Eribulin versus S-1 as first or second-line chemotherapy to

assess health-related quality of life and overall survival in

HER2-negative metastatic breast cancer (RESQ study): a noninferiority, randomised, controlled, open-label, phase 3 trial

Masato Takahashi,a,∗ Yuichiro Kikawa,b Kosuke Kashiwabara,c Naruto Taira,d Tsuguo Iwatani,e Kojiro Shimozuma,f Shoichiro Ohtani,g

Tetsuhiro Yoshinami,h Junichiro Watanabe,i

Masahiro Kashiwaba,j

Ken-ichi Watanabe,k Masahiro Kitada,l

Koichi Sakaguchi,m Yuko Tanabe,n

Tomohiko Aihara,o and Hirofumi Mukaip

a

Hokkaido University Hospital, Sapporo, Japan

b

Kansai Medical University Hospital, Osaka, Japan

c

The University of Tokyo Hospital, Tokyo, Japan

d

Kawasaki Medical School, Kurashiki, Japan

e

Okayama University Hospital, Okayama, Japan

f

Ritsumeikan University, Kusatsu, Japan

g

Ohtani_S Breast Clinic, Hiroshima, Japan

h

Osaka University Hospital, Osaka, Japan

i

Juntendo University Graduate School of Medicine, Tokyo, Japan

j

Sagara Hospital Miyazaki, Miyazaki, Japan

k

NHO Hokkaido Cancer Centre, Sapporo, Japan

l

Asahikawa Medical University Hospital, Asahikawa, Japan

mKyoto Prefectural University of Medicine, Kyoto, Japan

n

Toranomon Hospital, Tokyo, Japan

o

Aihara Hospital, Osaka, Japan

p

National Cancer Centre Hospital East, Chiba, Japan

Summary

Background Eribulin prolongs overall survival (OS) of patients with human epidermal growth factor receptor 2

(HER2)-negative metastatic breast cancer (MBC), particularly in later chemotherapy (ChT) treatment. However, the

health-related quality of life (HRQoL) and efficacy of first or second-line therapy in eribulin-treated patients remain

unknown. Using eribulin in the first- or second-line may demonstrate the non-inferiority of HRQoL compared to S-1,

an oral 5-fluorouracil derivative, while maintaining OS.

Methods This randomised, controlled, open-label, phase III trial was conducted at 50 hospitals in Japan. Patients were

enrolled from June 2016 and October 2019. Patients with HER2-negative MBC once under or no previous ChT were

randomly assigned (1:1) to receive eribulin or S-1. HRQoL was assessed using the European Organization for

Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire-Core 30 (QLQ-C30) every six weeks

until week 24 and every nine weeks until week 42. The primary endpoint was the deterioration defined as more

than 10 points worsening of the general health score of QLQ-C30 or death within one year after randomisation.

The secondary endpoints included OS. (Trial ID: UMIN000021398).

Findings Three hundred and two patients were enrolled, with 152 and 148 assigned to the eribulin and S-1 groups,

respectively. The questionnaire compliance rate was 85.6%. Risk difference of global health status deterioration

through one year was −0.66% (95% CI: −12.47–11.16; non-inferiority P = 0.077) for eribulin compared to S-1 groups.

Median time to first deterioration for global health status score was 5.64 (95% CI: 3.51–8.00) and 5.28 months (95%

CI: 3.28–7.80) in the eribulin and S-1 groups, respectively. The median OS was 34.7 and 27.8 months, (HR: 0.72, 95%

CI: 0.54–0.96; P = 0.026); the median progression-free survival was 7.57 and 6.75 months in the eribulin and S-1

groups, (HR: 0.88, 95% CI: 0.67–1.16; P = 0.35), respectively. No new adverse events occurred.

Interpretation The time of the first clinical deterioration was similar between the two groups and OS significantly

increased in eribulin-treated patients.

*Corresponding author. Department of Breast Surgery, Hokkaido University Hospital, Kita14 Nishi5, Kita-ku, Sapporo, 060-8648, Japan.

E-mail address: masato.takahashi@huhp.hokudai.ac.jp (M. Takahashi).

eClinicalMedicine

2024;74: 102715

Published Online 16 July

2024

https://doi.org/10.

1016/j.eclinm.2024.

102715

www.thelancet.com Vol 74 August, 2024 1

Articles

Funding This study was funded by CSPOR-BC and Eisai CO., Ltd.

Copyright © 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords: Eribulin; S-1; Metastatic breast cancer; Health-related quality of life; Overall survival

Introduction

Breast cancer is the most prevalent malignancy affecting

women worldwide, despite advances in chemotherapy

and targeted therapies.1 Metastatic breast cancer (MBC)

remains incurable despite drug developments. Chemotherapy is the standard treatment for both oestrogen

receptor-negative and positive tumours of this lifethreatening disease. Chemotherapy for MBC prolongs

survival and improves quality of life (QoL).2

Overall survival (OS), or progression-free survival

(PFS) as a surrogate for OS, is considered the primary

endpoint of most phase III clinical studies on MBC.3

Therefore, differences in effects of chemotherapy regimens have been based solely on prolonging survival as a

primary endpoint, with QoL improvement considered a

secondary endpoint because evaluating QoL changes is

difficult.

However, maintaining QoL is as important as prolonging survival under the current circumstances,

where MBC cannot be cured. The European Society for

Medical Oncology (ESMO) developed the ESMO

Magnitude of Clinical Benefit Scale (ESMO-MCBS), a

well-validated tool for stratifying the magnitude of

clinical benefits of cancer treatment. QoL evaluation in

clinical trials has become increasingly important,4

because evaluations using validated questionnaires,

such as the European Organization for Research and

Treatment of Cancer Quality of Life Questionnaire

(EORTC QLQ-C30) and Functional Assessment of

Cancer Therapy-General (FACT-G), have been found to

be quantitative, reproducible, and reliable.5 Therefore,

we arranged a phase III clinical trial for MBC, with

health-related QoL (HRQoL) as the primary trial

endpoint.6,7

The possibility that choosing between first- or

second-line chemotherapy regimens for MBC results in

HRQoL differences is unclear. SELECT BC is a randomised controlled trial in which orally administered

fluorinated pyrimidine S-1 was compared with a standard intravenous chemotherapy regimen, (taxanes) in

women with MBC.8 S-1 was not inferior to the taxanes in

OS. Additionally, the global health status after S-1

treatment was significantly increased than that after

taxane treatment according to the EORTC QLQ-C30

questionnaire. Hence, S-1 is considered a first-line

chemotherapy option for patients with human

Research in context

Evidence before this study

We conducted a literature review of the MEDLINE and

EMBASE databases on March 30, 2023, using the following

search terms (all fields) with no publication date or language

restrictions: “eribulin” AND (“patient-reported outcomes” OR

“PROs” OR “health-related quality of life” OR “HRQoL”) AND

“metastatic breast cancer” AND “random”. The search yielded

19 articles. Of these, nine papers described patient-reported

outcomes of eribulin-containing treatment in patients with

metastatic breast cancer in a randomised study.

