Two-year overall survival rates from a randomised phase 2 trial
evaluating the combination of nivolumab and ipilimumab versus
ipilimumab alone in patients with advanced melanoma
F Stephen Hodi, MD
1
, Jason Chesney, MD
2,*
, Anna C Pavlick, MD
3
, Caroline Robert, MD
4,*
,
Kenneth F. Grossmann, MD
5
, David F McDermott, MD
6
, Gerald P Linette, MD
7
, Nicolas
Meyer, MD
8,*
, Jeffrey K Giguere, MD
9
, Sanjiv S Agarwala, MD
10,*
, Montaser Shaheen, MD
11
,
Marc S Ernstoff, MD
12,*
, David R Minor, MD
13
, April K Salama, MD
14
, Matthew H Taylor,
MD
15
, Patrick A Ott, MD
1
, Christine Horak, PhD
16
, Paul Gagnier, MD
16
, Joel Jiang, PhD
16
,
Jedd D Wolchok, MD
17,*
, and Michael A Postow, MD
17
1
Dana-Farber Cancer Institute, Boston, MA, USA
2
University of Louisville, Louisville, KY, USA
3
New York University, New York, NY, USA
4
Gustave Roussy, INSERM U981, Villejuif-Paris-Sud, France
5
Huntsman Cancer Institute, Salt Lake City, UT, USA
6
Beth Israel Deaconess Medical Center, Boston, MA, USA
7
Washington University School of Medicine, St Louis, MO, USA
8
Institut Universitaire du Cancer, Toulouse, France
9
Greenville Health System Cancer Institute, Greenville, SC, USA
10
St Luke’s Cancer Center and Temple University, Bethlehem, PA, USA
11
University of New Mexico, Albuquerque, NM, USA
12
Dartmouth-Hitchcock Medical Center, Lebanon, NH, USA
13
California Pacific Center for Melanoma Research, San Francisco, CA, USA
14
Duke University Medical Center, Durham, NC, USA
15
Oregon Health & Science University, Portland, OR, USA
Corresponding author: F Stephen Hodi, MD, Dana-Farber Cancer Institute, 450 Brookline Ave., Boston, MA 02115, USA,
Stephen_Hodi@dfci.harvard.edu, Phone: 617-632-4715, Fax: 617-632-6727.
*
Indicates author is a full professor.
JDW and MAP contributed equally to this study.
Contributors
FSH contributed to the conception and design of the study, data collection, data interpretation, and writing of the report. JC, ACP, KG,
DFMcD, GL, NM, JFG, SSA, MS, MSE, DRM, AKS, MHT, and PAO contributed to data collection and data interpretation. CR
contributed to the conception and design of the study, data collection, and data interpretation. CH was the biomarker lead for the study.
PG was the medical monitor for the study and JJ was the lead statistician for the current analyses. JDW and MAP contributed to the
conception and design of the study, data collection, data interpretation, and writing of the report.
Declaration of interests
KG, GL, JKG, SSA, and MS declare no competing interests.
HHS Public Access
Author manuscript
Lancet Oncol
. Author manuscript; available in PMC 2017 October 07.
Published in final edited form as:
Lancet Oncol
. 2016 November ; 17(11): 1558–1568. doi:10.1016/S1470-2045(16)30366-7.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
16
Bristol-Myers Squibb, Princeton, NJ, USA
17
Memorial Sloan Kettering Cancer Center and Weill Cornell Medical College, New York, NY, USA
Summary
Background—Previously reported results of phase 2 and phase 3 trials showed a significant
improvement in the rate of objective response and progression-free survival with nivolumab (anti-
PD-1 antibody) plus ipilimumab (anti-CTLA-4 antibody) vs ipilimumab alone in patients with
advanced melanoma. To our knowledge, this is the first report of overall survival data from a
randomised, controlled trial evaluating the combination of nivolumab and ipilimumab in advanced
melanoma.
