Survival Outcomes in Patients With Previously Untreated BRAF Wild-Type Advanced Melanoma Treated With Nivolumab Therapy: Three-Year Follow-up of a Randomized Phase 3 Trial

Key Points

Question  What were 3-year outcomes with nivolumab vs dacarbazine in patients with previously untreated BRAF wild-type advanced melanoma?

Findings  In this follow-up of a randomized phase 3 trial, 3-year overall survival rates for nivolumab and dacarbazine were 51.2% and 21.6%, respectively, with median overall survival of 37.5 months and 11.2 months, respectively. Treatment-related grade 3/4 adverse events were reported in 15.0% (31 of 206) of nivolumab-treated patients and in 17.6% (36 of 205) of dacarbazine-treated patients.

Meaning  Nivolumab led to improved 3-year overall survival vs dacarbazine in patients with previously untreated BRAF wild-type advanced melanoma, with no new safety signals observed.

Importance  This analysis provides long-term follow-up in patients with BRAF wild-type advanced melanoma receiving first-line therapy based on anti–programmed cell death 1 receptor inhibitors.

Objective  To compare the 3-year survival with nivolumab vs that with dacarbazine in patients with previously untreated BRAF wild-type advanced melanoma.

Design, Setting, and Participants  This follow-up of a randomized phase 3 trial analyzed 3-year overall survival data from the randomized, controlled, double-blind CheckMate 066 phase 3 clinical trial. For this ongoing, multicenter academic institution trial, patients were enrolled from January 2013 through February 2014. Eligible patients were 18 years or older with confirmed unresectable previously untreated stage III or IV melanoma and an Eastern Cooperative Oncology Group performance status of 0 or 1 but without a BRAF mutation.

Interventions  Patients were treated until progression or unacceptable toxic events with nivolumab (3 mg/kg every 2 weeks plus dacarbazine-matched placebo every 3 weeks) or dacarbazine (1000 mg/m² every 3 weeks plus nivolumab-matched placebo every 2 weeks).

Main Outcome and Measure  Overall survival.

Results  At minimum follow-ups of 38.4 months among 210 participants in the nivolumab group (median age, 64 years [range, 18-86 years]; 57.6% male) and 38.5 months among 208 participants in the dacarbazine group (median age, 66 years [range, 25-87 years]; 60.1% male), 3-year overall survival rates were 51.2% (95% CI, 44.1%-57.9%) and 21.6% (95% CI, 16.1%-27.6%), respectively. The median overall survival was 37.5 months (95% CI, 25.5 months–not reached) in the nivolumab group and 11.2 months (95% CI, 9.6-13.0 months) in the dacarbazine group (hazard ratio, 0.46; 95% CI, 0.36-0.59; P < .001). Complete and partial responses, respectively, were reported for 19.0% (40 of 210) and 23.8% (50 of 210) of patients in the nivolumab group compared with 1.4% (3 of 208) and 13.0% (27 of 208) of patients in the dacarbazine group. Additional analyses were performed on outcomes with subsequent therapies. Treatment-related grade 3/4 adverse events occurred in 15.0% (31 of 206) of nivolumab-treated patients and in 17.6% (36 of 205) of dacarbazine-treated patients. There were no deaths due to study drug toxic effects.

Conclusions and Relevance  Nivolumab led to improved 3-year overall survival vs dacarbazine in patients with previously untreated BRAF wild-type advanced melanoma.

Trial Registration  ClinicalTrials.gov identifier: NCT01721772

The programmed cell death 1 (PD-1) receptor inhibitors nivolumab and pembrolizumab have demonstrated superior efficacy compared with chemotherapy or the cytotoxic T-lymphocyte–associated antigen 4 inhibitor ipilimumab in advanced melanoma, with a lower incidence of treatment-related grade 3/4 adverse events (AEs).^1-6 In phase 2 and phase 3 trials, the combination of nivolumab and ipilimumab has demonstrated significantly longer progression-free survival and a higher objective response rate compared with ipilimumab alone.^1,7-9

