On November 23, 2015, the FDA approved nivolumab (OPDIVO; Bristol-Myers Squibb) as a single agent for the first-line treatment of patients with BRAFV600 wild-type, unresectable or metastatic melanoma. An international, double-blind, randomized (1:1) trial conducted outside of the United States allocated 418 patients to receive nivolumab 3 mg/kg intravenously every 2 weeks (n = 210) or dacarbazine 1,000 mg/m2 intravenously every 3 weeks (n = 208). Patients with disease progression who met protocol-specified criteria (∼25% of each trial arm) were permitted to continue with the assigned treatment in a blinded fashion until further disease progression is documented. Overall survival was statistically significantly improved in the nivolumab arm compared with the dacarbazine arm [hazard ratio (HR), 0.42; 95% confidence interval (CI), 0.30–0.60; P < 0.0001]. Progression-free survival was also statistically significantly improved in the nivolumab arm (HR, 0.43; 95% CI, 0.34–0.56; P < 0.0001). The most common adverse reactions (≥20%) of nivolumab were fatigue, diarrhea, constipation, nausea, musculoskeletal pain, rash, and pruritus. Nivolumab demonstrated a favorable benefit–risk profile compared with dacarbazine, supporting regular approval; however, it remains unclear whether treatment beyond disease progression contributes to the overall clinical benefit of nivolumab. Clin Cancer Res; 23(14); 3479–83. ©2017 AACR.

Melanoma is the fifth leading cancer type in men and seventh in women, with an estimated 76,380 new cases and 10,130 deaths in 2016 (1). The prognosis of patients with metastatic melanoma is poor, with an estimated 5-year survival of approximately 15% (2). With the approval of seven new drugs since 2011 as systemic treatments for patients with unresectable or metastatic melanoma—ipilimumab, pembrolizumab, and nivolumab for unresectable metastatic melanoma (regardless of BRAFV600 mutation status) and vemurafenib, dabrafenib, cobimetinib, and trametinib for BRAFV600 mutation–positive unresectable metastatic melanoma—the treatment paradigm for melanoma has rapidly evolved (3).

Nivolumab is a human IgG4 programmed death receptor-1 (PD-1) mAb that blocks the interaction between PD-1 and its ligands, PD-L1 and PD-L2, releasing PD-1 pathway–mediated inhibition of the antitumor immune response. The FDA granted accelerated approval to nivolumab in December 2014 based on demonstration of an objective response rate (ORR) of 32% [95% confidence interval (CI), 23–41] with prolonged durations of response in a non-comparative, interim analysis of the first 120 patients with unresectable or metastatic melanoma who had progressed following treatment with ipilimumab and, if BRAFV600 mutation positive, a BRAF inhibitor regimen, and with a minimum follow-up of 6 months randomized to receive nivolumab in an ongoing, randomized clinical trial (4, 5). Under the accelerated approval provisions, submission of a clinical trial or trials confirming the clinical benefit of nivolumab was required (6). The results of the trial submitted to confirm clinical benefit, which formed the FDA basis for the regular approval of nivolumab for the first-line treatment of patients with BRAFV600 wild-type (WT) unresectable or metastatic melanoma, are described in this article.

The safety and efficacy of nivolumab was evaluated in a single clinical trial, trial CA209066, an international, double-blind, randomized controlled trial conducted in patients with BRAFV600 WT unresectable or metastatic melanoma who had not previously received treatment for their disease (7). The primary endpoint of the trial was overall survival (OS), with key secondary efficacy endpoints of progression-free survival (PFS) and ORR as assessed by the investigator per RECIST v1.1 (8). The trial excluded patients with ocular melanoma, Eastern Cooperative Oncology Group (ECOG) performance status >1, active brain or leptomeningeal metastases, and autoimmune disease and patients requiring chronic systemic treatment with corticosteroids (>10 mg daily prednisone equivalent) or other immunosuppressive medications.

