Nivolumab Plus Ipilimumab vs Nivolumab for Previously Treated Patients With Stage IV Squamous Cell Lung Cancer

This phase 3 randomized clinical trial investigates if the addition of ipilimumab to nivolumab improves the survival outcome of patients with chemotherapy-pretreated immunotherapy-naive squamous cell lung cancer.

P rogrammed death 1 (PD-1) axis inhibitor monotherapies, including nivolumab, are currently standard salvage therapy for patients with immunotherapy-naive advanced non-small cell lung cancer (NSCLC) that has progressed on platinum-doublet chemotherapy. [1][2][3][4] The PD-1 axis inhibitors pembrolizumab and atezolizumab have additionally been approved for use as first-line monotherapy in patients with advanced NSCLC expressing programmed deathligand 1 (PD-L1) or in combination with chemotherapy regardless of tumor PD-L1 expression. [5][6][7][8][9][10][11] Ipilimumab is an immune checkpoint inhibitor targeting cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) that is approved for use as monotherapy in metastatic melanoma. The combination of ipilimumab with nivolumab has demonstrated superior efficacy compared with nivolumab alone in patients with advanced melanoma and is additionally approved in this setting. 12,13 Phase 1 and 2 studies in patients with untreated advanced NSCLC showed promising early results with nivolumab plus ipilimumab, and 2 recent phase 3 trials in this population demonstrated superiority of the combination either alone or with chemotherapy compared with chemotherapy alone. [14][15][16][17] Nivolumab plus ipilimumab is currently approved in the first-line setting for patients with metastatic PD-L1-expressing NSCLC and in combination with 2 cycles of platinum-doublet chemotherapy regardless of tumor PD-L1 expression.
The Lung Cancer Master Protocol (Lung-MAP) was designed as a biomarker-driven protocol to evaluate molecularly targeted therapies in biomarker-defined populations of patients with chemotherapy-pretreated advanced squamous (Sq) NSCLC. The initial iteration of the Lung-MAP protocol was designated S1400. 18 To enrich the probability of a patient participating, Lung-MAP includes both biomarker-driven studies and nonmatch studies for patients not eligible to participate in a biomarker-driven study. The second nonmatch trial within Lung-MAP (S1400), S1400I, evaluated the efficacy of nivolumab plus ipilimumab vs nivolumab alone in patients with immune checkpoint inhibitor-naive SqNSCLC. At the time this study was designed, nivolumab was approved as salvage therapy but not as initial treatment in patients with advanced NSCLC.

Study Design
This multicenter, open-label, phase 3 randomized clinical trial and substudy of Lung-MAP (S1400) was conducted from December 18, 2015, to April 23, 2018, through the National Clinical Trials Network and led by the SWOG Cancer Research Network. The trial protocol is available in Supplement 1, and the S1400 master protocol is available in Supplement 2. The trial compared nivolumab plus ipilimumab with nivolumab monotherapy in patients with chemotherapy-pretreated, immunotherapy-naive advanced SqNSCLC. To participate, each site required approval by the US National Cancer Institute central institutional review board or approval by their local institutional review board. Written informed consent was required from all patients before registration. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.

Eligibility
Patients must have been eligible for Lung-MAP (S1400) and not eligible for any of the actively accruing biomarker-driven substudies. 18,19 Biomarker/drug pairs evaluated during this study have been described. [19][20][21][22][23][24] Patients had histologically confirmed SqNSCLC with measurable disease by computed tomography or magnetic resonance imaging per Response Evaluation Criteria in Solid Tumors (RECIST) guidelines, version 1.1; were previously treated with platinum-doublet chemotherapy (within 1 year if given as part of curative-intent therapy without subsequent systemic therapy); and had a Zubrod performance status score of 0 (asymptomatic) to 1 (symptomatic but completely ambulatory). Patients must not have had prior treatment with an antibody or other drug specifically targeting T-cell costimulation or immune checkpoint pathways, including PD-1 and CTLA-4. Systemic treatment with either corticosteroids or other immunosuppressive medications within 14 days before substudy registration was not allowed. Patients could not have an active, known, or suspected autoimmune disease, HIV, hepatitis B, hepatitis C, or interstitial lung disease. Patients registered after September 1, 2016, who were able to complete a questionnaire in English participated in a patient-reported outcomes study.