Of the nine papers, two studies were randomised comparisons

between eribulin monotherapy and other chemotherapy

regimens. The EMBRACE trial was a late-line efficacy study of

eribulin. The efficacy of eribulin in OS has been validated;

however, its QoL outcomes are inadequate. The 301 study,

which mainly included patients undergoing second-line

chemotherapy, verified QoL maintenance but did not verify

OS superiority of eribulin. Prior anthracycline and taxane

treatment was required in both studies.

From these research results, the earlier use of eribulin, such as

in first-line treatment, is expected to maintain QoL without

worsening OS; however, evidence for this claim is insufficient.

The efficacy of eribulin in patients not previously treated with

taxanes remains unproved.

Added value of this study

Our study is the first randomised phase III trial to compare

eribulin with other chemotherapeutic agents in patients not

previously treated with anthracyclines or taxanes. The noninferiority of HRQoL in eribulin was not statistically

demonstrated and the median time of deterioration for

global status was similar in the two groups. Eribulin, used

primarily as a first-line therapy, showed superiority in OS

compared to S-1.

Implications of all the available evidence

We could not conclude that eribulin is non-inferior to S-1

with respect to HRQoL with the non-inferiority margin of

10% but we could show that Eribulin, as an earlier line of

treatment, did not worsen QoL deterioration time and may

prolong OS. The results of this study may be utilised for

decision-making during treatment selection for patients with

HER2-negative metastatic breast cancer.

Articles

2 www.thelancet.com Vol 74 August, 2024

epidermal growth factor receptor 2 (HER2)-negative

MBC.9–11

Eribulin, a macrocyclic ketone analogue of halichondrin B, exerts anticancer effects via disrupting the

tubulin structure, thus inhibiting cancer cells.12 In a

phase III clinical trial named Study 301, eribulin was

compared with capecitabine as second-line treatments

for MBC, with a median OS of 15.9 and 14.5 months for

eribulin and capecitabine, respectively.13 Eribulin has

been proven to be superior or result in equal OS for

MBC compared to conventional chemotherapeutic

agents, including oral fluoropyrimidine anticancer

drugs. In Study 301, HRQoL associated with eribulin

and capecitabine was investigated.14 In clinical trials,

assessing changes over time to quantify patient QoL and

comparing differences between groups is essential. One

objective method for detecting these quantitative temporal differences is using the minimal important difference (MID).5,15 MID is a predefined minimum

threshold at which a clinical difference is assumed to

exist. Using MID as a QoL measure, no difference in

HRQoL was detected between eribulin and capecitabine

in this clinical study.14

Therefore, eribulin is expected to have the same effect as capecitabine on life prognosis as a second-line

treatment, which is the same oral fluoropyrimidine

drug as S-1. For example, for gastrointestinal toxicity,

eribulin showed a good trend in patient QoL compared

with capecitabine.14 Thus, eribulin may be equivalent to

S-1 in QoL maintenance.

Based on this background, we planned a randomised,

controlled, open-label phase III trial. Our hypothesis was

that using eribulin in the first- or second-line would

demonstrate the non-inferiority of HRQoL compared to

S-1 while maintaining OS, using the EORTC QLQ-C30

questionnaire. The study aims to provide further insights into the impact of eribulin and S-1 on the QoL of

patients with MBC. The findings of this study show eribulin as a potential additional treatment option for

patients with HER2-negative MBC.

Methods

Study design and participants

This randomised, controlled, open-label, phase III trial

was conducted in 50 hospitals in Japan. Women aged

20–75 years were enrolled with informed consent. Patients were to have passed a specific period since the end

of the previous treatment during registration. The inclusion criteria for this study were patients who were

histologically diagnosed with breast cancer regardless of

histologic subtype. Patients has to have been diagnosed

with MBC with measurable lesions, either as stage IV

breast cancer with distant metastasis not indicated for

surgery at the first visit or as distant metastasis after

initial treatment for curative purposes. Patients must

also have had planned chemotherapy for MBC and an

Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 or 1. Patients were to have had

satisfactory maintenance of major organ functions at the

final examination performed within 14 days prior to

enrolment and met the following inclusion criteria:

neutrophil count (ANC): 1500/mm3 or higher, platelet

count: 100,000/mm3 or higher, haemoglobin level of

9.0 g/dL or more, aspartate transaminase/alanine

transaminase: less than 100 IU/L (less than 150 IU/L in

patients with liver metastasis), serum total bilirubin:

<2.1 mg/dL, serum creatinine: <1.6 mg/dL, and creatinine clearance >50 mL/min; creatinine clearance estimate = [(140-age) × body weight]/[72 × serum creatinine

level] × 0.85. Additionally, patients who complied with

scheduled visits, treatment plans, tests, QoL assessments, and other procedures were eligible for inclusion

in this study.

Representative exclusion criteria included being

HER2-positive, having undergone previous treatment

with S-1 or Eribulin, having residual Grade 3 or higher

adverse events from previous treatment, having

concomitant malignancies, or having bilateral breast

cancer. However, patients with ductal carcinoma in situ

on one side or bilateral breast cancer with the same

oestrogen receptor, progesterone receptor, and HER2

status were eligible for participation.

Ethics

The study protocol (Supplementary protocol file) was

approved by Hokkaido Cancer Center Ethics Committeethe in 2016. Subsequently, the Japanese Clinical

Research Law was enacted in 2018, making it necessary

to obtain ethical approval under this law for clinical

studies during the patient registration period. The study

approved by the Central Research Ethics Review Committee (Hokkaido University Certified Review Board) in

2019, and the approval number was CRB1180001. This

study was conducted following the ethical principles of

the Declaration of Helsinki and all applicable local regulations. All patients provided written informed consent

prior to enrolment.

Randomisation and masking

Patients were randomly assigned (1:1) to receive eribulin

or S-1. Registration was from June 2016 to October 2019.

Randomisation was centrally generated with the minimisation method, stratified by institution, age (<50

versus ≥50), treatment line (first versus second line),

oestrogen receptor status, and time from surgery to

recurrence (<24 months versus ≥24 months). A random

sequence was generated using a validated computer

system operated by the data management team (DATA

Centre, Comprehensive Support Project for Oncology

Research, Tokyo, Japan; independent of the trial sponsor

and investigators). Local investigators entered the data

into the system during enrolment and follow-ups. Participants, investigators, and data analysts were not

Articles

www.thelancet.com Vol 74 August, 2024 3

blinded to the treatment assignment because an oralbased regimen was compared with an infusion regimen.

Procedures

S-1 was administered orally to patients assigned to the S1 group at doses of 40–60 mg twice daily for 14

consecutive days and then discontinued for seven days.

The next course began 21 days (day 22) following the

previous course (day 1). The dose was determined according per body surface area; those with a body surface

area less than 1.25 m2

, between 1.25 and 1.5 m2

, and

1.5 m2 or greater received 80 mg, 100 mg, and 120 mg

per day, respectively.

Eribulin was administered intravenously to patients

in the eribulin group in doses of 1.4 mg/m2 for 2–5 min

on days 1 and 8 of a 21-day cycle.

Both treatments were repeated until tumour progression or for at least six courses (24 weeks). Continuation of chemotherapy beyond this was permitted at

the physician’s discretion. If the allocated regimen

failed, another cytotoxic or endocrine drug of the investigator’s choice was administered. Crossover and

generic drugs were allowed.