The trial randomized (1:1) patients to receive nivolumab 3 mg/kg by intravenous infusion every 2 weeks or dacarbazine 1,000 mg/m2 by intravenous infusion every 3 weeks. Patients also received either nivolumab- or dacarbazine-matched placebo infusions every 2 or 3 weeks, respectively, to maintain the treatment blind. Randomization stratification factors were PD-L1 status [positive (≥5% tumor cell membrane staining) vs. negative (<5% tumor cell membrane staining) or indeterminate (tumor cell membrane scoring hampered by high cytoplasmic staining or melanin content)] and M stage (M0/M1a/M1b versus M1c).

Patients were allowed to remain on the assigned therapy in a blinded fashion beyond initial investigator-assessed disease progression per RECIST v1.1 if the patient was experiencing “investigator-assessed clinical benefit and tolerating study therapy.” Such patients were subsequently required to discontinue therapy when further disease progression, defined as an additional 10% or greater increase in tumor burden volume from time of initial progression (including all target lesions and new measurable lesions), was documented by the investigator.

Trial CA209066 was designed with 90% power and an overall type I error of 0.05 (two sided) to detect a hazard ratio (HR) of 0.69 (assuming a median OS of 10 months in the dacarbazine arm and 14.49 months in the nivolumab arm) following the occurrence of at least 312 deaths. One interim analysis for efficacy was planned following the occurrence of 218 deaths (70% of total deaths needed for final analysis). Lan–DeMets alpha spending function with O'Brien–Fleming boundaries was used for adjustment of the type I error rate at the interim and final analysis. In addition to the formal planned interim analysis for OS, the data monitoring committee would have access to periodic unblinded interim reports of efficacy and safety to allow a risk–benefit assessment.

The trial was conducted entirely outside of the United States; a total of 418 patients were enrolled and randomized to nivolumab (n = 210) or dacarbazine (n = 208). There were 411 patients who received at least one dose of study-specified treatment. Among the intent-to-treat population (all randomized), the majority (69%) was enrolled in Western Europe and Canada, 59% were male, 99.5% were white, and median age was 65 years (range, 18–87 years). More patients in the nivolumab arm had an ECOG performance status of 0 (71% vs. 58%). Other demographics and disease characteristics appeared to be similar between arms. The majority of patients had a diagnosis of cutaneous melanoma (74%), 61% had M1c stage disease, 37% had an elevated lactate dehydrogenase (LDH) level, and 35% had PD-L1 greater than or equal to 5% tumor cell membrane expression (35%) according to a proprietary clinical trial assay. No patients had previously received treatment in the metastatic setting, and all were BRAFV600 WT by regionally acceptable V600 mutational status testing. At the time of the database lock, no patients in the control arm had crossed over to receive nivolumab, although this was allowed in a subsequent protocol amendment. A greater proportion of patients in the dacarbazine arm compared with the nivolumab arm received subsequent anticancer therapy (55% vs. 30%, respectively), and a greater proportion of patients in the dacarbazine arm received subsequent ipilimumab (38% vs. 21%, respectively).

Efficacy

The trial was halted prior to the planned interim analysis based on the recommendations of an independent data monitoring committee following an unplanned interim analysis, which demonstrated an improvement in OS based on 110 deaths (35% of the total planned events for the final analysis), which crossed the O'Brien–Fleming boundary based on 110 events and assuming two interim analyses. The results presented in this report were based on a clinical cutoff date of June 24, 2014, with 146 deaths (47% of the total planned events) in the study report generated by Bristol-Myers Squibb. A statistically significant and clinically meaningful improvement in OS (HR, 0.42; 95% CI, 0.30–0.60; P < 0.0001) was observed in the nivolumab arm as compared with the dacarbazine arm; the median OS was not reached in the nivolumab arm and was 10.8 months in the dacarbazine arm (see Table 1). Key secondary endpoints also demonstrated statistically significant improvements in PFS and ORR favoring the nivolumab arm (see Table 1). At the time of the FDA analysis, 88% (63/72) of responding patients in the nivolumab arm had ongoing responses; 43 of the 72 responding patients had ongoing response of 6 months or longer. In exploratory analyses based on demographics and baseline disease characteristics, all subgroups, including PD-L1 positive and negative/indeterminate status, appeared to achieve an improvement in OS with nivolumab treatment.

