The n e w e ng l a n d j o u r na l of m e dic i n e

original article

Pertuzumab, Trastuzumab, and Docetaxel

in HER2-Positive Metastatic Breast Cancer
Sandra M. Swain, M.D., José Baselga, M.D., Sung-Bae Kim, M.D., Jungsil Ro, M.D.,
Vladimir Semiglazov, M.D., Mario Campone, M.D., Eva Ciruelos, M.D.,
Jean-Marc Ferrero, M.D., Andreas Schneeweiss, M.D., Sarah Heeson, B.Sc.,
Emma Clark, M.Sc., Graham Ross, F.F.P.M., Mark C. Benyunes, M.D.,
and Javier Cortés, M.D., for the CLEOPATRA Study Group*

A BS T R AC T

Background
The authors’ affiliations are listed in the In patients with metastatic breast cancer that is positive for human epidermal
Appendix. Address reprint requests to growth factor receptor 2 (HER2), progression-free survival was significantly im-
Dr. Swain at the Washington Cancer Insti-
tute, MedStar Washington Hospital Cen- proved after first-line therapy with pertuzumab, trastuzumab, and docetaxel, as
ter, 110 Irving St., Washington, DC 20010, compared with placebo, trastuzumab, and docetaxel. Overall survival was signifi-
or at [email protected]. cantly improved with pertuzumab in an interim analysis without the median being
* A complete list of investigators in the reached. We report final prespecified overall survival results with a median follow-
Clinical Evaluation of Pertuzumab and up of 50 months.
Trastuzumab (CLEOPATRA) study is pro-
vided in the Supplementary Appendix, Methods
available at NEJM.org. We randomly assigned patients with metastatic breast cancer who had not received
N Engl J Med 2015;372:724-34. previous chemotherapy or anti-HER2 therapy for their metastatic disease to receive
DOI: 10.1056/NEJMoa1413513 the pertuzumab combination or the placebo combination. The secondary end
Copyright © 2015 Massachusetts Medical Society.
points of overall survival, investigator-assessed progression-free survival, indepen-
dently assessed duration of response, and safety are reported. Sensitivity analyses
were adjusted for patients who crossed over from placebo to pertuzumab after the
interim analysis.
Results
The median overall survival was 56.5 months (95% confidence interval [CI], 49.3 to
not reached) in the group receiving the pertuzumab combination, as compared with
40.8 months (95% CI, 35.8 to 48.3) in the group receiving the placebo combination
(hazard ratio favoring the pertuzumab group, 0.68; 95% CI, 0.56 to 0.84; P<0.001),
a difference of 15.7 months. This analysis was not adjusted for crossover to the per-
tuzumab group and is therefore conservative. Results of sensitivity analyses after ad-
justment for crossover were consistent. Median progression-free survival as assessed
by investigators improved by 6.3 months in the pertuzumab group (hazard ratio, 0.68;
95% CI, 0.58 to 0.80). Pertuzumab extended the median duration of response by
7.7 months, as independently assessed. Most adverse events occurred during the admin-
istration of docetaxel in the two groups, with long-term cardiac safety maintained.
Conclusions
In patients with HER2-positive metastatic breast cancer, the addition of per­tuzumab
to trastuzumab and docetaxel, as compared with the addition of placebo, signifi-
cantly improved the median overall survival to 56.5 months and extended the re-
sults of previous analyses showing the efficacy of this drug combination. (Funded
by F. Hoffmann–La Roche and Genentech; CLEOPATRA ClinicalTrials.gov number,
NCT00567190.)

724 n engl j med 372;8 nejm.org february 19, 2015

The New England Journal of Medicine

Downloaded from nejm.org on February 29, 2024. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
 Pertuzumab Combination in HER2-Positive Breast Cancer