Biomarker Analysis
All patients in the Lung-MAP (S1400) study had biomarker screening by next-generation sequencing (Foundation Medicine) as previously described. 19 Tumor mutational burden (TMB) was calculated as the number of somatic, coding, short variants, excluding known driver mutations, per megabase of the genome interrogated. Immunohistochemistry for PD-L1 staining and scoring was performed by the Clinical Laboratory Improvement Amendments-certified Biomarker Analysis Laboratory at the University of Colorado using the PD-L1 28-8 pharmDx kit (Dako) on the Dako Link 48 platform. The PD-L1 analysis was done after entry into the study and not used for stratification in randomization. For all stained specimens, 1 pathologist (H.Y.) scored all of the specimens, and a quality control pathologist scored 20% of the specimens. For discrepant results, a final score was determined by a consensus con-

Key Points
Question Does the addition of ipilimumab to nivolumab improve survival in patients with advanced chemotherapy-pretreated immunotherapy-naive squamous cell lung cancer?
Findings In this randomized clinical trial of 252 patients, the addition of ipilimumab to nivolumab did not lead to improved survival in patients with advanced chemotherapy-pretreated squamous cell carcinoma.
Meaning Combination therapy with nivolumab and ipilimumab is currently only indicated as first-line therapy in patients with advanced non-small cell lung cancer.
ference of the pathologists. Scoring was determined by the percentage of tumor cells with partial or complete cell membrane staining at any intensity.

Randomization and Treatment
Patients were randomized in a 1:1 ratio to receive either nivolumab combined with ipilimumab or nivolumab alone. Randomization was stratified by sex (men vs women) and number of prior therapies (1 vs 2 or more) using a dynamic balancing algorithm. 25 In both groups, nivolumab was administered intravenously at a dose of 3 mg/kg over 30 minutes on day 1 of 14day cycles. For patients randomized to the nivolumab plus ipilimumab group, ipilimumab was administered intravenously at a dose of 1 mg/kg over 60 minutes on day 1 of every third cycle (ie, cycle 1, cycle 4, etc) starting 30 minutes after the end of nivolumab infusion. Disease assessments occurred every 6 weeks for the first year and every 3 months thereafter. Treatment was continued until disease progression or intolerable toxic effects. Postprogression therapy was allowed if the patient was experiencing clinical benefit in the opinion of the treating investigator.

End Points
The primary end point for the trial was overall survival (OS), which was defined as the duration from randomization to death due to any cause. The OS for patients last known to be alive was censored at the date of last contact. Secondary end points included investigator-assessed progression-free survival (IA-PFS), response by RECIST 1.1, and toxic effects by the US National Cancer Institute's new Patient-Reported Outcome Common Terminology Criteria for Adverse Events (PRO-CTCAE), version 4.03. The IA-PFS was defined as the duration from randomization to first occurrence of progression by RECIST 1.1, symptomatic deterioration, or death due to any cause. The IA-PFS for patients last known to be alive and free of progression or symptomatic deterioration was censored at the date of last disease assessment. Response was defined as the occurrence of a complete or partial response, confirmed or unconfirmed per RECIST 1.1 criteria. Response for patients not known to have a response was coded as nonresponse. Patient-reported symptoms were evaluated using the PRO-CTCAE items for diarrhea frequency, itching severity, fatigue severity, and interference with daily activities due to fatigue. 26

Statistical Analysis
The S1400I iteration of Lung-MAP was designed to have 90% power to rule out a hazard ratio (HR) equal to 1 if the true HR was 0.67 using a stratified log-rank test at the 1-sided 0.025 level. Assuming an accrual duration of 27 to 36 months, 9 months of follow-up, and a median OS of 9 months with nivolumab alone, 332 eligible patients were needed to achieve the 256 events required by design. Two interim analyses were planned at 50% and 75% of the expected deaths to evaluate early stopping for either efficacy or futility. Assessment of futility was based on testing the alternative hypothesis (HR = 0.67) using a modified stratified log-rank test for testing the non-null hypothesis at the 1-sided 0.0025 level.
Assessment of early signs of efficacy was based on testing the null hypothesis using a stratified log-rank test at the 1-sided 0.0025 level.
Efficacy analyses (OS, IA-PFS, and response) were conducted on all randomized patients determined to be eligible (modified intention to treat). As is standard for National Clinical Trials Network trials, eligibility was confirmed after randomization. Safety analyses were conducted on all eligible patients who received at least 1 dose of study drug. Stratified logrank tests using randomization stratification factors were used to analyze time-to-event data. Distribution of time-to-event outcomes was estimated using the Kaplan-Meier method, and the CI for medians was estimated using the Brookmeyer-Crowley method. Hazard ratios and associated 95% CIs were estimated with the use of a stratified Cox proportional hazards regression model. Descriptive statistics were used to summarize the characteristics of the patients. Analysis of the association between baseline factors, including TMB score and PD-L1 expression levels, with clinical outcomes was also done using a Cox proportional hazards regression model for OS and IA-PFS and logistic regression for response. The proportion of patients by arm experiencing a maximum postrandomization PRO-CTCAE score greater than 0 or greater than 3, both unadjusted and adjusted for the baseline PRO-CTCAE score, was compared using Fisher exact tests among patients with both baseline and follow-up scores. 27 Prespecified biomarker analyses separately evaluated OS in the subset of patients with PD-L1 levels greater than or equal to 5% and for TMB levels greater than or equal to 10 mutations (mt)/Mb using a stratified log-rank test. Post hoc exploratory analysis of PD-L1 evaluated subgroup effects for 0%, 1% to 4%, 5% to 49%, 50% to 74%, and 75% and greater tumor PD-L1 expression to expand beyond the prespecified 5% cut point and for TMB levels by decile to evaluate whether there was evidence of a cut point for activity beyond the 10 mt/Mb cut point. Subsequent analyses evaluated subgroup effects of combined PD-L1 and TMB levels by using a Cox proportional hazard regression model including an interaction between continuous TMB score and treatment within the exploratory PD-L1 subgroups mentioned previously. If a signal of activity was observed (P < .10), evaluation of subgroup effects by TMB scores between the 20th and 70th percentiles would be pursued (to protect against small numbers, we did not use the full decile range) to evaluate whether the association with increasing TMB levels was increasing or whether there was a threshold above which activity was isolated. All statistical analyses were performed using SAS software, version 9.4 (SAS Institute Inc) and RStudio, version 1.4.1106 (R Core Team); data were analyzed from May 3, 2018, to February 1, 2021. ). One hun-dred twenty-five patients were randomized to the nivolumab plus ipilimumab group and 127 to the nivolumab group. Details of patient enrollment, eligibility, and treatment are provided in Figure 1, and baseline characteristics are described in Table 1. The study was closed to accrual by the Data and Safety Monitoring Committee on April 23, 2018, after the first interim analysis demonstrated futility. The frequency of detection of individual alterations from the next-generation sequencing screening is detailed in eTable 1 in Supplement 3.

Patient Characteristics and Treatment
A total of 124 patients in the nivolumab plus ipilimumab group and 123 patients in the nivolumab group received at least 1 dose of protocol treatment, with a median of 7 doses in both groups (ranges, 1-84 doses for nivolumab plus ipilimumab and 1-83 doses for nivolumab alone). Trial therapy was continued after progression in 38 of 116 patients (33%) in the nivolumab group and 37 of 104 patients (36%) in the nivolumab plus ipilimumab group. As of December 19, 2019, 9 patients were still receiving trial therapy (6 in the nivolumab plus ipilimumab group and 3 in the nivolumab group). A total of 50 patients (20%) discontinued due to toxic effects (31 of 124 [25%] in the nivolumab plus ipilimumab group and 19 of 123 [15%] in the nivolumab group). Median time to discontinuation of therapy due to toxic effects was 5.6 months (95% CI, 3.3-10.4 months) with nivolumab plus ipilimumab and 4.0 months (95% CI, 2.0-6.4 months) with nivolumab alone.
The TMB levels and PD-L1 score were not correlated (correlation coefficient = 0.09; P = .29) (eFigure 2 in Supplement 3). Treatment was not associated with OS in either of the prespecified thresholds (TMB≥10 mt/Mb and PD-L1 expression ≥5%) or for the exploratory thresholds for PD-L1 expression or TMB (P > .05 for all). Higher TMB levels, evaluated as a continuous measure, were associated with improved survival in the subset of patients with PD-L1 expression less than 1% (56 patients and 47 events; P = .06) (eTable 4 in Supplement 3), and the benefit was likely best isolated to the subset  with TMB greater than or equal to 10 mt/Mb (eFigure 3 in Supplement 3). Subsequent analysis showed improved survival with nivolumab plus ipilimumab over nivolumab alone in patients with TMB greater than or equal to 10 mt/Mb and PD-L1 expression less than 1% (n = 15 for both arms; HR, 0.37; 95% CI, 0.15-0.89; P = .02) but survival detriment in patients with TMB less than 10 mt/Mb and PD-L1 expression less than 1% (n = 17 for nivolumab alone and n = 9 for nivolumab plus ipilimumab; HR, 2.46; 95% CI, 0.95-6.34; P = .06) (Figure 3). Of note, the exploratory nature of these combined analyses and the small number of patients in each subset limit conclusions. Tumor response in the 2 subsets is provided in eTable 5 in Supplement 3.

Discussion
Although nivolumab combined with ipilimumab has demonstrated superior efficacy to nivolumab alone in advanced melanoma and to first-line platinum-based chemotherapy in advanced NSCLC, the combination did not improve OS over nivolumab alone in patients with immunotherapynaïve, pretreated advanced SqNSCLC. Prior trials have suggested that high TMB may predict greater benefit with nivolumab plus ipilimumab in advanced NSCLC, although optimal cutoff for TMB differs across studies. In our study, high TMB defined as a cutoff of 10 mt/Mb did not lead to a better outcome with combination therapy. Additional biomarker analyses from this trial, including tumor PD-L1 expression and tumor sequencing, did not identify individual tumor characteristics predictive of benefit from either treatment over the other. However, in patients with no tumor PD-L1 expression, high TMB predicted a benefit with nivolumab plus ipilimumab, whereas low TMB predicted inferior survival compared with nivolumab alone. Consistent with previous studies, TMB did not correlate with tumor PD-L1 status. Subsequent to this trial, treatment paradigms for advanced NSCLC changed such that most patients will receive immunotherapy as front-line therapy. However, although the question posed by this trial is no longer as relevant, there is still a need to understand potential differences among various PD-1 axis inhibitors and whether firstline combination therapy with an anti-CTLA-4 agent will improve outcomes over PD-1 axis inhibitor therapy, either alone or with concurrent chemotherapy. In CheckMate 227, 16 first-line therapy with nivolumab plus ipilimumab in  patients with PD-L1-expressing NSCLC resulted in a numerically higher response rate and median duration of response (36.4% and 23.2 months, respectively) than with nivolumab monotherapy (27.5% and 15.5 months, respectively).
Although not a primary end point of that trial, risk of progression was decreased with nivolumab plus ipilimumab compared with nivolumab alone (HR, 0.83; 95% CI, 0.71-0.97). Difference in OS did not reach statistical significance (HR, 0.90; 95% CI, 0.76-1.07). Among patients with tumor PD-L1 expression less than 1%, nivolumab plus ipilimumab resulted in longer OS and higher 2-year PFS rates (20% and 6%, respectively) than did chemotherapy, highlighting the need for other biomarkers in this population. Interestingly, exploratory analysis in the subgroup with PD-L1 less than 1% found that the PFS benefit with nivolumab plus ipilimumab was limited to patients with high TMB.

Limitations
As discussed, the main limitation of our study is that the control group of nivolumab is no longer a commonly used salvage therapy for patients with advanced lung cancer; most patients will receive PD-1 axis inhibitor therapy in the first-line setting. Another limitation is that adequate tissue for determining tumor PD-L1 status was not required. Indeed, only 64% of patients had successful tumor PD-L1 analysis; among these patients, there did not appear to be major differences in PD-L1 expression between arms. Reduced numbers of patients with available results from both tumor PD-L1 and mutational burden assessment do limit conclusions from the combined subset analyses that evaluated the outcome presented above.

Conclusions
In this multicenter, open-label, phase 3 randomized clinical trial and substudy of Lung-MAP (S1400), ipilimumab added to nivolumab did not improve outcomes in patients with advanced, pretreated, immune checkpoint inhibitor-naive SqNSCLC. At present, there is no immunotherapy option for patients who experience disease progression on PD-1 axis inhibitor therapy. At least 2 ongoing trials are evaluating  nivolumab plus ipilimumab as salvage therapy in patients with progressive advanced NSCLC after PD-1 axis inhibitor therapy. 28, 29 These studies enroll both patients with acquired and primary resistance to PD-1 axis inhibitor therapy. Until these studies and others are completed, there will continue to be no role for combination therapy with nivolumab and ipilimumab in patients with pretreated advanced NSCLC.   Funding/Support: This study was supported by grants from the Foundation for the National Institutes of Health to SWOG.

Role of the Funder/Sponsor:
The funder had no role in the conduct of the study; collection, management, analysis, and interpretation of the data; preparation or approval of the manuscript; and decision to submit the manuscript for publication.
Data Sharing Statement: See Supplement 4.