Table 1.

Efficacy results from trial CA209066

Nivolumab (n = 210)Dacarbazine (n = 208)
OS 
 Events (%) 50 (24) 96 (46) 
 Median, months (95% CI) Not reached 10.8 (9.3–12.1) 
 HR (95% CI) 0.42 (0.30–0.60) 
P <0.0001a 
PFS 
 Events (%) 108 (51) 163 (78) 
 Median, months (95% CI) 5.1 (3.5–10.8) 2.2 (2.1–2.4) 
 HR (95% CI) 0.43 (0.34–0.56) 
P <0.0001a 
ORR 34% 9% 
95% CI 28–41 5–13 
Complete response rate 4% 1% 
Partial response rate 30% 8% 
Nivolumab (n = 210)Dacarbazine (n = 208)
OS 
 Events (%) 50 (24) 96 (46) 
 Median, months (95% CI) Not reached 10.8 (9.3–12.1) 
 HR (95% CI) 0.42 (0.30–0.60) 
P <0.0001a 
PFS 
 Events (%) 108 (51) 163 (78) 
 Median, months (95% CI) 5.1 (3.5–10.8) 2.2 (2.1–2.4) 
 HR (95% CI) 0.43 (0.34–0.56) 
P <0.0001a 
ORR 34% 9% 
95% CI 28–41 5–13 
Complete response rate 4% 1% 
Partial response rate 30% 8% 

aP value is compared with the allocated alpha of 0.0021 for this interim analysis.

The FDA conducted exploratory summary analyses characterizing patient characteristics and outcomes for those treated beyond investigator-assessed, RECIST-defined progression. Among the 411 treated patients, 94 were treated beyond progression (defined as at least one dose of nivolumab or dacarbazine after initial documentation of disease progression as assessed by the investigator per RECIST v1.1). Because this subgroup is not a randomized subgroup of patients, no inference may be drawn. Of these patients, 51 were in the nivolumab arm, and 43 were in the dacarbazine arm. A summary of the demographics for the subgroups of patients treated beyond progression and those not treated beyond progression is shown in Table 2. The majority of patients in both arms who were treated beyond progression progressed at the first assessment—69% on the nivolumab arm and 74% on the dacarbazine arm. The median duration of treatment beyond progression in the nivolumab arm was 42 days (minimum–maximum, 2–415) and in the dacarbazine arm 28 days (minimum–maximum, 2–111; see Fig. 1). The median number of doses (including both study drug and matched placebo) in each arm was 5 (95% CI, 4–6; range, 1–46) in the nivolumab arm and 3 (95% CI, 2–5; range, 1–18) in the dacarbazine arm. In the group of patients treated beyond progression, 58.8% and 53.5% of patients on the nivolumab and dacarbazine arms, respectively, had either a decrease in tumor burden or no increase in tumor burden compared with baseline for at least one assessment period; the remaining 41.2% and 46.5% had continued increase in tumor burden (see Fig. 2).

Table 2.