Methods—In this phase 2 trial (CheckMate 069), 142 patients aged ≥18 years with previously
untreated, unresectable stage III or IV melanoma, with an Eastern Cooperative Oncology Group
performance status of 0 or 1, were randomly assigned 2:1 to receive an intravenous infusion of
nivolumab 1 mg/kg plus ipilimumab 3 mg/kg or ipilimumab 3 mg/kg plus placebo, every 3 weeks
for 4 doses, followed by nivolumab 3 mg/kg or placebo, respectively, every 2 weeks until disease
progression or unacceptable toxicity. Randomisation was done by an interactive voice response
system with a permuted block schedule and stratification by
BRAF
mutation status. The primary
endpoint (previously reported) was the rate of investigator-assessed objective response among
patients with
BRAF
V600 wild-type melanoma. Overall survival was an exploratory endpoint.
Efficacy analyses were done on the intention-to-treat population, where safety was evaluated in all
treated patients. This study is registered with ClinicalTrials.gov, number NCT01927419, and is
ongoing but no longer enrolling patients.
Findings—Between September 16, 2013, and February 6, 2014, we screened 179 patients,
randomly allocating 95 patients to nivolumab plus ipilimumab and 47 to ipilimumab (72 [76%]
and 37 [79%] patients with
BRAF
V600 wild-type tumors, respectively). At a median follow-up of
24 months, overall survival rates in all randomized patients were 63·8% (95% CI 53·3–72·6) for
nivolumab plus ipilimumab vs 53·6% (95% CI 38·1–66·8) for ipilimumab alone; median overall
survival had not been reached in either group (hazard ratio 0·74, 95% CI 0·43–1·26; p=0.26).
Grade 3–4 adverse events related to nivolumab plus ipilimumab were reported in 51 [54%] of 94
patients vs 9 [20%] of 46 patients related to ipilimumab alone. The most common treatment-
related grade 3–4 adverse events in the combination group were colitis (12 [13%] of 94 patients)
and increased alanine aminotransferase (10 [11%]), and for ipilimumab alone, were diarrhoea (five
[11%] of 46 patients) and hypophysitis (two [4%]). Serious grade 3–4 adverse events related to
nivolumab plus ipilimumab were reported in 34 [36%] of 94 patients vs 4 [9%] of 46 patients
related to ipilimumab alone, which included colitis (10 [11%]) and diarrhoea (5 [5%]) in the
combination group and diarrhoea (2 [4%]), colitis (1 [2%]), and hypophysitis (1[2%]) in the
ipilimumab alone group.
Interpretation—While follow-up of the patients continues, the results of this analysis suggest
that the combination of first-line nivolumab plus ipilimumab may lead to a higher overall survival
rate vs first-line ipilimumab in patients with advanced melanoma. The results suggest encouraging
survival outcomes with immunotherapy in this patient population.
Funding—Bristol-Myers Squibb.
Hodi et al.
Page 2
Lancet Oncol
. Author manuscript; available in PMC 2017 October 07.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Introduction
Survival outcomes for patients with advanced melanoma have, historically, been very poor,
with a median overall survival of ~8 months and a 5-year survival rate from diagnosis of
metastatic disease of ~10%.
1
Ipilimumab, which blocks cytotoxic T-lymphocyte antigen-4,
was the first agent to demonstrate an improvement in overall survival in a randomised,
controlled, phase 3 trial of patients with advanced melanoma.
2
In this phase 3 trial, the two-
year overall survival rate of ipilimumab-treated patients was 25·3%.
3
A pooled analysis of
data from 12 clinical trials in advanced melanoma, in which some ipilimumab-treated
patients were followed up to 10 years, showed durable long-term overall survival with a 3-
year survival rate of 22%.
4
Newer immune checkpoint inhibitors, which block the
programmed death 1 receptor, include nivolumab and pembrolizumab. In a phase 3 trial
(CheckMate 066), nivolumab monotherapy demonstrated an improvement in overall survival
vs dacarbazine in treatment-naïve patients with
BRAF
wild-type tumours.
5
Follow-up of
patients in this study has shown 2-year overall survival rates of 58% with nivolumab and
27% with dacarbazine.