T
he overexpression of human epi- area or its equivalent. All randomization and
dermal growth factor receptor 2 (HER2) in masking processes and assessments have been
breast cancer results in more aggressive reported previously.4,5
disease with a poor prognosis.1 The humanized After significant improvement in overall
anti-HER2 monoclonal antibodies pertuzumab survival was reported with pertuzumab (Perjeta,
and trastuzumab are more active in combination F. Hoffmann–La Roche and Genentech),5 pa-
than alone because of more comprehensive sig- tients who were still receiving the placebo com-
naling blockade.2,3 We investigated combination bination in whom disease had not progressed
therapy with docetaxel for first-line treatment of were invited to cross over to receive pertuzumab.
HER2-positive metastatic breast cancer in the Overall survival results presented here are from
Clinical Evaluation of Pertuzumab and Tras­ the intention-to-treat population; therefore, data
tuzumab (CLEOPATRA) trial. Analysis of the pri- from crossover patients were analyzed in the
mary end point showed that patients who re- control group. Two sensitivity analyses of overall
ceived pertuzumab, trastuzumab, and docetaxel survival were conducted: one in which data from
(pertuzumab group) had a significantly longer crossover patients were censored at the time of
median progression-free survival, as assessed by the first pertuzumab dose and one in which these
independent reviewers, than did those who re- patients were excluded from the results. Post-
ceived placebo, trastuzumab, and docetaxel (con- crossover safety data are reported separately.
trol group) (hazard ratio favoring the pertuzu­ The CLEOPATRA trial was performed in ac-
mab group, 0.62).4 The second interim analysis cordance with Good Clinical Practice guidelines
of overall survival confirmed significantly longer and the provisions of the Declaration of Helsinki.
survival in the pertuzumab group (hazard ratio, Protocol approval was obtained from an inde-
0.66).5 Safety profiles (including cardiac) were pendent ethics committee at each study site. All
similar and consistent across the two study patients provided written informed consent.
groups and the analysis time points.4-6 Here we
report follow-up data at a median of 50 months Study Oversight
regarding overall survival, investigator-assessed The study was designed by the senior academic
progression-free survival and safety, and inde- authors and representatives of the sponsors,
pendently assessed duration of response. F. Hoffmann–La Roche and Genentech. Data
were collected by the sponsors and analyzed in
Me thods collaboration with the senior academic authors,
who vouch for the completeness and accuracy of
Study Design and Patients the data and analyses and for the fidelity of the
In this randomized, double-blind, placebo-con- study to the protocol. The protocol, including the
trolled, phase 3 trial,4,5 we enrolled patients who statistical analysis plan, is available with the full
were at least 18 years of age with locally recur- text of this article at NEJM.org. All investigators
rent, unresectable, or metastatic HER2-positive participating in the study agreed to confidential-
breast cancer that was centrally confirmed. Eligi- ity regarding the data. The first author prepared
ble patients had a left ventricular ejection fraction the initial draft of the manuscript with assistance
(LVEF) of 50% or more at baseline, had an East- from a medical writer employed by Health Inter-
ern Cooperative Oncology Group performance actions and funded by F. Hoffmann–La Roche.
status of 0 or 1, and had received no more than All authors contributed to subsequent drafts and
one hormonal treatment for metastatic disease. made the decision to submit the manuscript for
Adjuvant or neoadjuvant chemotherapy with or publication.
without trastuzumab was allowed. Patients were
excluded if they had central-nervous-system me- Study Procedures
tastases, if the LVEF was less than 50% during or Study drugs were administered intravenously
after previous trastuzumab therapy, or if they every 3 weeks. Pertuzumab (at a dose of 840 mg)
had received other anticancer therapy (with the or placebo was given on day 1 of cycle 1, followed
exception of one previous hormonal regimen) or by 420 mg on day 1 of each subsequent cycle;
had a cumulative exposure of more than 360 mg trastuzumab (Herceptin, F. Hoffmann–La Roche
of doxorubicin per square meter of body-surface and Genentech) at a dose of 8 mg per kilogram of

n engl j med 372;8 nejm.org february 19, 2015 725

The New England Journal of Medicine
Downloaded from nejm.org on February 29, 2024. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

body weight was administered on day 2 of cycle 1, the use of the O’Brien–Fleming approach and
followed by 6 mg per kilogram on day 1 of the was crossed at the second interim analysis; con-
remaining cycles. Pertuzumab or placebo and sequently, this final analysis is considered to be
trastuzumab were administered until disease pro- descriptive.
gression or the occurrence of unmanageable We used the log-rank test to compare overall
toxic effects; dose reductions were not allowed. survival between the two treatment groups, with
Docetaxel was given at a dose of 75 mg per stratification according to status with respect to
square meter on day 2 of cycle 1 and on day 1 of adjuvant or neoadjuvant chemotherapy and geo-
the remaining cycles; at least six cycles were rec- graphic region. Kaplan–Meier analyses were used
ommended, but fewer cycles were allowed in case to estimate medians. A Cox proportional-hazards
of disease progression or unmanageable toxic ef- model was used to estimate hazard ratios and
fects, and more cycles were allowed at the discre- 95% confidence intervals with the same stratifi-
tion of the investigator or patient. Docetaxel cation factors.
could be escalated to 100 mg per square meter if Subgroup analyses were performed to ensure
there were no unmanageable toxic effects and robustness of the treatment effect across pre-
could be reduced by 25% in the case of myelosup- specified categories. All analyses were performed
pression, hepatic dysfunction, or other toxic ef- with the use of SAS software, version 8.2 (SAS
fects. Growth-factor support could be added Institute).
according to prescribing information and the
guidelines of the American Society of Clinical R e sult s
Oncology.7 No concurrent hormonal therapy was
allowed before disease progression. Study Population
Patients were enrolled from February 12, 2008,
Study End Points through July 7, 2010. The cutoff for data collec-
Since independent review of progression-free sur- tion for this analysis was February 11, 2014. (En-
vival and the objective response rate were stopped rollment and randomization are shown in Fig. 1.)
after the first analysis,4 we report updated data Baseline characteristics were similar in the two
on the secondary end points of overall survival study groups; 630 patients (88.0%) had visceral
(time from randomization to death from any disease.4,5 A total of 389 patients had died at the
cause) and investigator-assessed progression-free time of this analysis. Median follow-up was 49.5
survival, which was defined as the time from months (range, 0 to 70) in the pertuzumab group
randomization to the first documented radio- and 50.6 months (range, 0 to 69) in the control
graphic evidence of progressive disease, accord- group. After the interim analysis of overall sur-
ing to Response Evaluation Criteria in Solid Tu- vival in May 2012, investigators were informed
mors (RECIST), version 1.0,8 or death from any about study-group assignments. Between July
cause within 18 weeks after the last investigator and November 2012, a total of 48 of the 406 pa-
assessment of tumors. We also present the dura- tients (11.8%) in the control group crossed over
tion of response from the time of the data cutoff to receive pertuzumab.
for the primary analysis.
Overall Survival
Statistical Analysis Intention-To-Treat Population
Data for patients who were alive or lost to follow- Deaths were reported in 168 of 402 patients
up were censored at the last date they were (41.8%) in the pertuzumab group and in 221 of
known to be alive or at the time of randomiza- 406 patients (54.4%) in the control group (haz-
tion plus 1 day, if no information was available ard ratio favoring the pertuzumab group, 0.68;
after baseline. For the final analysis of overall 95% confidence interval [CI], 0.56 to 0.84;
survival, we determined that the study would P<0.001) (Fig. 2A). The median overall survival
have a power of 80% to detect a 33% improve- was 56.5 months (95% CI, 49.3 to not reached) in
ment in the pertuzumab group (hazard ratio for the pertuzumab group and 40.8 months (95% CI,
death from any cause, 0.75) after the occurrence 35.8 to 48.3) in the control group, a difference of
of 385 deaths. The statistical stopping boundary 15.7 months. The estimated Kaplan–Meier over-
for each interim analysis was established with all survival rate was 94.4% (95% CI, 92.1 to 96.7)

726 n engl j med 372;8 nejm.org february 19, 2015

The New England Journal of Medicine

Downloaded from nejm.org on February 29, 2024. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
 Pertuzumab Combination in HER2-Positive Breast Cancer

1196 Patients were screened

388 Did not meet inclusion

criteria or declined to
participate

808 Were enrolled and underwent

randomization

406 Were assigned to receive 402 Were assigned to receive

placebo, trastuzumab, pertuzumab, trastuzumab,
and docetaxel and docetaxel

9 Received nonran- 1 Received nonran-

domized treatment domized treatment

2 Did not receive 2 Did not receive

any treatment any treatment

406 Were included in the 396 Were included in the 408 Were included in the 402 Were included in the
intention-to-treat safety population safety population intention-to-treat
population population
358 Did not cross over to 67 Were alive and re-
pertuzumab group ceiving randomized
6 Were alive and 345 Withdrew from study treatment
receiving random- treatment 125 Were alive and in
ized treatment 36 Withdrew for safety 338 Withdrew from study survival follow-up
90 Were alive and in reasons treatment 42 Withdrew consent or
survival follow-up 23 Had adverse events 41 Withdrew for safety were lost to follow-up
42 Withdrew consent 13 Died reasons 168 Died
or were lost to 309 Withdrew for non- 34 Had adverse events
follow-up safety reasons 7 Died
220 Died 281 Had progressive 297 Withdrew for non-
48 Crossed over to per- disease safety reasons
tuzumab group 1 Had violation of 264 Had progressive
31 Were alive and re- selection criteria disease
ceiving crossover at entry 2 Had violation of
treatment 1 Had other proto- selection criteria
15 Were alive and in col violation at entry
survival follow-up 23 Declined treatment 21 Declined treatment
1 Withdrew consent 1 Failed to return 4 Failed to return
1 Died 2 Had other reasons 6 Had other reasons

Figure 1. Enrollment and Outcomes.

All patients who underwent randomization were included in the intention-to-treat population, and patients who received at least one
dose of a study drug were included in the safety analysis. Reasons for withdrawals are shown. Included in the safety population are nine
patients in the control group who received pertuzumab and one patient in the pertuzumab group who received placebo, as indicated.

in the pertuzumab group and 89.0% (95% CI, 4 years. Exploratory analyses in predefined sub-
85.9 to 92.1) in the control group at 1 year; 80.5% groups showed a consistent benefit with pertuz­
(95% CI, 76.5 to 84.4) and 69.7% (95% CI, 65.0 to umab (Fig. 2B). The hazard ratio for death from
74.3), respectively, at 2 years; 68.2% (95% CI, 63.4 any cause among patients who had previously
to 72.9) and 54.3% (95% CI, 49.2 to 59.4), respec- been treated with trastuzumab (47 patients in
tively, at 3 years; and 57.6% (95% CI, 52.4 to 62.7) the pertuzumab group and 41 patients in the
and 45.4% (95% CI, 40.2 to 50.6), respectively, at control group) was 0.80 (95% CI, 0.44 to 1.47).

n engl j med 372;8 nejm.org february 19, 2015 727

The New England Journal of Medicine
Downloaded from nejm.org on February 29, 2024. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

A Overall Survival
100
90
80

Overall Survival (%)

70 Pertuzumab, 168 events
60
50
40
30 Control, 221 events
20
Hazard ratio, 0.68 (95% CI, 0.56–0.84)
10 P<0.001
0
0 10 20 30 40 50 60 70 80
Months
No. at Risk
Pertuzumab 402 371 318 268 226 104 28 1 0
Control 406 350 289 230 179 91 23 0 0

B Subgroup Analysis of Overall Survival

No. of P Value for
Subgroup Patients Hazard Ratio (95% CI) Interaction
All patients 808 0.67 (0.55–0.82)
Previous adjuvant or
neoadjuvant treatment 0.63
No 432 0.64 (0.48–0.85)
Yes 376 0.70 (0.53–0.93)
Region 0.36
Europe 306 0.65 (0.47–0.91)
North America 135 0.63 (0.37–1.07)
South America 114 0.50 (0.30–0.85)
Asia 253 0.82 (0.57–1.16)
Age 0.27
<65 yr 681 0.70 (0.56–0.87)
≥65 yr 127 0.53 (0.31–0.90)
Age 0.84
<75 yr 789 0.68 (0.55–0.83)
≥75 yr 19 0.85 (0.26–2.73)
Race or ethnic group 0.40
White 480 0.63 (0.49–0.82)
Black 30 0.41 (0.11–1.45)
Asian 261 0.82 (0.58–1.17)
Other 37 0.37 (0.13–1.06)
Disease type 0.03
Visceral 630 0.59 (0.48–0.74)
Nonvisceral 178 1.11 (0.66–1.85)
ER or PgR status 0.47
Positive 388 0.71 (0.53–0.96)
Negative 408 0.61 (0.47–0.81)
HER2 status
IHC 3+ 721 0.66 (0.53–0.81) 0.52
FISH-positive 767 0.69 (0.56–0.85) 0.14