Demographics of subgroups by treatment beyond progression from trial CA209066

Patients who progressed and were treated beyond progression (n = 94)Patients who progressed and were not treated beyond progression (n = 167)
Age (in years) 
 Median 63.5 66 
 Min–Max 18–84 25–86 
Gender—Male (%) 59 58 
PD-L1 status (%) 
 Positive 33 32.9 
 Neg/Indeter. 67 67.1 
M stage (%) 
 M1c 60.6 58.7 
 M0/M1a/M1b 39.4 41.3 
AJCC (%) 
 Stage III 8.5 12 
 Stage IV 91.5 88 
Baseline ECOG PS (%) 
 0 74.5 65.3 
 1 25.5 34.1 
 2 0.6 
ECOG at time of progressiona (%) 
 Number of patients 89/94 132/167 
 0 70.8 55.3 
 1 28.1 41.7 
 2 1.1 
Baseline LDH (%) 
 ≤ULN 64.9 61.7 
 >ULN 35.1 35.3 
LDH at time of progressiona (%) 
 Number of patients 77/94 128/167 
 ≤ULN (%) 70.1 51.6 
 >ULN (%) 29.9 48.4 
Patients who progressed and were treated beyond progression (n = 94)Patients who progressed and were not treated beyond progression (n = 167)
Age (in years) 
 Median 63.5 66 
 Min–Max 18–84 25–86 
Gender—Male (%) 59 58 
PD-L1 status (%) 
 Positive 33 32.9 
 Neg/Indeter. 67 67.1 
M stage (%) 
 M1c 60.6 58.7 
 M0/M1a/M1b 39.4 41.3 
AJCC (%) 
 Stage III 8.5 12 
 Stage IV 91.5 88 
Baseline ECOG PS (%) 
 0 74.5 65.3 
 1 25.5 34.1 
 2 0.6 
ECOG at time of progressiona (%) 
 Number of patients 89/94 132/167 
 0 70.8 55.3 
 1 28.1 41.7 
 2 1.1 
Baseline LDH (%) 
 ≤ULN 64.9 61.7 
 >ULN 35.1 35.3 
LDH at time of progressiona (%) 
 Number of patients 77/94 128/167 
 ≤ULN (%) 70.1 51.6 
 >ULN (%) 29.9 48.4 

Abbreviations: AJCC, American Joint Committee on Cancer; Min–Max, minimum–maximum; Neg/Indeter., negative/indeterminate; PS, performance status; ULN, upper limit of normal.

aLDH test or ECOG status performed/recorded within 4 weeks before date of progressive disease.

Figure 1.

Swimmer plot of patients treated beyond progression on trial CA209066.

Figure 1.

Swimmer plot of patients treated beyond progression on trial CA209066.

Close modal
Figure 2.

Target lesion tumor burden change from baseline in patients treated beyond progression on trial CA209066. A, Nivolumab-treated patients. B, Dacarbazine-treated patients.

Figure 2.

Target lesion tumor burden change from baseline in patients treated beyond progression on trial CA209066. A, Nivolumab-treated patients. B, Dacarbazine-treated patients.

Close modal

Of the 51 patients treated beyond RECIST-defined progression with nivolumab, 4 (7.8%) subsequently developed a reduction of greater than 30% in the sum of the diameters of all target lesions compared with baseline. Three of these patients had progression at first assessment (new lesion and target lesion progression, new lesion, and unequivocal progression in nontarget lesion) and, subsequently, with continuation of nivolumab had 46%, 46%, or 33% reduction in target lesions by weeks 21, 49, and 50, respectively. These patients' responses were durable and ongoing at 4.5, 3, and 6 months, respectively, at the time of data cutoff. The fourth patient met criteria for RECIST v1.1–defined progression at the second tumor response assessment (week 16) based on development of a new lesion as well as an increase in the size of target lesions and then experienced a 67% decrease in the size of target lesions at the following assessment; however, the target lesion sum of reference diameters was 15 mm at baseline, and this patient had progression in nontarget lesions at the time of 67% decrease in target lesions and so was removed from treatment. In the dacarbazine arm, 1 of 43 patients had a 30% or greater reduction in target lesions after progression with continuation of dacarbazine. This dacarbazine-treated patient had progression at first assessment with a new lesion and, subsequently, had target lesion reduction of 52% as well as a new lesion at next assessment and was, therefore, taken off treatment.