Treatment efficacy was determined via computerised

tomography CT scans of the chest and abdomen every

six weeks from enrolment to week 24 and every nine

weeks thereafter, according to RECIST (version 1.1).

Physical findings, PS, grade 2 or higher nonhaematologic toxicities, peripheral blood (white blood

cells, neutrophils, platelets, and haemoglobin), and

biochemical parameters (total protein, total bilirubin,

aspartate aminotransferase, alanine aminotransferase,

lactate dehydrogenase, blood urea nitrogen, and creatinine) were assessed at baseline and on day 1 of each

treatment cycle. HRQoL was assessed using the EORTC

QLQ-C30 (version 3.0), the breast module of the EORTC

QLQ (QLQ-BR23) (version 1.0), and EuroQol 5 dimensions 5-level (EQ-5D-5L) at enrolment and every six

weeks until week 24 from the start of study treatment,

and every nine weeks thereafter until week 42. After the

first year, only the EQ-5D-5L was assessed every year

during the study observation period.16–18 These questionnaires were completed by the patients themselves at

the hospital and handed to the person in charge or at

home and mailed. In both cases, the attending physicians were not allowed to confirm the content.

Outcome

The primary endpoint was the clinically meaningful

deterioration of the global health status (GHS) score of

the EORTC QLQ-C30 within one year of enrolment.

Secondary endpoints included time to first deterioration

of GHS or death, OS, PFS, time to treatment failure

(TTF), new metastasis-free survival (nMFS), adverse

events, patient preference, cost-effectiveness, and functional and symptom scales of the EORTC QLQ-C30 and

QLQ-BR23. OS was defined as the time from

randomisation to death from any cause. PFS and TTF

were defined as the time from the randomisation date to

the earliest date of disease progression or death from

any cause and the interval between the randomisation

date and earliest date of disease progression, study

treatment withdrawal for any reason, or death from any

cause, respectively. Time from randomisation to disease

progression due to new metastases or death from any

cause, whichever occurred earlier, was the definition for

nMFS.

Stastics

This study was designed to assess the non-inferiority of

eribulin with S-1 regarding the primary endpoint. A

patient was defined as improved, stable, or deteriorated

if the patient reported functioning scores, symptom

scores, and GHS of EORTC-QLQ C30 showed a change

≥+10, between −10 and +10, and ≤−10 points, respectively, from the baseline assessment.19 Death was

included in the definition of deterioration. The proportion of improved or stable patients who did not experience deterioration within one year of randomisation was

compared between the eribulin and S-1 groups, which

was estimated using the Kaplan–Meier method to account for patient dropouts. The fixed sequence procedure was used to test the non-inferiority and superiority

of eribulin to S-1 at two-sided 5% significance levels

each, in that order. The non-inferiority margin for the

proportion was defined as 10 percentage points.20,21 No

multiplicity adjustment for any other endpoints

including secondary endpoints were applied.

In the SELECT BC trial, 60% S-1 patients were

observed to have improved or were stable within one

year of randomisation. An improved/stable proportion

for eribulin than for S-1 was suggested after analysing

the SELECT BC trial and the published results of 221

trials, which led to an estimated improved/stable proportion of 65% for the eribulin group.22 These estimates

resulted in 328 patients, obtaining 80% power for noninferiority. The power of superiority was 15.3%. The

more optimistically estimated improved/stable proportion of 70% in the eribulin group had a 47.8% power for

superiority.

The functional and symptom scores of HRQoL was

analysed using a mixed-effects model for repeated

measures (MMRM) with post-treatment QoL scores

assessed during each visit as a response value, treatment

group, visit (time) and treatment-by-visit interaction as

fixed effects in which first-order autoregressive correlation structure was specified and death was treated as

censoring.23 The coefficient for treatment represented

the average difference in QoL scores between groups

throughout the entire treatment period (12 months),

adjusting for baseline scores. Normality of residuals, a

required assumption for MMRM, was tested by

Shapiro–Wilk test based on the pooled studentized residuals. If normality was rejected, a weighted

Articles

4 www.thelancet.com Vol 74 August, 2024

generalized estimating equation (wGEE) was performed

as a sensitivity analysis as follows. The censoring probability was estimated by the logistic regression, in which

the response variable was censoring indicator of each

visit and the explanatory variables were the treatment,

the baseline score of global health score, the score of

global health score just before each visit, and visits. To

obtain the stabilized weight, the same logistic regression

as the previous one except that the score of global health

score just before each visit was excluded from the

explanatory variables was also implemented. The structural modelling was the same as the MMRM and compound symmetry covariance was adopted to the working

covariance matrix of the response variable across visits.

The treatment effects were tested with the significance level set at 0.05. Exploratory time-to-event analyses

of OS and PFS were performed. Each patient was followed up until death or at least two years after randomisation. A landmark analysis with the landmark

time of 15 months was conducted for OS because of the

possible violation of proportional hazard assumption.

All statistical analyses were performed using the SAS

software (version 9.4; SAS Institute, NC, USA). This

trial was registered with the Japan Registry of Clinical

Trials (Trial ID: UMIN000021398, jRCTs011180026).

Role of funding source

The funder (Eisai Co., Ltd.) had no role in the study

design, data collection, analysis, interpretation, or

writing of the manuscript.

Results

Between June 2016 and October 2019, we enrolled 302

patients. Of these, 300 were randomly assigned (152 and

148 to the eribulin and S-1 groups, respectively). Fifteen

patients did not start the protocol treatment and the full

analysis set comprised 285 patients in which PFS and OS

were analysed: 141 and 144 in the eribulin and S-1

groups, respectively. Patient demographics are presented

in the table for the 270 cases in which HRQoL was

analysed (Fig. 1). No significant differences were

observed in patient characteristics, including age (<60

versus ≥60 years), ECOG performance status, number of

Fig. 1: CONSORT flowchart of participants.

Articles

www.thelancet.com Vol 74 August, 2024 5

prior chemotherapy regimens for advanced disease, prior

treatment with anthracycline and/or taxane in the adjuvant or advanced setting, hormone receptor status (oestrogen positive/negative), presence of liver metastasis, or

disease-free interval (<2 years versus ≥2 years), between

the eribulin and S-1 groups. These findings were

consistent in both the HRQoL and full analysis sets

comprising 285 patients for the PFS and OS analyses

(Table 1). In the full analysis set of 285 patients, 41.1%

and 44.4% of patients in the eribulin and S-1 groups,

respectively, were <60 years old. The patient proportion

with no prior chemotherapy regimens for advanced disease was 70.2 and 70.1% in the eribulin and S-1 groups,

respectively. When examining the treatment period for

cases in which a single regimen of chemotherapy was

administered for advanced disease, the median duration

of prior treatment in the Eribulin group was 9.2 months,

with a standard deviation of 6.1 months. In the S-1 group,

the median duration was 7.8 months, with a standard

deviation of 7.2 months. While previous treatments with

anthracyclines and taxanes was required for entry into

Studies 301 and 305, our study did not specify this

requirement. In this study, 32.6 and 40.3% in the

eribulin and S-1 groups, respectively, were previously

treated with anthracyclines and taxanes. Hormone status

indicated that 79.4% of patients in the eribulin group and

78.5% of patients in the S-1 group were ER-positive.