0.2 0.4 0.6 1.0 2.0 3.0 4.0

Pertuzumab Better Placebo Better

Sensitivity Analyses censored at the time of the first pertuzumab

The 48 patients without disease progression who dose, the median overall survival was 56.5
opted to cross over from the control group to months (95% CI, 49.3 to not reached) in the per-
receive pertuzumab had all been receiving treat- tuzumab group and 39.6 months (95% CI, 35.0 to
ment for 2 years or longer. When their data were 45.1) in the control group (hazard ratio, 0.63;

728 n engl j med 372;8 nejm.org february 19, 2015

The New England Journal of Medicine

Downloaded from nejm.org on February 29, 2024. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
 Pertuzumab Combination in HER2-Positive Breast Cancer

of response were 20.2 months in the pertuzumab

Figure 2 (facing page). Overall Survival.
group (95% CI, 16.0 to 24.0) and 12.5 months in
Panel A shows Kaplan–Meier estimates of overall sur-
vival in the intention-to-treat population, stratified ac- the control group (95% CI, 10.0 to 15.0).
cording to adjuvant or neoadjuvant therapy and geo-
graphic region. The median overall survival among Breast-Cancer Therapies after
patients receiving pertuzumab, trastuzumab, and doce­ Discontinuation of study treatment
taxel (pertuzumab group) was 56.5 months, 15.7 months
Among the 704 patients in the intention-to-treat
longer than survival among patients receiving placebo,
trastuzumab, and docetaxel (control group). The tick population who received breast-cancer therapies
marks indicate censoring events. Panel B shows hazard after the discontinuation of the study treatment,
ratios and 95% confidence intervals for overall survival the proportions of patients who received various
in all prespecified subgroups according to baseline drugs was balanced in the two study groups (Ta-
characteristics, without stratification. Race or ethnic
ble 1). A total of 72.9% patients in the pertuz­
group was determined by the investigator. “Other” in-
cludes American Indian or Alaska Native, Native Ha- umab group and 71.5% in the control group re-
waiian or other Pacific Islander, and other populations. ceived HER2-targeted therapy.
A grade of 3+ on immunohistochemical (IHC) analysis
indicates positivity for human epidermal growth factor Treatment Exposure
receptor 2 (HER2). For the assessment of HER2 status,
The median number of study-treatment cycles re-
prespecified subgroup analyses were restricted to pa-
tients whose tumors had an IHC score of 3+ or fluores- ceived by patients in the safety population was 24
cence in situ hybridization (FISH) positive status, since in the pertuzumab group (range, 1 to 96; 197
these categories accounted for approximately 90% of patients received more than the median number)
patients. CI denotes confidence interval, ER estrogen and 15 in the control group (range, 1 to 67). Pa-
receptor, and PgR progesterone receptor.
tients who crossed over to the pertuzumab group
received a median of 22.5 cycles of pertuzumab
(range, 1 to 28), which was similar to the median
95% CI, 0.52 to 0.78; P<0.001) (Fig. S1A in the number of cycles received by patients in the per-
Supplementary Appendix, available at NEJM.org). tuzumab safety population. Docetaxel exposure
When crossover patients were excluded, the me- did not change between data cutoffs (median, 8
dian overall survival was 56.5 months (95% CI, cycles in each group).
49.3 to not reached) in the pertuzumab group
and 34.7 months (95% CI, 31.2 to 39.4) in the Side-Effect Profile and Cardiac Safety
control group (hazard ratio for death from any All Patients
cause, 0.55; 95% CI, 0.45 to 0.67; P<0.001) (Fig. Rates of adverse events remained consistent with
S1B in the Supplementary Appendix). those in the primary analysis, with headache, up-
per respiratory tract infection, and muscle spasm
Investigator-Assessed Progression-Free reported as new adverse events with a difference
Survival of at least 5 percentage points between groups.
In the investigator-assessed intention-to-treat Most events were grade 1 or 2 and occurred dur-
analysis, events occurred in 284 of 402 patients ing docetaxel administration and declined after
(70.6%) in the pertuzumab group and 320 of 406 discontinuation. A list of the most common ad-
patients (78.8%) in the control group (hazard ra- verse events of all grades that occurred at a fre-
tio, 0.68; 95% CI, 0.58 to 0.80; P<0.001) (Fig. 3A). quency of 25% or more or showed a difference of
The results of subgroup analyses were consistent at least 5 percentage points between study groups
(Fig. 3B). The medians were unchanged from the overall is provided in Table S1 in the Supplemen-
May 2012 interim analysis (18.7 months in the tary Appendix. Table 2 shows the rates of the
pertuzumab group and 12.4 months in the con- same most common adverse events after the dis-
trol group, an improvement of 6.3 months).5 continuation of docetaxel.
The rate of left ventricular dysfunction, as
Duration of Response defined by the National Cancer Institute Com-
The duration of response was independently as- mon Terminology Criteria for Adverse Events,
sessed in patients with a confirmed partial or version 3.0, and the New York Heart Associa-
complete response at the time of the primary tion,6 was somewhat lower in the pertuzumab
analysis (275 patients in the pertuzumab group group than in the control group (6.6% [27 of 408
and 233 in the control group). Median durations patients] vs. 8.6% [34 of 396 patients]). There

n engl j med 372;8 nejm.org february 19, 2015 729

The New England Journal of Medicine
Downloaded from nejm.org on February 29, 2024. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

A Progression-free Survival
100
Hazard ratio, 0.68 (95% CI, 0.58–0.80)
90 P<0.001

Progression-free Survival (%)

80
70
60
50
40
Pertuzumab, 284 events
30
20
10
Control, 320 events
0
0 10 20 30 40 50 60 70 80
Months
No. at Risk
Pertuzumab 402 284 179 121 87 37 6 0 0
Control 406 223 110 75 51 21 6 0 0