6
Both nivolumab and pembrolizumab monotherapy have demonstrated superior efficacy
outcomes compared with ipilimumab alone in phase 3 trials of advanced melanoma.
7,8
In a
phase 2 trial of treatment-naïve patients with
BRAF
wild-type melanoma (CheckMate 069),
the combination of nivolumab and ipilimumab demonstrated a statistically significant
improvement in objective response rate and longer progression-free survival compared with
ipilimumab alone.
9
More recently, the results of a phase 3 trial (CheckMate 067) also
showed that nivolumab in combination with ipilimumab leads to longer progression-free
survival and a higher objective response rate than is achieved with ipilimumab alone in
treatment-naïve patients with advanced melanoma.
7
In a phase 1 dose-finding study of
nivolumab in combination with ipilimumab, patient follow-up has shown a 3-year overall
survival rate of 68% in previously treated and untreated patients with advanced melanoma.
10
Here we present 2-year overall survival data from the CheckMate 069 trial, which, to our
knowledge, is the first report of overall survival for the combination of nivolumab and
ipilimumab from a randomised, controlled trial in advanced melanoma.
Methods
Study design and patients
In this randomised, controlled, double-blind, phase 2 study, we recruited patients at 19 sites
in 2 countries (France, United States). Eligible patients were aged 18 years or older and had
histologically confirmed, unresectable stage III or stage IV metastatic melanoma with an
Eastern Cooperative Oncology Group performance status of 0 or 1, and known
BRAF
V600
mutation status. Patients were also required to have measurable disease by CT or MRI as per
Response Evaluation Criteria in Solid Tunors version 1.1 (RECIST v1.1) criteria, and to
provide tumour tissue adequate for biomarker analyses (assessment of PD-L1). We excluded
patients with active brain metastases or leptomeningeal metastases and those with ocular
melanoma Patients with mucosal melanoma were allowed to enroll. Patients who had
received prior systemic anticancer therapy for unresectable or metastatic melanoma were
excluded, but prior adjuvant or neoadjuvant treatment for melanoma was permitted (if
Hodi et al.
Page 3
Lancet Oncol
. Author manuscript; available in PMC 2017 October 07.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
completed at least 6 weeks prior to the date of first dose), and all related adverse events
either returned to baseline or stabilized.
BRAF
V600
mutation testing was done during the
screening period using a US Food and Drug Administration-approved test.
Patients provided consent to participate in this study, including follow-up for survival
outcomes, using written informed consent forms. The protocol, amendments, and patient
consent forms were approved by the Institutional Review Board/Independent Ethics
Committee at each study site prior to initiation of the trial.
Randomisation and masking
Patients were randomly allocated 2:1 to either nivolumab plus ipilimumab or ipilimumab
alone using an interactive voice response system. Once enrolled in the interactive voice
response system, patients that met all eligibility criteria were randomised provided that the
following information was provided: (1) patient number, (2) date of birth, and (3)
BRAF
V600 mutation status. We stratified randomisation by
BRAF
mutation status (V600
mutation-positive vs V600 wild-type). Randomisation procedures were carried out using
permuted blocks within each stratum. The study funder, patients, investigators, and study
site staff were blinded to the study drug administered. Each study site assigned an unblinded
pharmacist/designee who called the interactive voice response system to obtain the treatment
assignment for each patient. During the blinded portion of the study, unblinding of study
treatment could occur if the investigator determined it is necessary for immediate medical
management of the patient (who must have discontinued study treatment) or upon confirmed
disease progression. Upon disease progression and after unblinding, patients in the
ipilimumab group had the option to receive nivolumab 3 mg/kg every 2 weeks until further
disease progression.
Procedures
In the combination group, nivolumab was administered intravenously at a dose of 1 mg/kg
over a period of 60 minutes, once every 3 weeks for four doses. Thirty minutes after the
completion of each nivolumab infusion, patients received ipilimumab at 3 mg/kg over a
period of 90 minutes. After the fourth dose of both agents (induction phase), ipilimumab
was discontinued and nivolumab was then administered as a single agent at 3 mg/kg over a
period of 60 minutes, once every 2 weeks (maintenance phase). In the ipilimumab alone
group, the same dosing schedule was used, except that nivolumab was replaced with
matched placebo during both the combination and maintenance portions of the trial.