Regarding the expression rate of the estrogen receptor,

the proportion of ER-positive cases was divided into three

groups: low expression (1–9%), moderate expression

(10–50%), and high expression (51–100%). The Eribulin

group had 1.4% low, 18.6% moderate, and 80.0% high

ER expression. Similarly, the S1 group had 1.3% low,

25.3% moderate, and 73.3% high ER expression. Liver

metastases were observed in 47.5 and 45.8% of patients

in the eribulin and S1 groups, respectively.

Of the 270 patients in the HRQoL analysis set, the

overall compliance rate for the completion of the

HRQoL measure was >85%. At baseline, the mean

scores of EORTC-QLQ C30 and BR23 were similar in

both eribulin and S-1 groups (Table 2). No baseline

differences were observed between the eribulin and S1

groups in terms of the number of defects and scores for

detailed items (Table 3).

In the time to deterioration graph for the EORTC

QLQ-C30 global health status, using a 10-point decrease

HRQoL analysis set Full analysis set

Eribulin group S-1 group Eribulin group S-1 group

n = 134 n = 136 n = 141 n = 144

Age (years) 61.5 (37–74) 60.5 (29–74) 60.8 (37–74) 59.5 (29–74)

<60 years 55 (41.0) 59 (43.4) 58 (41.1) 64 (44.4)

≥60 years 79 (59.0) 77 (56.6) 83 (58.9) 80 (55.6)

ECOG performance status

0 94 (70.1) 103 (75.7) 100 (70.9) 109 (75.7)

1 40 (29.9) 32 (23.5) 41 (29.1) 34 (23.6)

2 0 (0.0) 1 (0.7) 0 (0.0) 1 (0.7)

No. of prior chemotherapy regimens for advanced disease

0 92 (68.7) 98 (72.1) 99 (70.2) 101 (70.1)

1 42 (31.3) 38 (27.9) 42 (29.8) 43 (29.9)

Prior treatment for either adjuvant or advanced setting with

Neither Anthracycline nor Taxane 70 (52.2) 68 (50.0) 76 (53.9) 73 (50.7)

Either Anthracyclin or Taxane 19 (14.2) 12 (8.8) 19 (13.5) 13 (9.0)

Both Anthracycline and Taxane 45 (33.6) 56 (41.2) 46 (32.6) 58 (40.3)

Hormone receptor status

ER positive 106 (79.1) 106 (77.9) 112 (79.4) 113 (78.5)

ER negative and PgR negative 27 (20.1) 30 (22.1) 28 (19.9) 31 (21.5)

Unknown 1 (0.7) 0 (0.0) 1 (0.7) 0 (0.0)

Liver metastasis

Yes 62 (46.3) 65 (47.8) 67 (47.5) 66 (45.8)

No 72 (53.7) 71 (52.2) 74 (52.5) 78 (54.2)

Disease free interval

<2 years 28 (20.9) 35 (25.7) 31 (22.0) 36 (25.0)

≥2 years 70 (52.2) 69 (50.7) 72 (52.2) 73 (50.7)

Without surgery 36 (26.9) 32 (23.5) 38 (27.5) 35 (24.3)

Data are median (range) or n (%).

Table 1: Baseline characteristics of patients.

Articles

6 www.thelancet.com Vol 74 August, 2024

as the event, the curves of the eribulin and S-1 groups

changed almost identically and no significant difference

was detected. The median time to first deterioration of

GHS score in the eribulin group was 5.64 months (95%

confidence interval, CI: 3.51–8.00), whereas that in the

S-1 group was 5.28 months (95% CI: 3.28–7.80) (hazard

ratio, HR: 1.07 [95% CI: 0.79–1.45]; P = 0.67) (Fig. 2).

The proportions of patients who did not experience

deterioration of GHS score in one year were 33.8% (95%

CI: 25.1–42.6) and 33.0% (95% CI: 24.6–41.6) in the

eribulin and S-1 groups, respectively. The risk difference of GHS deterioration in one year for the eribulin

group versus the S-1 group was–0.66% (95%

CI: −12.47–11.16; P non-inferiority = 0.077).

Fig. 3 illustrates the changes in QoL of the eribulin

and S-1 groups every six weeks, with a 10-point increase

Eribulin group S-1 group

(n = 134) (n = 136)

EORTC QLQ-C30 (number of responses, mean [SD])

GHS/QoL 134 59.6 (22.7) 136 60.5 (22.6)

Physical functioning 134 79.3 (19.7) 136 81.3 (18.0)

Role functioning 134 76.5 (28.0) 136 81.9 (22.7)

Emotional functioning 134 74.3 (19.3) 136 75.0 (20.8)

Cognitive functioning 134 77.6 (21.1) 136 81.6 (17.8)

Social functioning 133 80.3 (25.6) 136 84.1 (22.7)

Fatigue 134 36.5 (23.6) 136 32.6 (20.7)

Nausea and vomiting 134 5.0 (12.9) 136 5.5 (13.3)

Pain 134 27.0 (22.7) 136 23.7 (24.1)

Dyspnea 132 20.2 (25.3) 135 18.3 (22.2)

Insomnia 134 23.1 (26.2) 135 21.6 (24.2)

Appetite loss 132 20.5 (26.6) 136 17.9 (22.2)

Constipation 134 18.2 (28.5) 136 15.4 (21.0)

Diarrhea 134 6.5 (18.5) 136 7.8 (16.8)

Financial difficulties 131 27.0 (30.4) 136 20.8 (27.8)

EORTC QLQ-BR23 (number of responses, mean [SD])

Body image 132 67.2 (25.2) 133 69.9 (23.6)

Sexual functioning 128 2.3 (7.8) 125 2.1 (7.3)

Future perspective 132 44.2 (29.8) 133 43.4 (28.1)

Systemic therapy side effects 133 19.7 (13.9) 136 17.2 (11.8)

Breast symptoms 134 19.7 (21.7) 134 20.1 (21.1)

Arm symptoms 132 22.1 (22.6) 134 20.9 (22.5)

Upset by hair loss 132 7.3 (18.1) 136 5.1 (17.6)

Table 2: Baseline health-related quality of life.

Eribulin group (%) S-1 group (%)

(n = 134) (n = 136)

Baseline 100 134/134 100 136/136

6 weeks 94.2 126/133 91.9 124/135

12 weeks 88.9 116/131 85.5 112/130

18 weeks 83.6 109/130 80.2 100/125

24 weeks 78.7 97/123 87 105/122

33 weeks 73.2 88/119 74.6 88/117

42 weeks 79.3 85/108 71.8 78/109

Table 3: Proportion of patients completing GHS at scheduled points in

HRQOL analysis set.

Fig. 2: Time to first deterioration for EORTC QLQ-C30 GHS. EORTC

QLQ-C30, the European Organization for Research and Treatment of

Cancer Quality of Life Questionnaire-Core 30 questions; GHS, global

health status; TTD, time to deterioration; HR, hazard ratio; CI, confidence interval.