Safety

In trial CA209066, 411 patients received treatment with nivolumab (n = 206) or dacarbazine (n = 205). The median duration of exposure was 6.5 months (range, 1 day–16.6 months) in patients treated with nivolumab. The most common adverse reactions, which occurred in at least 20% of nivolumab-treated patients, were fatigue, diarrhea, constipation, nausea, musculoskeletal pain, rash, and pruritus. Adverse reactions led to permanent discontinuation of nivolumab in 7% of patients and dose interruption in 26% of patients; no single organ-specific adverse reaction accounted for the majority of discontinuations. Serious adverse reactions and grade 3/4 adverse reactions occurred in 36% and 41%, respectively, of nivolumab-treated patients. Immune-mediated adverse reactions occurring in nivolumab-treated patients in this trial were consistent in incidence with those observed across the nivolumab development program and included hypothyroidism (7%), immune-mediated colitis/diarrhea (4.9%), hyperthyroidism (4.4%), pancreatitis (1.9%), pneumonitis (1.4%), diabetes mellitus (0.9%), immune-mediated hepatitis (0.9%), uveitis (0.9%), immune-mediated renal dysfunction (0.5%), facial paresis (0.5%), and Guillain–Barre syndrome (0.5%).

The results of the CA209066 trial clearly demonstrated the clinical benefit of nivolumab as a first-line treatment of BRAFV600 WT unresectable or metastatic melanoma based on an improvement in OS, which appeared to be present across subgroups defined by demographic and baseline disease characteristics. Analyses of secondary efficacy endpoints of PFS and ORR were supportive of the clinical benefit of nivolumab. The safety profile of nivolumab was acceptable for the patient population.

As the trial was conducted entirely outside of the United States, the FDA needed to determine whether the data were applicable to the U.S. population with consideration of the population enrolled, appropriateness of the control arms, endpoints, and relevance to U.S. medical practice in the context of available and subsequent therapies. Dacarbazine was not commonly used as first-line therapy for metastatic melanoma in the United States during the conduct of the trial, given the availability of ipilimumab, nivolumab, and pembrolizumab and given dacarbazine's modest activity. Thus, the FDA advised against dacarbazine as a comparator arm based on standard U.S. medical practice. The trial was conducted outside the United States and included locations where dacarbazine was considered standard first-line treatment and ipilimumab was approved only in the second-line setting. Despite this, the demographics and baseline disease characteristics of the enrolled patients were consistent with that expected in the U.S. patient population. The trial's primary endpoint of OS is appropriate and clinically meaningful. The FDA review confirmed that trial CA209066 met its primary endpoint with demonstration of a clinically important improvement in OS in patients receiving nivolumab, which was observed in the context that almost 40% of the patients treated with dacarbazine went on to receive ipilimumab and were, therefore, exposed to a subsequent immunotherapy. Altogether, the FDA determined that the trial data were generalizable to the U.S. population, supporting an expanded indication for nivolumab in patients with BRAFV600 WT unresectable or metastatic melanoma.

In light of the evolving information regarding the unusual patterns of tumor response to immune-oncology therapy, the FDA conducted exploratory analyses to evaluate the incidence and characteristics of tumor responses in patients treated beyond progression in both treatment arms. Trial CA209066 allowed evaluation of the uncommon scenario in which both the blinded immunotherapy and chemotherapy arms could be continued beyond progression. The FDA's analysis demonstrated that a limited number of patients in both treatment arms [4/210 (1.9%) patients in the nivolumab arm and 1/208 (0.4%) in the dacarbazine arm] achieved a clinically meaningful reduction in tumor based on investigator-assessed target lesion reduction of greater than 30% after initial progression. Notably, the best response to initial treatment was progressive disease in all 5 patients, with 4 of 5 demonstrating disease progression at the first assessment and one at the second assessment following initiation of treatment. Previously published analyses of the same trial using different criteria reported that 17 patients on the nivolumab arm and 8 on the dacarbazine arm maintained or achieved a reduction in target lesions of 30% or greater (7). The difference in results comes from the fact that the analysis by Roberts and colleagues included patients who developed a best reduction of target lesions of 30% or greater prior to (or at the same time as) first documentation of progression and prior to treatment beyond progression, whereas our analysis only included patients with target lesion reductions of 30% or greater after treatment beyond progression was initiated. The different criteria used for describing response in patients treated beyond progression can be confusing; until uniform criteria are established and the outcomes are characterized, any inference about clinical benefit of treatment beyond progression is premature. Further investigation is required to determine which, if any, patients should be treated beyond progression and to demonstrate the contribution of this approach to the overall treatment effect of nivolumab.