B Subgroup Analysis of Progression-free Survival

No. of P Value for
Subgroup Patients Hazard Ratio (95% CI) Interaction
All patients 0.68 (0.58–0.80)
Previous adjuvant or 808
neoadjuvant treatment 0.60
No 432 0.66 (0.53–0.82)
Yes 376 0.71 (0.56–0.90)
Region 0.36
Europe 306 0.72 (0.55–0.93)
North America 135 0.52 (0.35–0.78)
South America 114 0.60 (0.39–0.91)
Asia 253 0.77 (0.58–1.03)
Age 0.08
<65 yr 681 0.72 (0.61–0.86)
≥65 yr 127 0.50 (0.32–0.77)
Age 0.76
<75 yr 789 0.69 (0.58–0.81)
≥75 yr 19 0.62 (0.16–2.40)
Race or ethnic group 0.50
White 480 0.65 (0.53–0.80)
Black 30 0.54 (0.19–1.51)
Asian 261 0.77 (0.58–1.03)
Other 37 0.45 (0.20–1.01)
Disease type 0.19
Visceral 630 0.64 (0.53–0.76)
Nonvisceral 178 0.83 (0.58–1.18)
ER or PgR status 0.38
Positive 388 0.73 (0.58–0.91)
Negative 408 0.64 (0.51–0.81)
HER2 status
IHC 3+ 721 0.66 (0.55–0.78) 0.21
FISH-positive 767 0.68 (0.58–0.81) 0.02

0.2 0.4 0.6 1.0 2.0

Pertuzumab Better Placebo Better

was one new event of symptomatic left ventricu- with both antibodies discontinued. Reductions
lar dysfunction in the pertuzumab group after in the LVEF of 10% or more from baseline to an
40 months, an event that resolved after 3 months absolute value of less than 50% occurred in 24 of

730 n engl j med 372;8 nejm.org february 19, 2015

The New England Journal of Medicine

Downloaded from nejm.org on February 29, 2024. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
 Pertuzumab Combination in HER2-Positive Breast Cancer

Figure 3 (facing page). Progression-free Survival. Table 1. Breast-Cancer Treatments Received by Patients Who Discontinued
Panel A shows Kaplan–Meier estimates of investigator- Study Treatment.*
assessed progression-free survival in the intention-to-
Pertuzumab
treat population, stratified according to adjuvant or
Control Group Group
neoadjuvant therapy and geographic region. The medi- Treatment (N = 369) (N = 335)
an progression-free survival was 18.7 months in the
pertuzumab group as compared with 12.4 months in no. of patients (%)
the control group, an improvement of 6.3 months. The Any treatment received after discon­ 291 (78.9) 258 (77.0)
tick marks indicate censoring events. Panel B shows tinuing study treatment
hazard ratios and 95% confidence intervals for investi-
HER2-targeted treatment 208 (71.5) 188 (72.9)
gator-assessed progression-free survival in all prespeci-
fied subgroups according to baseline characteristics, Trastuzumab 121 (41.6) 117 (45.3)
without stratification. Pertuzumab 4 (1.4) 2 (0.8)
Lapatinib 142 (48.8) 124 (48.1)
394 patients (6.1%) in the pertuzumab group Trastuzumab emtansine 34 (11.7) 32 (12.4)
and 28 of 378 patients (7.4%) in the control Capecitabine 170 (58.4) 142 (55.0)
group. Declines were reversed in 21 of 24 pa- Vinorelbine 88 (30.2) 67 (26.0)
tients (87.5%) in the pertuzumab group and 22
Doxorubicin 56 (19.2) 41 (15.9)
of 28 patients (78.6%) in the control group.
Most deaths were due to disease progression Cyclophosphamide 49 (16.8) 41 (15.9)
(in 150 of 408 patients [36.8%] in the per­ Taxanes 56 (19.2) 39 (15.1)
tuzumab group and 196 of 396 patients [49.5%] Hormonal treatments 56 (19.2) 69 (26.7)
in the control group). Other causes of death
were febrile neutropenia or infection (in 7 of 408 * Percentages are based on the numbers of patients in the intention-to-treat pop­
ulation who received any treatment after the discontinuation of study drugs.
patients [1.7%] in the pertuzumab group and 6 Patients who discontinued treatment include 4 patients who were not treated
of 396 patients [1.5%] in the control group) and and 17 patients who crossed over to ­receive pertuzumab and subsequently
causes that were classified as “other” or “un- withdrew from crossover treatment.
known” (in 12 of 408 patients [2.9%] in the
pertuzumab group and 15 of 396 patients [3.8%] tween-group separation in the Kaplan–Meier
in the control group). curves occurred early and was maintained over
time. Findings in subgroup analyses were consis-
Crossover Population tent with the final results and the results of pre-
No new safety signals were identified among pa- vious analyses.
tients in the control group who crossed over to We were able to estimate the treatment effect
receive pertuzumab. Most events were of grade 1 on the basis of a median overall survival of 56.5
or 2. Of the 221 events in the crossover group, months in the pertuzumab group, a duration
7 were grade 3 events, and 2 were grade 4 events that is exceptionally long in this population of
(diarrhea and dehydration in the same patient). patients. In previous prospective studies of first-
There was one death from an unknown cause. line therapy with trastuzumab plus chemother-
No symptomatic left ventricular dysfunction was apy in patients with HER2-positive metastatic
reported after crossover. Two patients had as- breast cancer, the median overall survival has
ymptomatic reductions in the LVEF. ranged from 25.1 months to 38.1 months.9-13 The
progression-free survival benefit was also main-
Discussion tained over time in our study.4,5
The majority of adverse events occurred dur-
First-line therapy with pertuzumab, trastuzumab, ing docetaxel treatment. Long-term cardiac safe-
and docetaxel significantly improved overall sur- ty was maintained, since treatment with pertuz­
vival among patients with HER2-positive meta- umab did not increase cardiac toxicity (including
static breast cancer, as compared with placebo, asymptomatic and symptomatic left ventricular
trastuzumab, and docetaxel. Since this was an dysfunction), as compared with placebo. Symp-
intention-to-treat analysis, patients in the control toms developed at a later time with pertuzumab
group who crossed over to receive pertuzumab and quality of life did not differ from that in the
were included in the control group, a factor that control group during the chemotherapy period
added to the strength of the findings. The be- in the pertuzumab group.14