Emerging evidence shows encouraging long-term survival outcomes for patients with advanced melanoma who received first-line therapy based on anti–PD-1 receptor inhibitors. The randomized, controlled, double-blind CheckMate 066 clinical trial was one of the first phase 3 studies to evaluate anti–PD-1 therapy in advanced melanoma and compared nivolumab with dacarbazine in patients with previously untreated melanoma without BRAF mutation.³ The primary results were previously reported from that study, which demonstrated a significant improvement in the 1-year survival rate (73% with nivolumab vs 42% with dacarbazine), progression-free survival (5.1 months with nivolumab vs 2.2 months with dacarbazine), and objective response rate (40% with nivolumab vs 14% with dacarbazine).³ In this follow-up of a randomized phase 3 trial, we report 3-year overall survival data from the CheckMate 066 trial. This ongoing, multicenter academic institution trial enrolled patients from January 2013 through February 2014.

The CheckMate 066 trial design and patient eligibility criteria have been previously reported.³ In brief, eligible patients were 18 years or older with an Eastern Cooperative Oncology Group performance status of 0 or 1 and had histologically confirmed unresectable previously untreated stage III or IV melanoma but without a BRAF mutation.³ Patients were randomly assigned 1:1 to receive either nivolumab (3 mg/kg intravenously every 2 weeks plus dacarbazine-matched placebo intravenously every 3 weeks) or dacarbazine (1000 mg/m² intravenously every 3 weeks plus nivolumab-matched placebo intravenously every 2 weeks).³ Patients were treated until progression or unacceptable toxic effects occurred but could be treated beyond initial progression defined by the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 guideline¹⁰ if considered by a trial investigator to be experiencing clinical benefit and tolerating study drug. Patients must have discontinued therapy when further progression was documented. A protocol amendment on July 9, 2014, after unmasking of the study and based on recommendations of the data monitoring committee, allowed patients who discontinued dacarbazine to cross over to receive nivolumab in an open-label extension phase, in which they were treated until progression or unacceptable toxic effects. The study protocol was approved by the institutional review board at each participating center. The study was conducted in accord with the Declaration of Helsinki¹¹ and the International Conference on Harmonisation Guideline for Good Clinical Practice. Trial Protocol is available in Supplement 1.

Tumor response was assessed by the investigators on computed tomography or magnetic resonance imaging scan in accord with RECIST version 1.1 at the following time points: within 28 days before the first dose (baseline), 9 weeks from randomization, every 6 weeks thereafter for the first year, and then every 12 weeks until disease progression or discontinuation of treatment. The severity of AEs was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (version 4.0).¹² Exposure-adjusted evaluation of AEs over time was performed for patients receiving nivolumab, in which incidence rate per 100 person-years of exposure was reported as event count times 100 per person-years of exposure. Tumor expression of programmed cell death 1 ligand 1 (PD-L1) was assessed in pretreatment samples at a central laboratory by use of a validated, automated immunohistochemical assay (PD-L1 IHC 28-2 pharmDx; Dako) as previously described.¹³

The primary end point was overall survival. Secondary end points included investigator-assessed progression-free survival and objective response, as well as tumor PD-L1 expression as a predictive biomarker of overall survival. A post hoc analysis was conducted to assess outcomes in patients who discontinued study treatment and received subsequent therapy, including nivolumab to subsequent therapy that included any ipilimumab (nivolumab to ipilimumab), dacarbazine to subsequent therapy that included any nivolumab (dacarbazine to nivolumab), and dacarbazine to subsequent therapy that included any ipilimumab (dacarbazine to ipilimumab). Overall survival was available for these 3 patient groups because of a protocol amendment (May 6, 2015) allowing survival data to be requested outside of the protocol-defined window. Response data were also collected for patients who received any ipilimumab therapy after discontinuation of nivolumab.

Overall survival and progression-free survival were compared between the 2 treatment groups with the use of a 2-sided log-rank test stratified according to PD-L1 status and metastasis stage. The hazard ratios for the nivolumab group compared with the dacarbazine group and corresponding 95% CIs were estimated with the use of a stratified Cox proportional hazards model. Survival curves for each treatment group were estimated with the use of the Kaplan-Meier product-limit method. Rates at fixed time points were derived from the Kaplan-Meier estimate, along with their corresponding 95% CIs that were log-log transformed.