In conclusion, nivolumab demonstrated a statistically significant and clinically meaningful improvement in OS compared with dacarbazine in a population of patients with BRAFV600 WT unresectable or metastatic melanoma who had not previously received treatment for their advanced disease. Secondary endpoints were supportive of clinical benefit, and the safety profile was acceptable for the patient population studied. Therefore, nivolumab met requirements for regular approval in patients with BRAFV600 WT unresectable or metastatic melanoma with a positive benefit–risk analysis.

No potential conflicts of interest were disclosed.

The Deputy Editor handling the peer review and decision-making process for this article has no relevant employment associations to disclose.

Conception and design: J.A. Beaver, M.R. Theoret, R. Pazdur

Development of methodology: J.A. Beaver, R. Pazdur

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Pazdur

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.A. Beaver, M.R. Theoret, S. Mushti, K. He, R. Sridhara, P. Keegan, R. Pazdur

Writing, review, and/or revision of the manuscript: J.A. Beaver, M.R. Theoret, S. Mushti, K. He, K. Goldberg, R. Sridhara, A.E. McKee, P. Keegan

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.A. Beaver, M. Libeg, R. Pazdur

1.
Siegel
R
,
Miller
K
,
Jemal
A
. 
Cancer statistics, 2016
.
CA Cancer J Clin
2016
;
66
:
7
30
.
2.
NCI
. 
SEER statistics: melanoma of the skin
.
Rockville (MD)
:
NCI
. Available from: http://seer.cancer.gov/statfacts/html/melan.html.
3.
U.S. Food and Drug Administration
. 
Drugs@FDA: FDA Approved Drug Products
.
Silver Spring (MD)
:
U.S. Food and Drug Administration
. Available from: https://www.accessdata.fda.gov/scripts/cder/drugsatfda/.
4.
Weber
JS
,
D'Angelo
SP
,
Minor
D
,
Hodi
FS
,
Gutzmer
R
,
Neyns
B
, et al
Nivolumab versus chemotherapy in patients with advanced melanoma who progressed after anti-CTLA-4 treatment (CheckMate 037): a randomised, controlled, open-label, phase 3 trial
.
Lancet Oncol
2015
;
16
:
375
84
.
5.
Hazarika
M
,
Chuk
MK
,
Theoret
MR
,
Mushti
S
,
He
K
,
Weis
SL
, et al
U.S. FDA approval summary: nivolumab for treatment of unresectable or metastatic melanoma following progression on ipilimumab
.
Clin Cancer Res
2017
;
23
:
3484
88
.
6.
Approval based on a surrogate endpoint or on an effect on a clinical endpoint other than survival or irreversible morbidity, 21 C.F.R. Sect. 601.41 (2016)
.
7.
Robert
C
,
Long
GV
,
Brady
B
,
Dutriaux
C
,
Maio
M
,
Mortier
L
, et al
Nivolumab in previously untreated melanoma without BRAF mutation
.
N Engl J Med
2015
;
372
:
320
9
.
8.
Eisenhauer
EA
,
Therasse
P
,
Bogaerts
J
,
Schwartz
LH
,
Sargent
D
,
Ford
R
et al 
New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1)
.
Eur J Cancer
2009
;
45
:
228
47
.