No Improvement in RFS With Adjuvant Nivolumab/Relatlimab in Advanced Melanoma

Compared with nivolumab alone, nivolumab/relatlimab did not improve RFS in patients with resected stage III to IV melanoma.

Results of the phase 3 RELATIVITY-098 study (NCT05002569) found nivolumab (Opdivo) plus relatlimab failed to improve RFS compared with nivolumab alone for patients with completely resected stage III to IV melanoma, according to the presentation from the 2025 American Society of Clinical Oncology (ASCO) Annual Meeting.

The median RFS with nivolumab/relatlimab (n = 547) was not reached (NR; 95% CI, 30.8-NR) by investigator assessment vs 33.1 months (95% CI, 31.0-NR) with nivolumab alone (n = 546; HR, 1.01; 95% CI, 0.83-1.22; P = .928). The 12-month RFS rates in the respective arms were 75% (95% CI, 72%-79%) and 72% (95% CI, 68%-75%), respectively; the 24-month rates were 62% (95% CI, 58%-66%) and 64% (95% CI, 59%-68%).

“Adjuvant nivolumab combined with relatlimab did not significantly improve the RFS compared with nivolumab alone in resected stage III or IV [melanoma]. The overall survival was not tested. The safety profile was similar to what we see in the advanced setting,” Georgina V. Long, MD, of Melanoma Institute Australia, The University of Sydney, Royal North Shore and Mater Hospitals, in Sydney, NSW, Australia, said in a presentation of the data. “We did see a higher number of LAG3 CD8 T-cell frequency in the tumor than in the periphery in the advanced setting; however, in blood samples, we saw higher levels of LAG3-expressing T cells in the advanced setting compared with [the] adjuvant [setting]. On treatment, we saw a greater magnitude increase in the LAG3-positive T cells.”

Reflecting on RELATIVITY-098: Design, Treatment, Objectives

The trial enrolled patients who were at least 12 years of age and had completely resected stage III to IV melanoma and an ECOG performance status of 0 or 1. Patients who previously received systemic anticancer treatment were excluded. Participants were randomly assigned 1:1 to receive 480 mg of nivolumab plus 160 mg of relatlimab every 4 weeks (n = 547) vs 480 mg of nivolumab every 4 weeks (n = 546). Treatment continued until 1 year, recurrence, intolerable toxicity, withdrawal, or death.

Stratification factors included disease stage per The American Joint Committee on Cancer v8 criteria (IIIA to IIIB vs IIIC vs IIID to IV) and geographic region (USA/Canada/Australia vs Europe vs rest of world). Patients were followed for at least 8 years or until death, loss to follow-up, withdrawn consent, or study completion. The primary point of the study was RFS, and a key secondary end point was overall survival (OS). Other end points included distant metastasis-free survival (DMFS), safety, and progression-free survival 2. Exploratory biomarker analyses were also performed.

RFS was defined as the time from randomization to recurrence—either distant, local, regional, or a new primary melanoma, or death from any cause. Monitoring with CT scans [was done] every 3 months for the first 3 years, and then 6 monthly thereafter. Those who did not undergo a complete lymph node dissection required ultrasound of the relevant nodal basin at those same intervals,” Long noted. “For this RFS analysis, we had 427 events and 410 were planned; this gives us a cumulative power of 91.5%. The OS end point was not analyzed as per the methods, which included a hierarchical testing strategy. There were 148 events at this analysis, which was only 48% maturity.”

The data cutoff date for the analysis was December 16, 2024. Long noted that the baseline patient characteristics were “broadly similar” between the arms. The median patient age in the relatlimab and nivolumab-alone arms was 59 years (range, 18-92), about 60% were male (60%; 58%), most had an ECOG performance status of 0 (90%; 91%), and 40% had BRAF-mutated disease. Moreover, 58% of patients in both arms were from Europe, around half in both arms had stage IIIC disease (49%; 50%), and most had cutaneous nonacral melanoma (80%; 83%).

“I will just draw your attention to the PD-L1 expression less than 1%, which was [noted in] 58% [of those in] the combination arm,” Long said. “I just want to make a comparison with RELATIVITY-047 [NCT03470922] in the advanced setting. If we look at [all] evaluable patients, the PD-L1 expression less than 1% was actually [noted in] 72% [of those] in RELATIVITY-098 and 59% in the advanced setting in RELATIVITY-047.”

With regard to patient disposition, fewer patients in the nivolumab/relatlimab arm completed treatment than those in the nivolumab-alone arm, at 52.1% and 60.0%, respectively. In the combination arm, 48% of patients discontinued treatment; 21% did so because of study drug toxicity and 19% did so because of disease recurrence. In the monotherapy arm, 40% discontinued with 12% doing so because of toxicity and 23% doing so because of recurrence.