n engl j med 372;8 nejm.org february 19, 2015 731

The New England Journal of Medicine
Downloaded from nejm.org on February 29, 2024. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

rates of complete response on pathological

Table 2. Adverse Events after the Discontinuation of Docetaxel in the Safety
Population.* analysis, as compared with trastuzumab plus
docetaxel, pertuzumab plus docetaxel, or pertuz­
Control Group Pertuzumab Group umab plus trastuzumab.17 These data supported
Adverse Event (N = 261) (N = 306)
the use of neoadjuvant pertuzumab with trastuz­
Most common events of any grade
— no. of patients (%)†
umab and docetaxel in patients with HER2-
positive breast cancer, an indication that was
Alopecia 6 (2.3) 5 (1.6)
recently approved by the Food and Drug Admin-
Diarrhea‡ 37 (14.2) 86 (28.1) istration.18
Neutropenia 13 (5.0) 10 (3.3) Despite the clinically meaningful survival re-
Nausea 30 (11.5) 39 (12.7) sults in our study, important questions remain
Fatigue 25 (9.6) 41 (13.4) to be addressed in this field. First, most deaths
Rash‡ 21 (8.0) 56 (18.3) in the study were due to breast cancer; therefore,
better treatments are still needed. This is particu-
Asthenia 23 (8.8) 41 (13.4)
larly important in patients with early disease,
Decreased appetite 14 (5.4) 22 (7.2)
such as those in the ongoing Adjuvant Pertuzu­
Peripheral edema 32 (12.3) 28 (9.2) mab and Herceptin in Initial Therapy of Breast
Vomiting 17 (6.5) 30 (9.8) Cancer (APHINITY) trial (ClinicalTrials.gov num-
Myalgia 19 (7.3) 25 (8.2) ber, NCT01358877). Second, it is not clear
Mucosal inflammation 4 (1.5) 11 (3.6) whether hormonal therapy plus pertuzumab and
Headache 32 (12.3) 52 (17.0)
trastuzumab is more effective than hormonal
therapy plus trastuzumab alone in patients with
Constipation 18 (6.9) 17 (5.6)
hormone receptor–positive disease. Third, there
Upper respiratory tract infection‡ 32 (12.3) 56 (18.3) is no biomarker in addition to HER2 that pre-
Pruritus‡ 15 (5.7) 42 (13.7) dicts which patients with HER2-positive disease
Febrile neutropenia 0 0 would benefit most from the combination of
Dry skin 10 (3.8) 10 (3.3) pertuzumab and trastuzumab.19-21 The presence
Muscle spasm‡ 6 (2.3) 24 (7.8) of mutated PIK3CA may indicate an adverse prog-
nosis for patients with metastatic disease but
* Data are for patients who received at least one dose of a study drug after com- did not have a predictive value in our study.21
pleting docetaxel treatment. Data for overall adverse events are provided in Whether other chemotherapeutic agents could
Table S1 in the Supplementary Appendix.
† The most common events are those that occurred with a frequency of 25% or improve on the safety profile is a fourth area for
more overall (including during the docetaxel treatment period) or that differed investigation. A phase 2 study assessing progres-
by 5 percentage points or more in frequency between the two groups overall. sion-free survival at 6 months with pertuzumab,
‡ The frequency of this event was at least 5 percentage points greater in the
pertuzumab group, as compared with the control group. trastuzumab, and paclitaxel showed few grade 3
or 4 adverse events and no cardiac events.22 Fi-
nally, the question of when, if ever, therapy with
Preclinical data showed that combining per­ pertuzumab plus trastuzumab for metastatic
tuzumab and trastuzumab resulted in greater ac- breast cancer should be stopped remains unan-
tivity than that with either antibody alone owing swered. With trastuzumab, continuing therapy
to more comprehensive signaling blockade, since after progression and adding a different chemo-
the antibodies bind to different HER2 epitopes.2,3 therapeutic agent resulted in prolonged progres-
The BO17929 study initially showed that the sion-free survival,23 and lapatinib added to tras­
combination of pertuzumab plus trastuzumab tuzumab was superior to lapatinib monotherapy
was active in patients with advanced breast with respect to progression-free and overall sur-
cancer following progression15 and subsequently vival.24 The current recommendation is to dis-
showed that it was the combination, rather than continue pertuzumab after progression, but clini-
pertuzumab alone, that was responsible for the cal trials are warranted to determine the most
activity.16 The efficacy of combination therapy effective duration of therapy.
was also shown among patients receiving neo- This analysis has several limitations. First,
adjuvant therapy, with the use of pertuzumab, patients who crossed over to receive pertuz­u­
trastuzumab, and docetaxel resulting in higher mab were included in the control group, which

