS disease was not related to previous treatment with whole brain radiotherapy. The rationale for potential positive effect of eribulin on brain metastases is not clear. Preclinical evidence suggests that the agent does not cross the blood–brain barrier (BBB) to a significant extent [15,26]. Radiation therapy can compromise the BBB, which may help drugs enter into the brain [27]. However, this BBB drug crossing can be also explained by the altered BBB due to the presence of brain metastases alone. In the 301 study, a lower proportion of patients in the eribulin arm developed CNS metastases compared to the capecitabine arm (2.4% vs 4.6%, respectively) [7]. Moreover, in the three patients with known stable brain metastases included in that trial and treated with eribulin, a reduction in the size of brain lesions was also observed. Some case reports have also shown efficacy with eribulin at the CNS level, both in patients previously treated with radiotherapy and not [28,29].
The pattern of disease progression in patients treated with eribulin is distinctive. Most patients in our series progressed in pre-existing M1 sites (75%). This fact may reflect the inhibitory effect of eribulin in the EMT, reducing cancer cell migration and invasiveness; however, we did not find any information about the pattern of progression in patients who received other oncospecific treatments [1–3]. It has been previously reported that patients with ABC who develop tumour progression with new metastases have worse outcomes compared to those patients whose disease progresses due to growth of pre-existing lesions [11,30].
Regarding toxicity, most patients in our series demonstrated acceptable tolerability of the eribulin treatment. The most frequent grade 3/4 toxicities were asthenia, neurotoxicity and neutropenia and were manageable. Remarkably, the proportion of grade 3 neutropenia was lower in our study than in pivotal trials (9.8% vs 23.4%), as well as grade 4 toxicities with 10.2% in our study (all of them neutropenia) versus 27.7% in the pooled analysis [8].
This study has several strengths and limitations. Our study reflects the everyday clinical practice use of eribulin and analyzes multiple variables such as BMI, age >70 years and activity in patients with CNS metastases that have not been previously addressed and studied together in patients outside of a clinical trial. The major limitations of this study are the limited number of patients included and its retrospective nature without having a dedicated pre-planned design of patient management and data collection, which can lead to potential biases. It is worth noting that, during the study period, several clinical trials evaluating new targeted drugs in combination with eribulin were ongoing and those patients were not included in this study.