The median duration of therapy in both arms was around 11 months; the mean duration was 8.11 months (range, 0.03-12.06) in the combination arm and 8.67 months (range, 0.03-12.06) in the monotherapy arm. The percentages of patients who received subsequent systemic treatment were similar between the arms, at 27% and 28%, respectively, although more patients in the monotherapy arm received PD-1/CTLA-4 inhibitors vs those in the combination arm (17% vs 14%), Long noted.

Additional Efficacy Revelations

“In terms of prespecified subgroups, there was no specific subgroup that seemed to benefit from one treatment over the other,” Long said. She did point out that the hazard ratio for RFS in those with nonacral cutaneous subtype (n = 890) was 0.87 (95% CI, 0.70-1.09), “which is tight, just crosses 1,” she said. “Similarly, if you look at geographic region, and this may be due to a predominance of nonacral cutaneous melanoma, the hazard ratio is 0.85 [95% CI, 0.56-1.31] for [USA/Canada/Australia].” She added that baseline PD-L1 and LAG3 status, as well as BRAF status, did not appear to favor one approach over the other.

Recurrence events were similar in the treatment arms. In the nivolumab/relatlimab arm, 37% of patients experienced recurrence vs 38% of those in the nivolumab-alone arm. Predominantly, patients experienced distant recurrences (19% vs 20%). Deaths occurred prior to recurrence for 13 patients in the combination arm and 7 patients in the monotherapy arm.

The median DMFS in the combination arm was NR (95% CI, NR-NR) vs 33.1 months (95% CI, 31.5-NR) in the monotherapy arm (HR, 1.07; 95% CI, 0.84-1.36). The 12-month DMFS rates in the respective arms were 85% (95% CI, 81%-88%) and 84% (95% CI, 80%-87%); the 24-month rates were 73% (95% CI, 69%-77%) and 76% (95% CI, 72%-80%).

Safety Spotlight

The rate of treatment-related adverse effects (TRAEs) occurred more frequently in the combination arm vs the monotherapy arm, at 89% vs 80%; these effects were grade 3 or 4 for 19% and 8% of patients, respectively. In the combination arm, the most common TRAEs were hypothyroidism (any grade, 25%; grade 3/4, 1%), fatigue (24%; <1%), pruritus (18%; 0%), hyperthyroidism (17%; 1%), rash (15%; 0%), arthralgia (13%; <1%), and diarrhea (11%; 1%).

Any-grade TRAEs led to treatment discontinuation for 17% of patients in the relatlimab arm vs 9% of those in the nivolumab-alone arm. Treatment-related deaths occurred in 2 patients in the combination arm vs 1 patient in the monotherapy arm.

“Endocrinopathies were higher in the combination arm than [the] nivolumab[-alone arm], particularly adrenal insufficiency and hypophysitis, as well as diabetes requiring insulin,” Long said with regard to immune-mediated toxicities. “Just to point out a significant toxicity, which is important to note, and those are related to muscles—myasthenic syndrome, myocarditis, or the myositis—was slightly more frequent in the combination arm. In fact, there was 1 death due to myocarditis and myositis; however, that occurred in the nivolumab-alone arm.”

Breaking Down Biomarker Analyses

Long then posed the question: Why was a significantly improvement in efficacy not observed in RELATIVITY-098 but was in RELATIVITY-047?

She noted that the CODEX/PhenoCycler was leveraged to compare T cell frequencies in tumor tissue samples collected at baseline in the latter trial. For both trials, longitudinal flow cytometry was used to conduct blood biomarker analyses in both trials; blood was collected at baseline and on treatment (day 1, cycle 1). They used the Wilcoxon test to compare T-cell frequencies in both the tumor and the blood and between patients in both studies.

“Let’s just look at RELATIVITY-047, so the advanced setting. You can see in the tumor, PD-1–positive CD8, LAG3 CD8, and double-positive CD8 was similar in frequency in the tumor. In fact, these were not significantly different when you look at LAG3 and PD-1,” Long reported. “When we look at the blood, we can see that there is significantly fewer peripheral LAG3 CD8 cells compared to the tumor, which is present when these bloods are taken. We also see significantly fewer LAG3 CD8 cells compared with PD-1 CD8 cells in the periphery blood.”

She then set out to make a comparison between adjuvant and advanced settings and noted comparable frequencies of PD-1 CD8 cells in the blood. “However, if you look at the baseline LAG CD8, you can see that they are higher in frequency in the advanced setting than the adjuvant. If we look at the double positive CD8s, again in the advanced setting, you can see these fatigued cells are higher in the advanced setting compared with adjuvant.”

She closed by pointing out a greater delta change from baseline to on-treatment in the advanced vs the adjuvant setting. “We have a target, LAG3, and we have a drug, relatlimab…More work is being done on a biomarker analysis and will be presented later this year,” Long said.

Reference

Long GV, Ascierto PA, Guo J, et al. Nivolumab plus relatlimab vs nivolumab alone for the adjuvant treatment of completely resected stage III-IV melanoma: primary results from RELATIVITY-098. J Clin Oncol. 2025;43(suppl 17):LBA9500. doi:10.1200/JCO.2025.43.17_suppl.LBA9500