The 95% CIs for binomial proportions were derived using the Clopper-Pearson method. The differences in objective response rates between the 2 treatment groups, along with their 2-sided 95% CIs, were estimated using the Cochran-Mantel-Haenszel method. Efficacy analyses were performed in the randomized (intent-to-treat) population, whereas safety analyses were performed in all patients who received at least 1 dose of study drug. The threshold for statistical significance was based on the number of deaths (269 at the time of the analysis) using the Lan-DeMets α spending function with O’Brien-Fleming boundaries. We performed all statistical analyses with a software program (SAS, version 9.2; SAS Institute Inc).

A total of 418 patients were randomized to receive either nivolumab (n = 210) or dacarbazine (n = 208) (Figure 1); 4 patients in the nivolumab group and 3 patients in the dacarbazine group did not receive treatment. In the nivolumab group, the median age of patients was 64 years (range, 18-86 years), with 57.6% (121 of 210) male; in the dacarbazine group, the median age of patients was 66 years (range, 25-87 years), with 60.1% (125 of 208) male (Table 1). Baseline characteristics were balanced between the 2 treatment groups,³ although a higher percentage of patients in the nivolumab group had an Eastern Cooperative Oncology Group performance status of 0 (70.5% [148 of 210] vs 58.2% [121 of 208]) (Table 1). Among randomized patients, 57.6% (121 of 210) in the nivolumab group and 76.4% (159 of 208) in the dacarbazine group received subsequent treatment. In the nivolumab and dacarbazine groups, respectively, subsequent treatment included radiotherapy (25.7% [54 of 210] and 30.8% [64 of 208]), surgery (15.2% [32 of 210] and 16.3% [34 of 208]), and systemic therapy (46.2% [97 of 210] and 63.5% [132 of 208]) (eTable 1 in Supplement 2).

At a minimum follow-up of 38.4 months in the nivolumab group and 38.5 months in the dacarbazine group (database lock, June 22, 2017), the median overall survival was 37.5 months (95% CI, 25.5 months–not reached [NR]) with nivolumab and 11.2 months (95% CI, 9.6-13.0 months) with dacarbazine (hazard ratio, 0.46; 95% CI, 0.36-0.59; P < .001) (Figure 2A). Three-year overall survival rates were 51.2% (95% CI, 44.1%-57.9%) and 21.6% (95% CI, 16.1%-27.6%) in the nivolumab and dacarbazine groups, respectively.

As of the data cutoff, 63.8% (134 of 210) of patients in the nivolumab group had disease progression or had died compared with 82.7% (172 of 208) of patients in the dacarbazine group. The median progression-free survival was 5.1 months (95% CI, 3.5-12.2 months) in the nivolumab group and 2.2 months (95% CI, 2.1-2.5 months) in the dacarbazine group (hazard ratio, 0.42; 95% CI, 0.33-0.53; P < .001) (Figure 2B), with 3-year progression-free survival rates of 32.2% (95% CI, 25.6%-39.0%) and 2.9% (95% CI, 0.7%-8.1%), respectively.

In a prespecified subgroup analysis, we investigated survival outcomes by PD-L1 tumor expression. In patients with PD-L1 expression of at least 5%, the median overall survival was not reached (95% CI, 42.4-NR) in the nivolumab group and was 9.7 months (95% CI, 6.7-13.5 months) in the dacarbazine group; for those with PD-L1 expression less than 5%, the median overall survival with nivolumab was 28.2 months (95% CI, 18.2-38.5 months) and with dacarbazine was 11.6 months (95% CI, 9.3-13.0 months) (eFigure 1 in Supplement 2). Similarly, regardless of PD-L1 expression, patients in the nivolumab group had numerically longer progression-free survival compared with patients in the dacarbazine group (eFigure 1 in Supplement 2).