Treatment was continued as long as clinical benefit (as defined by the investigator) was
observed, until there were unacceptable side effects, patient request to stop study treatment
or withdrawal of consent, pregnancy, or termination of the study by the sponsor Dosing
interruptions were allowed for the management of adverse events, but dose reductions were
not permitted. Patients who experienced disease progression, and were tolerating study
therapy, could be treated beyond progression (with blinding maintained) or have blinded
study therapy discontinued. After unblinding, patients in the ipilimumab alone group could
receive nivolumab at 3 mg/kg every 2 weeks until further disease progression; patients in the
combination group were required to discontinue treatment.
Hodi et al.
Page 4
Lancet Oncol
. Author manuscript; available in PMC 2017 October 07.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Tumour response was assessed by the investigators by CT or MRI according to RECIST
v1.1 criteria at the following time points: within 28 days prior to the first dose (baseline), 12
weeks after the first treatment, every 6 weeks thereafter for the first year, then every 12
weeks until disease progression or discontinuation of treatment. Responses required
confirmation through a subsequent scan at least 4 weeks later. Safety evaluations were
performed in patients who had received at least one dose of study treatment, and the severity
of adverse events was graded according to the National Cancer Institute Common
Terminology Criteria for Adverse Events, version 4.0.Safety was evaluated in all patients
who received at least one dose of study drug, and an adverse events was considered on-study
if it occurred within 30 days after the last dose of study treatment. On-study laboratory
assessments (including chemistry and haematology tests) were done within 72 hours prior to
each dose during the induction phase; during the nivolumab maintenance phase, laboratory
assessments were done within 72 hours prior to the first dose and every alternate dose
thereafter. Tumour expression of PD-L1 was assessed in pretreatment samples at a central
laboratory with the use of a validated, automated immunohistochemical assay (Bristol-
Myers Squibb and Dako), as described.
11
In each tumour tissue sample, at least 5% of
tumour cells showing cell-surface PD-L1 staining of any intensity in a section containing at
least 100 evaluable tumour cells was required to determine PD-L1–positivity.
Outcomes
The primary endpoint was the rate of confirmed objective response among patients with
BRAF
V600 wild-type tumours (as assessed by the investigators). Secondary endpoints
included investigator-assessed progression-free survival in patients with
BRAF
wild-type
tumours, the objective response rate and progression-free survival among patients with
BRAF
V600 mutation–positive tumours, and safety. Overall survival was an exploratory
endpoint.
Statistical analysis
We planned a sample size of 100 patients with
BRAF
wild-type tumours randomised 2:1 to
the two treatment groups. Given a two-sided alpha of 0.05, this number of patients provided
an 87% power to detect a significant difference in objective response rate between the
groups, assuming an objective response rate of 40% with the combination and 10% with
ipilimumab alone.
9
Assuming that 66% of the patients had
BRAF
wild-type tumours, we
planned to randomise approximately 150 patients (50 with
BRAF
mutation-positive
tumours). Analyses in the population with
BRAF
mutation-positive tumours were intended
to be descriptive only and thus were not part of the sample size calculations.
9
A hierarchical
testing approach was applied to key secondary endpoints after analysis of the primary
endpoint: the objective response rate among all randomly assigned patients was tested first,
followed by progression-free survival among randomized patients with
BRAF
wild-type
tumours, and then progression-free survival among all randomly assigned patients.
For the primary endpoint, the comparison of investigator-assessed objective response rate
between treatment groups was performed using Fisher’s exact test. An associated odds ratio
and corresponding two-sided 95% CI were calculated. Time-to-event distributions (i.e.,
progression-free survival, overall survival, time to response, and duration of response) and
Hodi et al.
Page 5
Lancet Oncol
. Author manuscript; available in PMC 2017 October 07.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript