Comparing First-Line Immunotherapy in Advanced Melanoma: No Clear Winner?

In this Perspective on Science piece, we reflect on the indirect treatment comparison analysis between RELATIVITY-047 and CheckMate 067 trial data.

Authors

Affiliations

Saito, Yoshine

National Institutes of Health

Miller, David M.

Massachusetts General Hospital

Harvard Medical School

Brownell, Isaac

National Institutes of Health

Keywords

Melanoma, Indirect Treatment Comparison

Metrics

PlumX

Comparing First-Line Immunotherapy in Advanced Melanoma: No Clear Winner?

Featured Article

Long, G. V. et al. First-Line Nivolumab Plus Relatlimab Versus Nivolumab Plus Ipilimumab in Advanced Melanoma: An Indirect Treatment Comparison Using RELATIVITY-047 and CheckMate 067 Trial Data. Journal of Clinical Oncology (2024) doi:10.1200/jco.24.01125.1

Introduction

On November 8^th, 2024, the multi-institutional Society of Cutaneous Oncology (SoCO) Journal Club (JC) reviewed the Journal of Clinical Oncology article titled “First-Line Nivolumab Plus Relatlimab Versus Nivolumab Plus Ipilimumab in Advanced Melanoma: An Indirect Treatment Comparison Using RELATIVITY-047 and CheckMate 067 Trial Data.” Participants in this session included non-clinician researchers, student trainees, and predominantly clinicians specializing in surgical oncology, medical dermatology, or medical oncology (Figure 1). Attendees represented institutions such as Massachusetts General Hospital, Mass Eye and Ear Infirmary, National Institutes of Health, and others. The participants’ relative experience in managing melanoma and metastatic melanoma varied (Figure 2).

Figure 1 - November 8^th 2024 SoCO journal club attendees

Figure 1: Attendees at the November 8^th SoCO Journal Club were surveyed as to their professional role. The percentage of the total respondents is shown to the right of each bar.

Figure 2 - Melanoma Experience

A. Melanoma Experience

B. Advanced Melanoma Experience

Figure 2: Attendees were surveyed about their experience with melanoma overall (A) and advanced melanoma (B), where systemic therapy may be indicated.The percentage of total respondents is displayed to the right of each bar.

In this perspective piece, we summarize the featured JC article and highlight key discussion points. Perspective on the Science reflects its authors’ views after the Journal Club and does not represent the views of other SoCO members or affiliated institutions.

Background

The use of immune checkpoint inhibitors (ICI) for the treatment of unresectable or metastatic melanoma has dramatically improved patient outcomes^2–8.In addition to single-agent therapies, multiple dual-regimen ICIs are now approved for melanoma (Table 1). The CheckMate 067 trial evaluated nivolumab plus iplimumab (Nivo/Ipi) – a dual PD-1 and CTLA-4 inhibitor regimen – against nivolumab alone in patients with advanced melanoma, leading to its subsequent FDA approval⁹. In the CheckMate 067 trial, patients were randomly assigned (1:1:1) to receive nivolumab 1 mg/kg once every 3 weeks plus ipilimumab 3 mg/kg once every 3 weeks for four doses followed by nivolumab 3 mg/kg once every 2 weeks, nivolumab 3 mg/kg once every 2 weeks, or ipilimumab 3 mg/kg once every 3 weeks for four doses.In similar fashion, RELATIVITY-047 trial demonstrated that nivolumab plus relatlimab (Nivo/Rela) produced favorable results compared to nivolumab alone in patients with advanced melanoma, resulting in its FDA approval as well⁶. RELATIVITY-047, patients were randomly assigned (1:1) to receive either nivolumab 480 mg and relatlimab 160 mg as a fixed-dose combination once every 4 weeks or nivolumab 480 mg once every 4 weeks. Nivo/Rela is currently the only dual PD-1 and LAG-3 inhibitor regimen approved for the treatment of advanced melanoma. Both dual-regimen therapies are now recommended as first-line systemic treatment options for metastatic or unresectable melanoma in the NCCN guidelines¹⁰.

Table 1 - Immune Checkpoint Inhibitors With FDA Approved Indications for Melanoma

Therapy	Study	Target	Year of Approval	Reference
Ipilimumab	MDX010-20	CTLA4	2011	Hodi et al. 2010
Pembrolizumab	KEYNOTE-001	PD-1	2014 (accelerated)	Robert et al. 2014
Nivolumab	CheckMate 037	PD-1	2014 (accelerated)	Weber et al. 2015
Nivolumab and ipilimumab	CheckMate 067	PD-1/CTLA4	2015 (accelerated)	Larkin et al. 2015
Atezolizumab in combination with Vemurafenib and Cobimetinib	IMspire150	PD-L1 (BRAF/MEK)	2020	Gutzmer et al. 2020
Nivolumab and relatlimab	RELATIVITY-047	PD-1/LAG3	2022 (accelerated)	Tawbi et al. 2022

Since no randomized clinical trials have directly compared these two dual regimens as first-line therapy for advanced melanoma, the authors of this paper instead performed an indirect treatment comparison (ITCs) using patient-level data from the two independent registration trials to investigate the relative efficacy and safety profiles of the two regimens.

Study Design

In this study, Long et al. analyzed patient-level data from RELATIVITY-047 and CheckMate 067. To increase comparability between the trials, the authors adjusted the follow-up time for CheckMate067 and applied a propensity score model with inverse probability of treatment weighting (IPTW) to balance baseline patient characteristics. The efficacy outcomes they analyzed included progression-free survival (PFS), overall survival (OS), confirmed objective response rate (cORR), duration of response (DOR), and melanoma-specific survival (MSS). Safety outcomes focused on treatment-related adverse events (TRAEs) that occurred within 30 days of the last dose of ICI over a 36-month follow-up period. Additionally, the authors conducted sensitivity analyses to examine the untruncated CheckMate067 data and performed internal validation analyses by comparing outcomes for the nivolumab monotherapy arms from both trials using the same methodology described above.

Main Findings

After balancing all covariates using IPTW, the effective sample sizes were 339 for the Nivo/Rela group and 297 for the Nivo/Ipi group. The authors observed no significant differences between the two regimens in PFS (36-month PFS, 36% v 39%; HR, 1.08 [95% CI, 0.88 to 1.33]), cORR (48% v 50%; OR, 0.91 [95% CI, 0.73 to 1.14]), median DOR (HR, 0.88 [95% CI, 0.60 to 1.28]), OS (36-month OS, 57% for both treatments; HR, 0.94 [95% CI, 0.75 to 1.19]), and MSS (36-month MSS, 65% v 62%; HR, 0.86 [95% CI, 0.67 to 1.12]).

In terms of safety, the authors found that the Nivo/Rela group experienced fewer grade 3 and 4 TRAEs (23% v 61%) as well as fewer any-grade TRAEs leading to treatment discontinuation (17% v 41%). When categorized by organ system, the Nivo/Rela group had lower rates of any-grade TRAEs affecting the endocrine, GI, hepatic, renal, and skin systems while less frequent grade 3 and 4 TRAEs were seen in endocrine, GI, hepatic, and skin systems.

Sensitivity analyses using the untruncated CheckMate 067 data confirmed that PFS, OS, and MSS remained similar between Nivo/Rela and Nivo/Ipi, consistent with the primary analysis. Finally, internal validation analyses showed similar PFS, cORR, OS, and MSS between the two nivolumab monotherapy groups after weighting.

Discussion

Advances in melanoma management have led to the development of multiple effective immunotherapy regimens, significantly improving patient outcomes. However, the absence of direct, head-to-head trials comparing these regimens presents a challenge for clinicians in selecting the most appropriate therapy. To address this gap, the authors conducted an indirect treatment comparison using propensity score weighting to evaluate the efficacy and safety of Nivo/Rela versus Nivo/Ipi in advanced melanoma. Their analysis did not detect significant differences in efficacy between the two regimens across key outcomes, including PFS, OS, and ORR. However, Nivo/Rela was associated with a lower incidence of high-grade treatment-related adverse events, suggesting a potential advantage in tolerability. Despite this, the consensus among the journal club participants was that the study was not particularly practice-changing (Figure 3). Of the 9 respondents who actively manage melanoma patients, four indicated their current approach was already aligned with the findings, while three felt the paper was not compelling enough to alter their practice.

Figure 3 - Practice Changing Paper?

Figure 3: Journal club attendees were queried as to whether or not this paper would change their practice.

During the discussion, both Nivo/Rela and Nivo/Ipi were considered reasonable first-line treatment options for advanced melanoma, with the study not providing strong enough evidence to favor one regimen definitively. Treatment selection was viewed as a multifaceted decision influenced by factors such as disease progression rate, BRAF mutation status, presence of brain metastases, patient preference, and potential toxicities. This was reflected in the survey responses (Figure 4), where participants were asked how they would manage a hypothetical patient with metastatic melanoma. In the case of a patient with regional and pulmonary metastases, the majority selected Nivo/Ipi (45%) or Nivo/Rela (36%), indicating a lack of consensus on the optimal regimen. However, when brain metastases were introduced, all respondents recommended Nivo/Ipi, reflecting the well-established efficacy of this regimen in patients with intracranial metastases and the paucity of data supporting Nivo/Rela in this setting¹¹. These findings highlight that while toxicity considerations play a role in treatment selection, clinicians weigh multiple factors—including central nervous system activity—when deciding between regimens.

Figure 4 - Treatment Selection in Advanced Melanoma

A.

B

Figure 4: Survey responses from journal club participants regarding the preferred first-line treatment for a patient with untreated metastatic melanoma to regional nodes and lungs (A) and a similar patient with an additional asymptomatic brain metastasis (B). While responses were divided in A, Nivo/Ipi was the overwhelming choice when brain metastases were present, reflecting the lack of supporting data for Nivo/Rela in this setting. Respondents who answered “I Do Not Feel Comfortable Making A Recommendation” (n=4) were omitted.

Beyond clinical factors, the potential role of biomarker-driven treatment selection warrants further exploration. Specifically, the impact of LAG-3 expression on treatment response remains unclear. In cases where LAG-3 expression is low (<1%), it is uncertain whether patients derive greater benefit from Nivo/Ipi over Nivo/Rela. However, routine LAG-3 testing has not been widely implemented, and LAG-3 level was not assessed in the CheckMate 067 trial cohort. Exploring the clinical utility of LAG-3 expression as a predictive biomarker could help refine patient selection and optimize treatment strategies.

It is also important to consider differences in treatment dosing between trials. The CheckMate 067 trial used a dual-therapy dosage of 1 mg/kg nivolumab and 3 mg/kg ipilimumab, compared to the fixed-dose combination of 480 mg nivolumab and 160 mg relatlimab used in RELATIVITY-047. Due to the higher toxicity associated with the 3 mg/kg ipilimumab regimen, an alternative dosing strategy (3 mg/kg nivolumab and 1 mg/kg ipilimumab) has been tested in advanced melanoma¹² and is now approved with improved tolerability in other malignancies, including unresectable malignant pleural mesothelioma, metastatic non-small cell lung cancer, and advanced renal cell carcinoma^13–16. This raises the question of whether Nivo/Rela truly offers a significant advantage in reducing TRAEs compared to the lower-dose Nivo/Ipi regimen.

Another key theme that emerged during the discussion was the role of clinician familiarity with statistical methodologies in interpreting and applying study findings. While indirect treatment comparisons offer a valuable approach when head-to-head trials are unavailable, their complexity may limit their influence on clinical practice. The majority of journal club participants were clinical experts and academic physicians with research expertise outside the field of biostatistics, and most had never personally conducted a propensity score analysis (Figure 5). Additionally, 26% of respondents reported being “Not at all” or “Not very” comfortable with the statistical methods used in the study (Figure 6). This may represent a barrier to changing practice, as clinicians could be more inclined to adopt findings from studies that use more familiar study designs or analytical approaches. As a result, while the methodology in this analysis was rigorous, its impact on real-world treatment decision-making may be tempered by the challenge of translating complex statistical models into actionable clinical insights.

Figure 5 - Experience with Propensity Score Analysis

Figure 5. Journal club participants were asked whether they had personally conducted a propensity score analysis. The majority (87%) had never performed such an analysis, highlighting a potential barrier to fully engaging with the statistical methodology used in the study.

Figure 6 - Comfort with Statistical Methods in the Study

Figure 6. Participants were asked to rate their comfort with the statistical methods used in the study. While 40% felt comfortable, a significant proportion (26%) reported being “Not at all” or “Not very” comfortable, suggesting that complex methodologies may limit the extent to which such analyses influence clinical decision-making.

In this context, an open science approach could have improved accessibility and transparency of the analysis. The data supplement provided with this study was informative, particularly in outlining the methodology behind the propensity score model and inverse probability of treatment weighting. However, given the complexity of these statistical methods, many readers would have found additional materials—such as a detailed supplementary methods section with open-source code—highly beneficial. Recognizing the sensitivity of individual patient-level data, the inclusion of a synthetic dataset allowing readers to replicate the analysis would have been a meaningful step toward transparency. While not yet standard practice across all areas of medicine, such efforts to promote reproducibility and open science would have strengthened the study, fostering greater confidence in the findings and their applicability to clinical decision-making.

A key strength of this study lies in its methodological approach. The investigators effectively demonstrated how patient-level data and propensity score matching can enhance cross-study comparisons in the absence of head-to-head randomized clinical trials. Given the expanding array of treatment options in medical oncology and the lack of direct comparisons among these therapies, ITCs—when applied to similarly designed trials and supported by robust sensitivity analyses and internal validation—offer a valuable alternative to large-scale, resource-intensive randomized trials.

Nonetheless, the inherent limitations of cross-trial comparisons cannot be overlooked. A major challenge is the availability of patient-level data, as the reliability of ITCs diminishes when such data are inaccessible. However, techniques like Matching-Adjusted Indirect Comparisons (MAIC) attempt to mitigate this limitation by reweighting individual patient data to better align baseline characteristics across trials¹⁷. Moreover, the clinical implications of these findings must be carefully interpreted in the context of evolving treatment landscapes. For example, the RELATIVITY-047 trial was conducted when Nivo/Ipi was already an established treatment option, whereas CheckMate 067 took place during the earlier phase of ICI trials in advanced melanoma. This temporal difference may have influenced clinician decision-making and patient selection, further complicating direct comparisons between the two trials.

In summary, this study employs a rigorous analytical approach and reinforces the similarities between Nivo/Rela and Nivo/Ipi, offering a valuable contribution to the ongoing evaluation of treatment options in advanced melanoma. While some discussants felt that the findings largely aligned with existing literature—including studies using less rigorous methodologies²⁰—others saw value in the study’s methodological strengths and its effort to provide a more structured comparison. Moving forward, direct randomized comparisons incorporating broader molecular profiling or further investigations into predictive biomarkers may help refine treatment selection for advanced melanoma.

Materials and Methods

This Perspectives on the Science piece was published using Quarto®. The survey was conducted using REDCap®. The figures depicting the survey data were created using R (version 4.0.0) and the tidyverse suite of packages, including ggplot2. The image on the “Perspectives on the Science” page was created by the authors (DMM) using the rosemary package. GPT-4, a language model developed by OpenAI, was employed in the drafting and editing of this manuscript. GPT-4 provided assistance in manuscript structuring, and generation of content, ensuring a comprehensive and cohesive presentation of the research and discussion points²¹.

Bibliography

References

1.

Long, G. V. et al. First-line nivolumab plus relatlimab versus nivolumab plus ipilimumab in advanced melanoma: An indirect treatment comparison using RELATIVITY-047 and CheckMate 067 trial data. J Clin Oncol 42, 3926–3934 (2024).

2.

Hodi, F. S. et al. Improved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med 363, 711–23 (2010).

3.

Robert, C. et al. Anti-programmed-death-receptor-1 treatment with pembrolizumab in ipilimumab-refractory advanced melanoma: A randomised dose-comparison cohort of a phase 1 trial. Lancet 384, 1109–17 (2014).

4.

Weber, J. S. 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 16, 375–84 (2015).

5.

Larkin, J. et al. Combined nivolumab and ipilimumab or monotherapy in untreated melanoma. N Engl J Med 373, 23–34 (2015).