Articles

www.thelancet.com Vol 74 August, 2024 7

indicating improvement, a 10-point decrease indicating

aggravation, and no other recorded changes. No significant differences were observed between the two groups.

This displays the results of average change from baseline in the GHS every six weeks. This observation

method did not reveal any significant differences in

GHS between the eribulin and S-1 groups.

The forest plot of time to first deterioration,

including the function and symptom score of EORTCQLQ C30 and BR23, demonstrated that nausea/vomiting and physical functioning favoured the eribulin

group (HR: 0.65, 95% CI: 0.47–0.89; P = 0.0073) and S-1

group (HR: 1.45, 95% CI: 1.04–2.03; P = 0.027),

respectively (Fig. 4). The mean difference in GHS score

from baseline to 42 weeks between the two groups was

0.44 (95% CI: −3.49–4.38; P = 0.82). Because the

normality assumption required for the residuals of

MMRM was rejected by Shapiro–Wilk test (test statistic:

0.98; P < 0.0001), a wGEE sensitivity analysis was conducted and a similar mean difference was observed

(Supplementary material).

The median PFS was 7.57 and 6.75 months for the

eribulin and S-1 groups (HR: 0.88, 95% CI: 0.67–1.16;

P = 0.35) (Fig. 5). As of the data cut-off date (October 31

2022), 218 of the 285 patients had died in the full

analysis set: 99 in the eribulin group and 119 in the S-1

group. The median OS was 34.7 (95% CI: 27.2–41.0)

and 27.8 months (24.7–30.7) in the eribulin and S-1

groups, respectively (HR: 0.72, 95% CI: 0.54–0.96;

P = 0.026) (Fig. 6). Apparently, uniformity of HR over

time would not be true. In the exploratory landmark

analysis from 15 months, a larger HR was observed

(HR: 0.48, 95% CI: 0.29–0.78). Our randomised phase

III trial comparing existing chemotherapeutics found

that eribulin did not prolong PFS but prolonged OS.

The median time to treatment failure was 5.18

months (95% CI: 3.09–5.64) and 4.93 months

(3.15–5.64) for the eribulin and S-1 groups, respectively

(HR: 0.95, 95% CI: 0.73–1.22; P = 0.67). The median

new metastasis-free survival was 9.67 months (95% CI:

7.80–14.7) and 8.03 months (7.54–9.80) for the eribulin

and S-1 groups, respectively (HR: 0.86, 95% CI:

0.61–1.22; P = 0.39).

The classification factors for the OS subgroup analysis in this study were facility, age, primary and

Fig. 3: Mean change from baseline for GHS. The circles represent

the mean change from baseline score. Error bars indicate 95%

confidence interval range. GHS, global health status; CI, confidence

interval.

Fig. 4: Forest plot of TTD for EORTC QLQ-C30 of GHS, function, and symptom. The circles represent hazard ratio. Error bars indicate 95%

confidence interval range. EORTC QLQ-C30, the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Core

30 questions; GHS, global health status; HR, hazard ratio; CI, confidence interval.

Articles

8 www.thelancet.com Vol 74 August, 2024

secondary therapies, hormone receptor status, and time

from surgery to recurrence (Fig. 7). Regardless of these

factors, eribulin prolonged survival compared to S-1. In

this study, we enrolled patients who had not received

chemotherapy after recurrence and those who had

received up to one regimen. In the 200 patients (eribulin

group: 99, S-1 group: 101) who did not receive prior

treatment, the HR was 0.68 (95% CI: 0.48–0.96); in the

second-line therapy cases, in which one prior therapy

was administered, the HR was 0.84 (95% CI: 0.50–1.42).

In addition, when OS was examined in terms of prior

history of anthracycline or taxane, the HR was 0.72 (95%

CI: 0.48–1.09) for the 149 anthracycline- and taxanenaive patients, compared to the HR for the group of

136 patients previously treated with anthracycline or

taxane, which was 0.84 (95% CI: 0.55–1.26). With regard

to hormone receptor expression, the number of ER

negative cases was small (61); however, the difference in

ER expression did not affect OS. In the 133 patients with

liver metastases, the HR was 0.89 (95% CI: 0.58–1.38).

In contrast, in 152 patients without liver metastasis, the

HR was 0.61 (95% CI: 0.40–0.94).

Table 4 shows the treatments selected by the full

analysis set of participants after completing the treatment protocol. In the eribulin and S-1 groups, 24.1 and

17.8% of the participants chose taxane or anthracycline

as their subsequent treatment, respectively. Furthermore, 29.1% from the eribulin group chose S-1 as their

next treatment, whereas 38.1% of the S-1 group chose S1 as their next treatment. Overall, 34.0% of patients

received cross-treatments. Additionally, 17.0% of the

eribulin and 17.8% of S-1 groups selected the best

supportive care after completing treatment protocol.

Adverse events were observed in 97.9% and 94.5% of

patients in the eribulin and S-1 groups, respectively,

with differences in occurrence frequency across multiple items (Table 5). Neutropenia occurred in 60.3% of

the eribulin group, with grade 3 or higher neutropenia

observed in 49.3% of cases. In contrast, the S-1 group

had a neutropenia rate of 26.7%, with grade 3 or higher

observed in 8.2% of the cases (P < 0.0001). Peripheral

sensory neuropathy was reported in 55.5% and 12.3% of

the eribulin and S-1 groups, respectively (P < 0.0001).

Alopecia (hair loss) was observed in 64.4% and 9.6% of

the eribulin and S-1 groups, respectively (P < 0.0001).

In contrast, diarrhea was observed more frequently

in the S-1 group, with grade 3 or higher diarrhea

observed in 0.7% and 4.8% of the eribulin and S-1

groups, respectively (P < 0.0001). Thrombocytopenia

was reported in 6.8% and 22.6% of the eribulin and S-1

groups (P = 0.0002). Additionally, hyperpigmentation

was more prominent in the S-1 (45.9%) than the eribulin group (3.4%) (P < 0.0001).

Discussion

When selecting a chemotherapy regimen for MBC,

therapeutic efficacy and adverse events must be

balanced. For evidence-based treatment selection,

referring to randomised phase III trial results is recommended. To assess both therapeutic efficacy and

adverse events in clinical trials, monitoring PFS or OS

as therapeutic effect indicators and QoL as a health

benefit indicator is crucial. In many previous clinical

trials, PFS has been used as a surrogate for OS as the

primary endpoint, whereas QoL is infrequently used for

this purpose.2 Historically, anthracyclines and taxanes

have been considered first-line chemotherapy options

for MBC treatment. Despite advancements in chemotherapeutic drug development, none have shown better

OS than conventional drugs. S-1, a combination of

tegafur, gimeracil, and oteracil potassium, has been

approved for MBC treatment in Japan.24 As a first-line

chemotherapy, S-1 was non-inferior to taxanes in OS,

with a primary endpoint of 37.2 months compared to

Fig. 5: Progression free survival. PFS, progression free survival; HR,

hazard ratio; CI, confidence interval.

Fig. 6: Overall survival. OS, overall survival; HR, hazard ratio; CI,

confidence interval.

Articles

www.thelancet.com Vol 74 August, 2024 9

35.0 months. In addition, S-1 demonstrated better QoL

maintenance than taxanes or anthracyclines.