732 n engl j med 372;8 nejm.org february 19, 2015

The New England Journal of Medicine

Downloaded from nejm.org on February 29, 2024. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
 Pertuzumab Combination in HER2-Positive Breast Cancer

in the intention-to-treat analysis favored the subgroup of patients who received adjuvant or
control group. Even so, the treatment effect fa- neoadjuvant trastuzumab was also small, proba-
vored pertuzumab in sensitivity analyses that bly because trastuzumab was not widely available
excluded the crossover group (hazard ratio, 0.55) for this indication during the recruitment for
and that censored data at the time of the first this trial. However, the hazard ratios for progres-
pertuzumab dose (hazard ratio, 0.63), since data sion-free and overall survival were similar (0.75
for patients with a good therapeutic response, and 0.80, respectively), suggesting a treatment
who had continued to receive study treatment benefit in favor of pertuzumab. One retrospec-
for a long time, were censored or excluded from tive study has shown better outcomes with
the control group but not from the pertuzumab trastuzumab in patients with metastatic breast
group. However, these analyses support the re- cancer who had received no previous adjuvant or
sults of the intention-to-treat analysis. neoadjuvant therapy with the drug,25 whereas a
Second, although the subgroups were pre- controlled cohort study showed no difference.26
defined, the analyses were exploratory and not Further research is therefore needed.
powered for a statistical result. Small numbers of In conclusion, in patients with HER2-positive
patients and data with wide confidence intervals metastatic breast cancer, the addition of pertuz­
limited interpretation of data among patients in umab to trastuzumab and docetaxel, as com-
the subgroup with nonvisceral disease, for ex- pared with the addition of placebo, significantly
ample. The hazard ratio for progression-free increased the median overall survival to 56.5
survival in this subgroup was 0.96 in the pri- months, an improvement of 15.7 months over
mary analysis,4 with an updated hazard ratio of survival in the control group, and extended the
0.83. A relatively small proportion of patients in results of previous analyses showing the efficacy
this subgroup died (58 of 178 patients [32.6%] of this drug combination.
overall), and survival was long, with a median
that was not reached in the pertuzumab group Supported by F. Hoffmann–La Roche and Genentech.
and that was 61.5 months in the control group Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
(data not shown). This study included a large We thank Daniel Clyde of Health Interactions for providing
proportion of patients with visceral disease. The writing assistance.

Appendix
The authors’ affiliations are as follows: the Washington Cancer Institute, MedStar Washington Hospital Center, Washington, DC
(S.M.S.); Memorial Sloan Kettering Cancer Center, Memorial Hospital, New York (J.B.); the Department of Oncology, Asan Medical
Center, University of Ulsan College of Medicine, Seoul (S.-B.K.), and the Center for Breast Cancer, National Cancer Center, Goyang
(J.R.) — both in South Korea; the N.N. Petrov Research Institute of Oncology, St. Petersburg, Russia (V.S.); Centre René Gauducheau,
Saint-Herblain (Nantes) (M.C.), and Centre Antoine Lacassagne, Nice (J.-M.F.) — both in France; 12 de Octubre University Hospital,
Medical Oncology Department, Madrid (E. Ciruelos), and Vall d’Hebron University Hospital, Vall d’Hebron Institute of Oncology, Bar-
celona (J.C.) — both in Spain; the National Center for Tumor Diseases, University Hospital, Heidelberg, Germany (A.S.); Roche Prod-
ucts, Welwyn, United Kingdom (S.H., E. Clark, G.R.); and Genentech, South San Francisco, CA (M.C.B.).

References
1. Ross JS, Slodkowska EA, Symmans 4. Baselga J, Cortés J, Kim SB, et al. Per- 7. Smith TJ, Khatcheressian J, Lyman
WF, Pusztai L, Ravdin PM, Hortobagyi GN. tuzumab plus trastuzumab plus docetaxel GH, et al. 2006 Update of recommenda-
The HER-2 receptor and breast cancer: ten for metastatic breast cancer. N Engl J Med tions for the use of white blood cell
years of targeted anti-HER-2 therapy and 2012;366:109-19. growth factors: an evidence-based clini-
personalized medicine. Oncologist 2009; 5. Swain SM, Kim SB, Cortés J, et al. Per- cal practice guideline. J Clin Oncol 2006;
14:320-68. tuzumab, trastuzumab, and docetaxel for 24:3187-205.
2. Nahta R, Hung MC, Esteva FJ. The HER2-positive metastatic breast cancer 8. Therasse P, Arbuck SG, Eisenhauer
HER-2-targeting antibodies trastuzumab (CLEOPATRA study): overall survival re- EA, et al. New guidelines to evaluate the
and pertuzumab synergistically inhibit sults from a randomised, double-blind, response to treatment in solid tumors.
the survival of breast cancer cells. Cancer placebo-controlled, phase 3 study. Lancet J Natl Cancer Inst 2000;92:205-16.
Res 2004;64:2343-6. Oncol 2013;14:461-71. 9. Slamon DJ, Leyland-Jones B, Shak S,
3. Scheuer W, Friess T, Burtscher H, 6. Swain SM, Ewer MS, Cortés J, et al. et al. Use of chemotherapy plus a mono-
Bossenmaier B, Endl J, Hasmann M. Cardiac tolerability of pertuzumab plus clonal antibody against HER2 for meta-
Strongly enhanced antitumor activity of trastuzumab plus docetaxel in patients static breast cancer that overexpresses
trastuzumab and pertuzumab combina- with HER2-positive metastatic breast can- HER2. N Engl J Med 2001;344:783-92.
tion treatment on HER2-positive human cer in CLEOPATRA: a randomized, double- 10. Marty M, Cognetti F, Maraninchi D,
xenograft tumor models. Cancer Res blind, placebo-controlled phase III study. et al. Randomized phase II trial of the ef-
2009;69:9330-6. Oncologist 2013;18:257-64. ficacy and safety of trastuzumab com-

n engl j med 372;8 nejm.org february 19, 2015 733

The New England Journal of Medicine
Downloaded from nejm.org on February 29, 2024. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.
 Pertuzumab Combination in HER2-Positive Breast Cancer