A significantly higher proportion of patients achieved objective responses in the nivolumab group (42.9% [90 of 210]) compared with the dacarbazine group (14.4% [30 of 208]) (Table 2 and eFigure 2 in Supplement 2). The median duration of response had not been reached in the nivolumab group (95% CI, 38.2-NR) and was 6.0 months (95% CI, 3.9-24.3 months) in the dacarbazine group. A complete response and a partial response, respectively, were reported for 19.0% (40 of 210) and 23.8% (50 of 210) of patients in the nivolumab group compared with 1.4% (3 of 208) and 13.0% (27 of 208) of patients in the dacarbazine group (Table 2). The median time to both complete response and partial response was 2.1 months with nivolumab and was 2.9 and 2.2 months, respectively, with dacarbazine. Among the 40 patients (19.0%) who achieved a complete response in the nivolumab group, 25.0% (n = 10) had lactate dehydrogenase exceeding upper limit of normal, 57.5% (n = 23) had stage M1c disease, 30.0% (n = 12) had at least 3 organs involved, and 52.5% (n = 21) had tumor PD-L1 expression of at least 5%. In the nivolumab group, 66.7% (60 of 90) of patients had ongoing responses at the time of the last assessment (Table 2 and eFigure 3 in Supplement 2). Most of the responses lasted more than 160 weeks after the start of treatment, and almost half were in patients who had discontinued nivolumab. In addition, 3 initial responses occurred after 74 weeks. In the dacarbazine group, 36.7% (11 of 30) of patients had ongoing responses at the time of the last assessment (eFigure 3 in Supplement 2).

In a post hoc analysis of overall survival for patients who discontinued study therapy and received subsequent systemic therapy, the median overall survival from randomization was 21.5 months (95% CI, 14.2-28.3 months) for the nivolumab to ipilimumab group, 35.4 months (95% CI, 22.4 months–NR) for the dacarbazine to nivolumab group, and 17.4 months (95% CI, 11.7-22.1 months) for the dacarbazine to ipilimumab group (eFigure 4 in Supplement 2). The median overall survival from the start of subsequent therapy was 8.8 months (95% CI, 7.1-14.3 months) for the nivolumab to ipilimumab group, 16.5 months (95% CI, 7.5 months–NR) for the dacarbazine to nivolumab group, and 10.6 months (95% CI 7.8-17.4 months) for the dacarbazine to ipilimumab group. Among the 68 patients who discontinued nivolumab and received subsequent therapy that included ipilimumab (92.6% [63 of 68] discontinued nivolumab because of disease progression), the objective response rate from the start of subsequent ipilimumab was 10.3% (7 of 68) (eTable 2 in Supplement 2); the median time between the last nivolumab dose and the first ipilimumab dose was 1.0 month (range, 0.0-26.9 months).

Treatment-related grade 3/4 AEs were reported in 15.0% (31 of 206) of nivolumab-treated patients and in 17.6% (36 of 205) of dacarbazine-treated patients, which led to discontinuation in 4.9% (10 of 206) and 2.0% (4 of 205) of patients, respectively (Table 3). There were no deaths due to study drug toxic effects in either group. Exposure-adjusted treatment-related AEs were investigated as incidence rate per 100 person-years (event count times 100 per person-years of exposure) in 6-month increments through 36 months. For this report, AEs that occurred in at least 3 different time points are presented (eTable 3 and eFigure 5 in Supplement 2). Most of the observed AEs occurring in at least 3 time points had a larger exposure-adjusted incidence early in treatment that decreased over time.

This analysis represents a long-term follow-up of patients with previously untreated BRAF wild-type advanced melanoma who received an anti–PD-1 agent in a phase 3 trial. With 3 years of follow-up, nivolumab treatment resulted in improved overall survival compared with dacarbazine (median, 37.5 vs 11.2 months, respectively). Previously demonstrated improvements in progression-free survival and higher objective response rates with nivolumab vs dacarbazine, regardless of tumor PD-L1 expression, were maintained at the 3-year follow-up. Collectively, our results showed durable responses and long-term survival with nivolumab monotherapy, with no new AEs developing at late time points.