6.

Tawbi, H. A. et al. Relatlimab and nivolumab versus nivolumab in untreated advanced melanoma. N Engl J Med 386, 24–34 (2022).

7.

Gutzmer, R. et al. Atezolizumab, vemurafenib, and cobimetinib as first-line treatment for unresectable advanced BRAFV600 mutation-positive melanoma (IMspire150): primary analysis of the randomised, double-blind, placebo-controlled, phase 3 trial. The Lancet 395, 1835–1844 (2020).

8.

Lamba, N., Ott, P. A. & Iorgulescu, J. B. Use of first-line immune checkpoint inhibitors and association with overall survival among patients with metastatic melanoma in the anti-PD-1 era. JAMA Netw Open 5, e2225459 (2022).

9.

Wolchok, J. D. et al. Long-term outcomes with nivolumab plus ipilimumab or nivolumab alone versus ipilimumab in patients with advanced melanoma. J Clin Oncol 40, 127–137 (2022).

10.

National Comprehensive Cancer Network, I. NCCN clinical practice guidelines in oncology (NCCN guidelines) for melanoma: cutaneous. (2025).

11.

Tawbi, H. A. et al. Long-term outcomes of patients with active melanoma brain metastases treated with combination nivolumab plus ipilimumab (CheckMate 204): Final results of an open-label, multicentre, phase 2 study. Lancet Oncol 22, 1692–1704 (2021).

12.

Lebbe, C. et al. Evaluation of two dosing regimens for nivolumab in combination with ipilimumab in patients with advanced melanoma: Results from the phase IIIb/IV CheckMate 511 trial. J Clin Oncol 37, 867–875 (2019).

13.

Baas, P. et al. First-line nivolumab plus ipilimumab in unresectable malignant pleural mesothelioma (CheckMate 743): A multicentre, randomised, open-label, phase 3 trial. Lancet 397, 375–386 (2021).

14.

Hellmann, M. D. et al. Nivolumab plus ipilimumab in advanced non-small-cell lung cancer. N Engl J Med 381, 2020–2031 (2019).

15.

Motzer, R. J. et al. Nivolumab plus ipilimumab versus sunitinib in advanced renal-cell carcinoma. N Engl J Med 378, 1277–1290 (2018).

16.

Bristol-Myers Squibb. Yervoy (Ipilimumab) Package Insert. (2023).

17.

Signorovitch, J. E. et al. Matching-Adjusted Indirect Comparisons: A New Tool for Timely Comparative Effectiveness Research. Value in Health 15, 940–947 (2012).

18.

Zhao, B. W. et al. LAG3-PD1 or CTLA4-PD1 inhibition in advanced melanoma: Indirect cross comparisons of the CheckMate-067 and RELATIVITY-047 trials. Cancers (Basel) 14, (2022).

19.

Boutros, A. et al. Activity and safety of first-line treatments for advanced melanoma: A network meta-analysis. Eur J Cancer 188, 64–79 (2023).

20.

Thosar, K., M. Network meta-analysis (NMA) of immunooncology (IO) treatments for first-line (1L) advanced or metastatic melanoma. vols. Poster CO62 (2023).

21.

OpenAI. GPT-4: Language model. (2023).

Appendix

Abbreviations, Disclosures, Disclaimer, License, Publication Stage

Abbreviations

cORR: confirmed objective response rate, DOR: duration of response, FDA: food and drug administration, GI: gastro-intestinal, HR: hazard ratio, ICI: immune checkpoint inhibitors, Ipi: ipilimumab, IPTW: inverse probability of treatment weighting, ITC: indirect treatment comparison, JoCO: Journal Club, MSS: melanoma-specific survival, Nivo: nivolumab, Nivo/Ipi: nivolumab plus iplimumab, OS: overall survival, PFS: progression-free survival, SoCO: Society of Cutaneous Oncology, TRAEs: treatment-related adverse events

Disclosures

DMM has received honoraria for participation in advisory boards from Bristol Myers Squibb and Merck.

Disclaimer

This site represents our opinions only. See our full Disclaimer

Reuse

This work is licensed under a creative commons BY-NC-ND license

Publication Stage

• Draft