Conversely, Eribulin, a microtubule inhibitor of the

halichondrin class derived from the black sea sponge,

has been developed as a novel chemotherapy agent.

Results from a Phase III clinical trial conducted overseas (Study 301) showed that second-line chemotherapy

targeting eribulin tended to improve OS compared with

using capecitabine in patients with MBC.13 Furthermore, QoL analysis using the EORTC QLQ-C30 in

Study 301 demonstrated that eribulin was comparable to

capecitabine in GHS.14

The ACCRU trial, which aimed to compare the OS of

eribulin with paclitaxel as a conventional agent in first or

second-line MBC treatment, initiated patient enrolment

in the United States. Consequently, S-1, a conventional

drug with favourable OS and QoL, was selected for

comparison with eribulin in the ACCRU trial. In that

trial, OS was established as the primary endpoint,

whereas our study design identified QoL as the primary

endpoint.

In EMBRACE and 301 trials, prior treatment with

anthracyclines and taxanes was required as an inclusion

criteria.13,25 As a result, evidence for eribulin efficacy

against MBC has been limited to patients who have

received prior treatment with these drugs. However, in

this clinical trial, prior treatment with anthracyclines

and taxanes was not required. Only 40% patients in this

trial had received prior treatment with these drugs and

Fig. 7: Subgroup analysis of overall survival. The circles represent hazard ratio. Error bars indicate 95% confidence interval range. HR, hazard

ratio; CI, confidence interval.

Eribulin

group

(n = 141)

S-1 group

(n = 146)

Eribulin – – 56 (38.4)

S-1 41 (29.1) – –

Taxan/Anthracylin 34 (24.1) 26 (17.8)

CDK4/6i plus hormone 5 (3.5) 7 (4.8)

Other 24 (17.0) 20 (13.7)

Best Supportive Care 24 (17.0) 26 (17.8)

Unknown 6 (4.3) 6 (4.1)

Ongoing Protocol treatment 7 (5.0) 5 (3.4)

Table 4: Treatment after completing protocol treatment.

Articles

10 www.thelancet.com Vol 74 August, 2024

more than 50% of the cases received neither anthracyclines nor taxanes in this trial. We believe that this study

demonstrates that eribulin influence is not restricted to

patients who previously received anthracycline and taxane treatments.

Eribulin did not show non-inferiority to S-1 in the

primary endpoint of global QoL. However, visual inspection of the data revealed that the QoL of both

treatments was virtually equivalent. Furthermore, QoL

result analysis for each cycle indicated no statistically

significant difference between the two treatments at any

time point.

In this study, 330 patients were deemed necessary to

demonstrate non-inferiority. However, the actual

enrolment rate was 10% lower and instances in which

the initial QoL questionnaire was not provided occurred.

Approximately 10% cases were affected, resulting in

20% fewer cases available for analysis than the required

number.

Consequently, non-inferiority was not established

because fewer cases were available for analysis than

originally planned. However, the results did not

conclusively indicate that eribulin was inferior to S-1 in

QoL.

Although the overall QoL showed little difference

between the eribulin and S-1 groups, distinct variations

were observed. For instance, eribulin improved the QoL

with regard to nausea and vomiting, whereas S-1 tended

to have better results for social and physical functioning.

This information can provide valuable insights into the

patient’s perspective and facilitate medical doctors to

tailor treatment decisions to the patient’s unique needs

and assess disease progression.

Although the primary endpoint of this study was not

OS, a statistically significant difference was observed in

the efficacy of eribulin treatment compared to S-1

treatment based on the prespecified exploratory analysis.

In contrast, no significant difference was observed between the two groups in PFS, which served as a surrogate for OS. This finding is not unique to this study, as

previous studies, such as EMBRACE (study 301) and

study of sarcoma (study 309) have also demonstrated

that eribulin extends PFS, which is generally regarded

as an OS surrogate marker, compared to conventional

treatments.26 These results further indicated that eribulin treatment resulted in a significant OS increase.

While we cannot completely rule out bias in interpreting

these results, a comparison of patient characteristics

between the eribulin group and the S-1 group showed

no significant differences. Specifically, there were no

disparities regarding whether or not prior chemotherapy

for advanced disease was administered, nor in the

duration of the treatment period for previous therapies

in patients who experienced disease progression postfirst-line chemotherapy. Similarly, Table 1 indicates no

significant difference in the number of ER positive cases

between the two groups. We further categorized the

positive rate of ER positive cases in greater detail to

Eribulin group (n = 146) S-1 group (n = 146) P valuea

Grade 1–2 Grade 3 Grade 4 Grade 1–2 Grade 3 Grade 4

Hematological

Leucopenia ne 29 (19.9%) 8 (5.5%) ne 3 (2.1%) 0 <0.0001

Neutropenia 16 (11.0%) 37 (25.3%) 35 (24.0%) 27 (18.5%) 11 (7.5%) 1 (0.7%) <0.0001

Thrombocytopenia 8 (5.5%) 2 (1.4%) 0 30 (20.5%) 2 (1.4%) 1 (0.7%) 0.0002

ALT increased ne 10 (6.8%) 0 ne 5 (3.4%) 0

Non-hematological

Fatigue 81 (55.5%) 1 (0.7%) 0 67 (45.9%) 1 (0.7%) 0

Alopecia 94 (64.4%) - - 14 (9.6%) - - <0.0001

Hyperpigmentation 5 (3.4%) - - 67 (45.9%) - - <0.0001

Motor neuropathy 19 (13.0%) 2 (1.4%) 0 4 (2.7%) 1 (0.7%) 0

Sensory neuropathy 69 (47.3%) 11 (7.5%) 1 (0.7%) 17 (11.6%) 1 (0.7%) 0 <0.0001

Arthralgia 27 (18.5%) 2 (1.4%) 0 10 (6.8%) 0 0

Febrile neutropenia - 11 (7.5%) 1 (0.7%) - 2 (1.4%) 0 0.011

Fever 30 (20.5%) 1 (0.7%) 1 (0.7%) 15 (10.3%) 1 (0.7%) 0

Diarrhea 25 (17.1%) 0 1 (0.7%) 56 (38.4%) 7 (4.8%) 0 <0.0001

Mucositis 47 (32.2%) 2 (1.4%) 0 26 (17.8%) 4 (2.7%) 0

Nausea 51 (35.0%) 4 (2.7%) 0 65 (44.5%) 5 (3.4%) 0

Vomiting 12 (8.3%) 3 (2.1%) 1 (0.7%) 24 (16.4%) 1 (0.7%) 0

Anorexia 55 (37.7%) 7 (4.8%) 1 (0.7%) 54 (37.0%) 8 (5.5%) 0

Data are n (%). Adverse events were defined according to the Common Toxicity Criteria of the National Cancer Institute, version 4.0. ne, not evaluated; ALT, alanine

aminotransferase. a

Fisher’s exact test (Grade 0 versus 1 versus 2 versus 3 versus 4). Fisher’s exact test (Grade 0–2 versus 3 versus 4).

Table 5: Adverse events in the safety population.