The median overall survival had not been reached in the nivolumab group at the time of the initial report (up to 16.7 months of follow-up)³ but was 37.5 months at the time of the present analysis, with 51.2% (95% CI, 44.1%-57.9%) of patients alive. In the phase 1b KEYNOTE-001 study,¹⁴ the 3-year overall survival rate with pembrolizumab monotherapy was 45% in treatment-naive patients and 40% in the overall population. In the phase 3 KEYNOTE-006 study,^15,16 which also enrolled previously untreated and previously treated patients, pembrolizumab-treated patients had survival rates of 55% at 2 years and 50% at 33 months (median overall survival, 32.3 months [with a 33.9-month median follow-up]). In contrast to the present study, patients with and without a BRAF mutation were included in KEYNOTE-001 and KEYNOTE-006. The CheckMate 067 trial⁸ also included patients with BRAF wild-type and BRAF-mutant melanoma, with a 3-year median overall survival of 37.6 months and a survival rate of 52% observed with nivolumab monotherapy. These findings demonstrate consistency in survival outcomes with anti–PD-1 monotherapy across studies, regardless of BRAF mutation status. Nivolumab plus ipilimumab from CheckMate 067 was associated with a 3-year overall survival rate of 58%, with the median not yet reached.⁸

Compared with the initial analysis,³ our 3-year results demonstrate an increase in the objective response rate with nivolumab (40.0% [84 of 210] vs 42.9% [90 of 210]). In addition, compared with previous analyses,^3,17 our results show an increase in the number of patients who achieved a complete response (16 [7.6%] at 1 year, 22 [11.0%] at 2 years, and 40 [19.0%] at 3 years). Time to response was generally similar between nivolumab and dacarbazine, but the median duration of response with nivolumab was longer compared with dacarbazine (NR vs 6 months, respectively). In addition, responses to nivolumab were long-lasting in many patients who discontinued treatment, with most patients who stopped treatment still alive and without disease progression at the time of the last assessment. Patients who achieved a complete response in the nivolumab group had, in general, good prognostic factors, although 25.0% (10 of 40) had elevated lactate dehydrogenase level, 57.5% (23 of 40) had stage M1c disease, and 30.0% (12 of 40) had at least 3 organ sites involved. However, achievement of an objective response with an immune checkpoint inhibitor is not necessarily required for long-term survival.¹⁸ Indeed, our results showed that the percentage of patients surviving to 3 years was higher than the percentage who achieved an objective response.

One limitation across studies is the use of subsequent therapies, which likely affected survival outcomes in the present analysis. The median overall survival from randomization was longer and 3-year survival rates were higher for nivolumab compared with dacarbazine when either was followed by subsequent therapy that included ipilimumab. These results indicate that patients can benefit from subsequent ipilimumab, regardless of prior therapy. Survival outcomes may also have been affected by the allowance of nivolumab treatment beyond progression in some patients, as suggested by the findings of prior analysis.¹⁹

The results of this 3-year follow-up analysis provided evidence for a durable survival benefit with nivolumab monotherapy in patients with previously untreated BRAF wild-type advanced melanoma. Improved rates of complete response and longer progression-free survival, as well as overall survival, demonstrated durable benefit with nivolumab monotherapy beyond 1 year. Novel combinations under evaluation using anti–PD-1 therapies as the backbone have the potential to further improve outcomes in patients with advanced melanoma.

Accepted for Publication: July 17, 2018.

Published Online: October 25, 2018. doi:10.1001/jamaoncol.2018.4514

Correction: This article was corrected on December 6, 2018, to fix an instance of the incorrect treatment group in the Results and data errors in the legend for Figure 2.

Corresponding Author: Paolo A. Ascierto, MD, Istituto Nazionale Tumori Fondazione Pascale, Via Mariano Semmola, 80131 Naples, Italy (p.ascierto@istitutotumori.na.it).

Author Contributions: Dr Ascierto had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Ascierto, Long, Robert, Saci.