bined with docetaxel in patients with hu- during prior trastuzumab therapy. J Clin 21. Baselga J, Cortés J, Im S-A, et al. Bio-
man epidermal growth factor receptor Oncol 2010;28:1138-44. marker analyses in CLEOPATRA: a phase
2-positive metastatic breast cancer admin- 16. Cortés J, Fumoleau P, Bianchi GV, et III, placebo-controlled study of pertuzu­
istered as first-line treatment: the M77001 al. Pertuzumab monotherapy after tras­ mab in human epidermal growth factor
Study Group. J Clin Oncol 2005;23:4265- tuzumab-based treatment and subse- receptor 2–positive, first-line metastatic
74. quent reintroduction of trastuzumab: ac- breast cancer. J Clin Oncol 2014;32:3753-
11. Andersson M, Lidbrink E, Bjerre K, et tivity and tolerability in patients with 61.
al. Phase III randomized study comparing advanced human epidermal growth factor 22. Dang C, Iyengar N, Datko F, et al.
docetaxel plus trastuzumab with vinorel- receptor 2-positive breast cancer. J Clin Phase II study of paclitaxel given once per
bine plus trastuzumab as first-line thera- Oncol 2012;30:1594-600. week along with trastuzumab and per­
py of metastatic or locally advanced hu- 17. Gianni L, Pienkowski T, Im YH, et al. tuzumab in patients with human epider-
man epidermal growth factor receptor Efficacy and safety of neoadjuvant per­ mal growth factor receptor 2-positive
2-positive breast cancer: the HERNATA tuzumab and trastuzumab in women metastatic breast cancer. J Clin Oncol
study. J Clin Oncol 2011;29:264-71. with locally advanced, inflammatory, or 2014 December 29 (Epub ahead of print).
12. Valero V, Forbes J, Pegram MD, et al. early HER2-positive breast cancer (Neo- 23. von Minckwitz G, du Bois A, Schmidt
Multicenter phase III randomized trial Sphere): a randomised multicentre, M, et al. Trastuzumab beyond progression
comparing docetaxel and trastuzumab open-label, phase 2 trial. Lancet Oncol in human epidermal growth factor recep-
with docetaxel, carboplatin, and tras­ 2012;13:25-32. tor 2-positive advanced breast cancer:
tuzumab as first-line chemotherapy for 18. Amiri-Kordestani L, Wedam S, Zhang a German Breast Group 26/Breast Inter-
patients with HER2-gene-amplified meta- L, et al. First FDA approval of neoadjuvant national Group 03-05 study. J Clin Oncol
static breast cancer (BCIRG 007 study): therapy for breast cancer: pertuzumab for 2009;27:1999-2006.
two highly active therapeutic regimens. the treatment of patients with HER2-pos- 24. Blackwell KL, Burstein HJ, Storniolo
J Clin Oncol 2011;29:149-56. itive breast cancer. Clin Cancer Res 2014; AM, et al. Overall survival benefit with
13. Baselga J, Manikhas A, Cortés J, et al. 20:5359-64. lapatinib in combination with trastuz­
Phase III trial of nonpegylated liposomal 19. Schneeweiss A, Chia S, Hegg R, et al. umab for patients with human epidermal
doxorubicin in combination with tras­ Evaluating the predictive value of bio- growth factor receptor 2-positive meta-
tuzumab and paclitaxel in HER2-positive markers for efficacy outcomes in response static breast cancer: final results from the
metastatic breast cancer. Ann Oncol 2014; to pertuzumab- and trastuzumab-based EGF104900 Study. J Clin Oncol 2012;30:
25:592-8. therapy: an exploratory analysis of the 2585-92.
14. Cortés J, Baselga J, Im YH, et al. TRYPHAENA study. Breast Cancer Res 25. Murthy RK, Varma A, Mishra P, et al.
Health-related quality-of-life assessment 2014;16:R73. Effect of adjuvant/neoadjuvant trastuz­
in CLEOPATRA, a phase III study combin- 20. Gianni L, Bianchini G, Kiermaier A, umab on clinical outcomes in patients
ing pertuzumab with trastuzumab and et al. Neoadjuvant pertuzumab (P) and with HER2-positive metastatic breast can-
docetaxel in metastatic breast cancer. trastuzumab (H): biomarker analyses of a cer. Cancer 2014;120:1932-8.
Ann Oncol 2013;24:2630-5. 4-arm randomized Phase II study (‘Neo- 26. Negri E, Zambelli A, Franchi M, et al.
15. Baselga J, Gelmon KA, Verma S, et al. Sphere’) in patients (pts) with HER2-pos- Effectiveness of trastuzumab in first-line
Phase II trial of pertuzumab and tras­ itive breast cancer (BC). Presented at the HER2+ metastatic breast cancer after fail-
tuzumab in patients with human epider- 34th annual San Antonio Breast Cancer ure in adjuvant setting: a controlled co-
mal growth factor receptor 2-positive Symposium, San Antonio, TX, December hort study. Oncologist 2014;19:1209-15.
metastatic breast cancer that progressed 6–10, 2011. abstract. Copyright © 2015 Massachusetts Medical Society.

journal archive at nejm.org

Every article published by the Journal is now available at NEJM.org, beginning
with the first article published in January 1812. The entire archive is fully searchable,
and browsing of titles and tables of contents is easy and available to all.
Individual subscribers are entitled to free 24-hour access to 50 archive articles per year.
Access to content in the archive is available on a per-article basis and is also
being provided through many institutional subscriptions.

734 n engl j med 372;8 nejm.org february 19, 2015

The New England Journal of Medicine

Downloaded from nejm.org on February 29, 2024. For personal use only. No other uses without permission.
Copyright © 2015 Massachusetts Medical Society. All rights reserved.