Articles

www.thelancet.com Vol 74 August, 2024 11

assess any differences between the groups; however, our

analysis did not reveal any notable discrepancies. Taken

together, these detailed analyses support the conclusion

that there are no differences in the backgrounds of the

registered cases.

In this study, we collected data on the subsequent

treatments received by patients after completing the

treatment protocol. Although 34% patients received

crossover treatment before their next treatment, we

detected a difference in OS. This finding suggests that

eribulin could potentially improve OS for the next

treatment. Eribulin treatment extends OS but not PFS;

the reason for this remains elusive. The following three

mechanisms have been postulated based on basic eribulin studies.

First, eribulin can reverse the epithelial–

mesenchymal transition (EMT) in cancer cells. Cancer

cells can proliferate, invade, and metastasise; however,

each cell can differ in its abilities.27 Cancer cells that

have undergone EMT, which transform into spindleshaped cells, tend to renew their invasive capabilities

and are more likely to metastasise.28 Eribulin reverses

EMT, resulting in rounder cell morphology, decreased

invasion ability, and altered malignancy.29 These

morphological changes enhance efficacy of subsequent

drug therapies and prevent cancer spread. In Study 301,

eribulin was found to suppress new metastasis formation compared to capecitabine, which suggests that its

effects may persist after treatment with eribulin.13

The second reason is improved oxygenation. Eribulin

is believed to transform the cancer cell environment

from hypoxic to oxygenated via affecting the surrounding microvessels.30 Hypoxia exacerbates cancer cell malignancy and hinders drug delivery.31 Restoring

oxygenation enables subsequent treatments to reach the

cancer cells and increase their cytotoxicity in an oxygenrich environment.

Finally, cancer immunity is enhanced by eribulin.

The EMBRACE trial revealed that a strong effect on OS

was observed in patients with an absolute lymphocyte

count (ALC) of 1500 or higher, but no OS improvement

was observed in patients with low ALC.32 Similarly, a

Japanese case series showed that eribulin extended OS

in patients with an ALC of 1500 or higher.33 This correlation between lymphocyte count and the therapeutic

effect of eribulin suggests a significant role of host immunity in prolonging OS. If eribulin enhances the host

immune response against cancer, it is likely to influence

subsequent treatments and ultimately prolong OS. We

believe that understanding the relationship between

ALC and OS is important. We plan to examine these

results in additional analyses in future studies.

A limitation of this study is that the number of patients whose QOL could be analysed was small

compared to the planned enrolment of 330 patients.

Consequently, the accuracy of statistical analysis of the

primary endpoint was compromised. Additionally, all

registered patients were limited to Japanese participants. The strength of this study lies in its ability to

compare eribulin efficacy with standard therapy in a

randomised phase III trial. Similar to previous studies,

it did not prolong PFS compared with standard care but

extended OS. Moreover, the study enrolled cases that

included some patients who had not been previously

treated with anthracyclines or taxanes. In addition, some

patients underwent crossover treatments in the subsequent treatment phase, which is crucial when considering the reasons for prolonged OS.

Despite the presence of some crossover treatments, a

significant finding of this study was the distinct prolongation of OS observed with eribulin compared with

S-1. These findings can be used in decision-making for

patient treatments. Currently, various new drugs are

being developed to treat HER2 negative breast cancer.

Therefore, a limitation of this study is that the impact of

cytotoxic chemotherapeutic agents, including oral fluorinated pyrimidines, may be less significant than before.

However, since the use of these drugs will definitely

continue in the future, we believe that the overall survival and QoL data presented in this study will remain

valuable not only now but in the future. In making

treatment decisions, it is essential to evaluate from

multiple perspectives. While this paper primarily

presents data on QoL and OS from the study, costeffectiveness is also a crucial perspective. We were unable to present a detailed analysis of cost-effectiveness at

this time, but we plan to do so in a forthcoming paper.

We plan to further examine the impact of subsequent

treatments on OS and aim to provide an optimal

chemotherapy sequence.

Contributors

The study was designed by the funder in collaboration with all the authors. MT, YK, KK, NT and HM contributed to the conceptualisation,

data curation, methodology and visualisation of the results of the study.

MT, YK and KK contributed to formal analysis. MT, YK, KK, TI and NT

contributed to the underlying data verification. MT and KK had full

access to all the data in the study. All authors contributed to the investigation. SO, TY, JW, MKa, KW, MKi, KSa, YT administered the project.

TA and HM contributed to preparation of resources and software. KSh,

TA and HM have supervised this study and funding acquisition. All

authors had access to all the data reported in the study and contributed

to data interpretation. The first draft of the manuscript was written by

MT. All authors reviewed and edited the manuscript and approved the

final manuscript for publication.

Data sharing statement

Anonymized patient data, the research protocol, and the analysis

methodology can be made available for sharing upon reasonable request

and provision of a comprehensive protocol and analysis plan to the

corresponding author.

Declaration of interests

KK, TI, KSh, SO, MKi, YT and TA report no coflicts of interest. MT

reports honoraria from Astra Zeneca, Daiichi Sankyo, Eisai, Eli Lilly,

MSD and Pfizer. YK reports honoraria from Astra Zeneca, Chugai,

Daiichi Sankyo, Eisai, Eli Lilly, Pfizer and Taiho. NT reports grant and

honoraria from Eisai. TY reports honoraria from Astra Zeneca, Chugai,

Eisai, Eli Lilly, Kyowa Kirin, MSD and Pfizer. JW reports grant from

Articles

12 www.thelancet.com Vol 74 August, 2024

Eisai and honoraria from Eisai and Taiho. MKa reports honoraria from

Eli Lilly and Pfizer. KW reports honoraria from Astra Zeneca, Chugai,

Daiichi Sankyo, Eisai, Eli Lilly, Kyowa Kirin, Novartis, Nippon Kayaku,

Pfizer, Shionogi and Taiho. KSa reports honoraria from Eisai and Taiho.

HM reports honoraria from Daiichi Sankyo, Taiho and Takeda.

Acknowledgements

We sincerely thank all patients who participated in the RESQ study and

their families, as well as the investigators and research coordinators at

the 50 institutions and CSPOR-BC(Comprehensive Support Project For

Oncological Research of Breast Cancer) along with the staff involved in

the RESQ study.

Funding: This study was funded by the CSPOR-BC and Eisai Co.,

Ltd. The funder of the study had no role in the study design, data

collection, data analysis, data interpretation, or writing of the report.

Appendix A. Supplementary data

Supplementary data related to this article can be found at https://doi.

org/10.1016/j.eclinm.2024.102715.

References

1 Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020:

GLOBOCAN estimates of incidence and mortality worldwide for 36

cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209–249.

2 Cardoso F, Paluch-Shimon S, Senkus E, et al. 5th ESO-ESMO international consensus guidelines for advanced breast cancer (ABC

5). Ann Oncol. 2020;31(12):1623–1649.

3 Lux MP, Böhme S, Hücherig S, Jeratsch U, Kürschner N,

Lüftner D. Surrogate threshold effect based on a meta-analysis for

the predictive value of progression-free survival for overall survival

in hormone receptor-positive, HER2-negative metastatic breast

cancer. Breast Cancer Res Treat. 2019;176(3):495–506.