Acquisition, analysis, or interpretation of data: Ascierto, Long, Robert, Brady, Dutriaux, Di Giacomo, Mortier, Hassel, Rutkowski, McNeil, Kalinka-Warzocha, Savage, Hernberg, Lebbé, Charles, Mihalcioiu, Chiarion-Sileni, Mauch, Cognetti, Ny, Arance, Svane, Schadendorf, Gogas, Jiang, Rizzo, Atkinson.

Drafting of the manuscript: Ascierto, Long, Di Giacomo, Mihalcioiu, Cognetti, Gogas, Atkinson.

Critical revision of the manuscript for important intellectual content: Ascierto, Long, Robert, Brady, Dutriaux, Di Giacomo, Mortier, Hassel, Rutkowski, McNeil, Kalinka-Warzocha, Savage, Hernberg, Lebbé, Charles, Mihalcioiu, Chiarion-Sileni, Mauch, Ny, Arance, Svane, Schadendorf, Gogas, Saci, Jiang, Rizzo, Atkinson.

Statistical analysis: Long, Chiarion-Sileni, Cognetti, Jiang.

Obtained funding: Kalinka-Warzocha.

Administrative, technical, or material support: Robert, Charles, Mauch, Arance, Schadendorf, Gogas, Saci.

Supervision: Ascierto, Brady, Di Giacomo, Mortier, Rutkowski, Lebbé, Charles, Arance, Gogas, Atkinson.