4 Oosting SF, Barriuso J, Bottomley A, et al. Methodological and

reporting standards for quality-of-life data eligible for European

society for medical oncology-magnitude of clinical benefit scale

(ESMO-MCBS) credit. Ann Oncol. 2023;34(4):431–439.

5 Smith AB, Cocks K, Parry D, Taylor M. Reporting of health-related

quality of life (HRQOL) data in oncology trials: a comparison of the

European organization for research and treatment of cancer quality

of life (EORTC QLQ-C30) and the functional assessment of cancer

therapy-general (FACT-G). Qual Life Res. 2014;23(3):971–976.

6 Aaronson NK, Ahmedzai S, Bergman B, et al. The European organization for research and treatment of cancer QLQ-C30: a qualityof-life instrument for use in international clinical trials in oncology.

J Natl Cancer Inst. 1993;85(5):365–376.

7 Cella DF, Tulsky DS, Gray G, et al. The functional assessment of

cancer therapy scale: development and validation of the general

measure. J Clin Oncol. 1993;11(3):570–579.

8 Takashima T, Mukai H, Hara F, et al. Taxanes versus S-1 as the

first-line chemotherapy for metastatic breast cancer (SELECT BC):

an open-label, non-inferiority, randomised phase 3 trial. Lancet

Oncol. 2016;17(1):90–98.

9 Shiroiwa T, Fukuda T, Shimozuma K, et al. Cost-effectiveness

analysis of the introduction of S-1 therapy for first-line metastatic

breast cancer treatment in Japan: results from the randomized

phase III SELECT BC trial. BMC Cancer. 2017;17(1):773.

10 Mukai H, Uemura Y, Akabane H, et al. Anthracycline-containing

regimens or taxane versus S-1 as first-line chemotherapy for metastatic breast cancer. Br J Cancer. 2021;125(9):1217–1225.

11 Takashima T, Hara F, Iwamoto T, et al. A correlation analysis between metabolism-related genes and treatment response to S-1 as

first-line chemotherapy for metastatic breast cancer: the SELECT

BC-EURECA study. Clin Breast Cancer. 2021;21(5):450–457.

12 Jordan MA, Kamath K, Manna T, et al. The primary antimitotic

mechanism of action of the synthetic halichondrin E7389 is suppression of microtubule growth. Mol Cancer Ther. 2005;4(7):1086–1095.

13 Kaufman PA, Awada A, Twelves C, et al. Phase III open-label randomized study of eribulin mesylate versus capecitabine in patients

with locally advanced or metastatic breast cancer previously treated

with an anthracycline and a taxane. J Clin Oncol. 2015;33(6):594–601.

14 Cortes J, Hudgens S, Twelves C, et al. Health-related quality of life

in patients with locally advanced or metastatic breast cancer treated

with eribulin mesylate or capecitabine in an open-label randomized

phase 3 trial. Breast Cancer Res Treat. 2015;154(3):509–520.

15 Kawahara T, Taira N, Shiroiwa T, et al. Minimal important differences of EORTC QLQ-C30 for metastatic breast cancer patients:

results from a randomized clinical trial. Qual Life Res.

2022;31(6):1829–1836.

16 Efficace F, Biganzoli L, Piccart M, et al. Baseline health-related

quality-of-life data as prognostic factors in a phase III multicentre

study of women with metastatic breast cancer. Eur J Cancer.

2004;40(7):1021–1030.

17 Okamoto T, Shimozuma K, Katsumata N, et al. Measuring quality

of life in patients with breast cancer: a systematic review of reliable

and valid instruments available in Japan. Breast Cancer.

2003;10(3):204–213.

18 Wallwiener M, Simoes E, Sokolov AN, Brucker SY, Fasching PA,

Graf J. Health-related quality of life in metastatic and adjuvant

breast cancer patients. Geburtshilfe Frauenheilkd. 2016;76(10):1065–

1073.

19 Musoro JZ, Coens C, Fiteni F, et al. Minimally important differences for interpreting EORTC QLQ-C30 scores in patients with

advanced breast cancer. JNCI Cancer Spectr. 2019;3(3):pkz037.

20 Chow E, van der Linden YM, Roos D, et al. Single versus multiple

fractions of repeat radiation for painful bone metastases: a randomised, controlled, non-inferiority trial. Lancet Oncol.

2014;15(2):164–171.

21 Pham C, Niu F, Delate T, et al. Real-world clinical outcomes of

bevacizumab-awwb biosimilar versus bevacizumab reference

product in patients with metastatic colorectal cancer. BioDrugs.

2023;37(6):891–899.

22 Aogi K, Iwata H, Masuda N, et al. A phase II study of eribulin in

Japanese patients with heavily pretreated metastatic breast cancer.

Ann Oncol. 2012;23(6):1441–1448.

23 Faries DELA, Haro JM, Obenchain RL. Analysis of observational

health care data using SAS.: Care NC. SAS Institute Inc; 2010.

24 Saeki T, Takashima S, Sano M, et al. A phase II study of S-1 in

patients with metastatic breast cancer–a Japanese trial by the S-1

cooperative study group, breast cancer working group. Breast

Cancer. 2004;11(2):194–202.

25 Cortes J, O’Shaughnessy J, Loesch D, et al. 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(9769):914–923.

26 Schöffski P, Chawla S, Maki RG, et al. Eribulin versus dacarbazine

in previously treated patients with advanced liposarcoma or leiomyosarcoma: a randomised, open-label, multicentre, phase 3 trial.

Lancet. 2016;387(10028):1629–1637.

27 Yoshida T, Ozawa Y, Kimura T, et al. Eribulin mesilate suppresses

experimental metastasis of breast cancer cells by reversing phenotype from epithelial-mesenchymal transition (EMT) to

mesenchymal-epithelial transition (MET) states. Br J Cancer.

2014;110(6):1497–1505.

28 Gavert N, Ben-Ze’ev A. Epithelial-mesenchymal transition and the

invasive potential of tumors. Trends Mol Med. 2008;14(5):199–209.

29 Zhao N, Powell RT, Yuan X, et al. Morphological screening of

mesenchymal mammary tumor organoids to identify drugs that

reverse epithelial-mesenchymal transition. Nat Commun.

2021;12(1):4262.

30 Funahashi Y, Okamoto K, Adachi Y, et al. Eribulin mesylate reduces tumor microenvironment abnormality by vascular remodeling in preclinical human breast cancer models. Cancer Sci.

2014;105(10):1334–1342.

31 Muz B, de la Puente P, Azab F, Azab AK. The role of hypoxia in

cancer progression, angiogenesis, metastasis, and resistance to

therapy. Hypoxia. 2015;3:83–92.

32 Miyoshi Y, Yoshimura Y, Saito K, et al. High absolute lymphocyte

counts are associated with longer overall survival in patients with

metastatic breast cancer treated with eribulin-but not with treatment of physician’s choice-in the EMBRACE study. Breast Cancer.

2020;27(4):706–715.

33 Takahashi M, Inoue K, Mukai H, et al. Indices of peripheral leukocytes predict longer overall survival in breast cancer patients on

eribulin in Japan. Breast Cancer. 2021;28(4):945–955.

Articles

www.thelancet.com Vol 74 August, 2024 13