Conflict of Interest Disclosures: Dr Ascierto reported having a consultant/advisory role with Bristol-Myers Squibb, Roche-Genentech, Merck Sharp & Dohme, Novartis, Amgen, Array, Merck, Pierre-Fabre, Incyte, Genmab, Newlink Genetics, MedImmune, Syndax, and AstraZeneca and reported receiving institutional research funding from Bristol-Myers Squibb, Roche-Genentech, and Array. Dr Long reported receiving honoraria from Merck MSD, Roche, and Bristol-Myers Squibb and reported having a paid consulting role with Merck MSD, Roche, Bristol-Myers Squibb, Amgen, Novartis, and Pierre-Fabre. Dr Robert reported having a paid consulting role with Bristol-Myers Squibb, Merck, Roche, and Amgen and reported receiving honoraria from Bristol-Myers Squibb, Merck, GlaxoSmithKline, Roche, Novartis, and Amgen. Dr Brady reported being paid to participate in speakers’ bureaus for Bristol-Myers Squibb and Merck; reported having a paid consulting role with Merck and Novartis; and reported having travel or other expenses paid or reimbursed by Bristol-Myers Squibb. Dr Di Giacomo reported receiving honoraria from Bristol-Myers Squibb, Roche, and Merck Sharp & Dohme and reported being a consultant for Incyte, Pierre-Fabre, and GlaxoSmithKline. Dr Mortier reported receiving clinical trials funding from Bristol-Myers Squibb, GlaxoSmithKline, Merck Sharp & Dohme, and Roche and reported participating on an advisory board for Bristol-Myers Squibb. Dr Hassel reported having a paid consulting role with Merck and Amgen and reported receiving honoraria from Bristol-Myers Squibb, Merck, Novartis, Roche, and Pfizer. Dr Rutkowski reported receiving honoraria from Bristol-Myers Squibb, Roche, Novartis, Merck Sharp & Dohme, and GlaxoSmithKline; reported having a paid consulting role with Bristol-Myers Squibb, Roche, Merck Sharp & Dohme, and Amgen; reported being paid to participate in speakers’ bureaus for Novartis and Merck Sharp & Dohme; reported receiving institutional research funding from Bristol-Myers Squibb; and reported having travel or other expenses paid or reimbursed by Novartis. Dr McNeil and members of her department reported having travel or other expenses paid or reimbursed by Merck Sharp & Dohme and Bristol-Myers Squibb; reported participating on advisory boards for Bristol-Myers Squibb and Merck Sharp & Dohme; and reported receiving research funding from Merck Sharp & Dohme for her institution. Dr Kalinka-Warzocha reported receiving payments for trial procedures according to study protocol from Bristol-Myers Squibb; reported receiving honoraria from Bristol-Myers Squibb, Roche, Novartis, Merck Sharp & Dohme, and AstraZeneca; and reported having a paid consultant role with Bristol-Myers Squibb, Merck Sharp & Dohme, and AstraZeneca. Dr Savage reported receiving honoraria from Seattle Genetics, Bristol-Myers Squibb, Celgene, and Takeda; reported having a paid consulting role with Seattle Genetics, Merck, and Bristol-Myers Squibb; reported being paid to participate in a speakers’ bureau for Seattle Genetics; and reported receiving research funding from Roche. Dr Lebbé reported participating on advisory boards for Bristol-Myers Squibb, Merck Sharp & Dohme, Roche, GlaxoSmithKline, and Novartis and reported receiving travel expenses from Bristol-Myers Squibb and Merck Sharp & Dohme. Dr Mihalcioiu reporting having a paid consulting role with Bristol-Myers Squibb, Novartis, and Merck and reported having an issued US patent on circulatory cells. Dr Chiarion-Sileni reported having a paid consulting role with Bristol-Myers Squibb, Roche, GlaxoSmithKline, and Merck Sharp & Dohme; reported being paid to participate in a speakers’ bureau for Bristol-Myers Squibb, Roche, and GlaxoSmithKline; and reported having travel or other expenses paid or reimbursed by Bristol-Myers Squibb, Roche, GlaxoSmithKline, and Merck Sharp & Dohme. Dr Cognetti reported participating on an advisory board for Astellas Oncology. Dr Ny reported receiving clinical institutional funding from Merck Sharpe & Dohme and Syndax; reported receiving payment for participating in advisory boards with Bristol-Myers Squibb and Novartis; and reported having a consulting role with Bristol-Myers Squibb and AstraZeneca. Dr Arance reported having a paid consulting role with GlaxoSmithKline and Roche; reported being paid to participate in speakers’ bureaus for GlaxoSmithKline, Roche, and Bristol-Myers Squibb; and reported having travel or other expenses paid or reimbursed by Bristol-Myers Squibb. Dr Svane reported receiving honoraria from Bristol-Myers Squibb, Merck, and Roche; reported receiving clinical research funding from Bristol-Myers Squibb, Novartis, and Roche; and reported receiving payment for participation in scientific advisory boards from Bristol-Myers Squibb, GlaxoSmithKline, and Roche. Dr Schadendorf reported receiving honoraria from Merck Sharp & Dohme, Roche, Amgen, Novartis, and Bristol-Myers Squibb; reported having paid consulting roles and being a member of speakers’ bureaus with Merck Sharp & Dohme, Roche, Amgen, Novartis, Bristol-Myers Squibb, Pierre-Fabre, and Merck-Serono; reported having a paid consultant role with Incyte; and reported receiving research funding from Merck Sharp & Dohme and Bristol-Myers Squibb. Dr Gogas reported receiving personal fees from GlaxoSmithKline, Roche, Bristol-Myers Squibb, Merck Sharp & Dohme, and Amgen. Drs Saci, Jiang, and Rizzo reported being employed by Bristol-Myers Squibb. Dr Atkinson reported receiving honoraria from Bristol-Myers Squibb, Merck Sharp & Dohme, and Novartis; reported having a paid consulting role with Bristol-Myers Squibb, Merck Sharp & Dohme, and Novartis; reported being paid to participate in speakers’ bureaus for Bristol-Myers Squibb, Merck Sharp & Dohme, and Novartis; and reported having travel or other expenses paid or reimbursed by Bristol-Myers Squibb, Merck Sharp & Dohme, and Novartis. No other disclosures were reported.

Funding/Support: Financial support for this study was provided by Bristol-Myers Squibb.

Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 3.

Additional Contributions: Medical writing and editorial support were provided by Melissa Kirk, PhD, Jennifer Dinieri, PhD, and Cara Hunsberger, MS, of StemScientific (an Ashfield Company, Lyndhurst, New Jersey) and were funded by Bristol-Myers Squibb. We thank the patients who participated in this study, the clinical study teams, and representatives of the sponsor who were involved in data collection and analyses. We thank Dako for collaborative development of the automated immunohistochemical assay (PD-L1 IHC 28-2 pharmDx).