Interpreting Nivolumab/Relatlimab PFS/ORR Benefit, Safety in Melanoma
During a Case-Based Roundtable® event, Michael A. Postow, MD, reviewed the use of nivolumab plus relatlimab for patients with advanced melanoma in the first article of a 2-part series.
Michael A. Postow, MD
Chief, Melanoma Service
Co-Director, Melanoma Disease Management Team
Memorial Sloan Kettering Cancer Center
New York, NY
Targeted Oncology: Could you discuss the significance of the RELATIVITY-047 study (NCT03470922) in treating patients with advanced BRAF-negative melanoma?
MICHAEL POSTOW, MD: The RELATIVITY-047 study, the most recent study that came out, was on nivolumab plus relatlimab [Opdualag], which is a LAG3-blocking antibody, vs nivolumab [Opdivo] alone. The primary end point was progression-free survival [PFS] by the blinded independent central review [BICR]. The baseline characteristics of the patients are very well balanced as we would expect from a randomized trial for demographics, M-stage, and BRAF mutational status across the different groups.1,2
[Looking at] PFS by BICR, the takeaway message is that when you add in relatlimab as LAG3 blockade to nivolumab, you improve PFS. This was a statistically significant P value and the HR was 0.79 [95% CI, 0.62-0.92; P = .0055].2 It's not the [best] HR; the Kaplan-Meier curves aren't totally separated, but they are separated by approximately 10%, so 1 in 10 patients who are treated with the combination won’t have disease progression compared with the patients whom you've treated with nivolumab monotherapy. This is proof that nivolumab/relatlimab is better in terms of PFS than nivolumab alone.
One of the things that is challenging to reconcile is although the PFS was statistically significantly improved, overall survival [OS] was not statistically significantly improved. It looked like the HR was favoring the nivolumab/relatlimab group and that arm is appearing better than that [nivolumab] arm on the Kaplan-Meier curve, but you can't say it's better because it was not statistically significantly different in the way that this trial was designed [HR, 0.80; 95% CI, 0.64-1.01; P = .0593].1,2 In my opinion, if there was an OS difference, then no one should get single-agent PD-1 [inhibitor] because you have drugs in combinations that improve OS. But in both the CheckMate 067 study [NCT01844505] and the RELATIVITY-047 study, the PFS and objective response rates [ORRs] were better for combinations, but the OS is still not significantly improved [vs nivolumab alone].
I like to think about nivolumab/relatlimab as a ‘10%’ drug. It has 10% better PFS and 10% better ORR. There was a 43.1% ORR of nivolumab/relatlimab vs 32.6% for nivolumab monotherapy in this trial. Nivolumab monotherapy had a lower ORR in this trial than in other studies. Some people say nivolumab/relatlimab looks just like other studies of single agent PD-1 with a 43% ORR, but I think that's a little unfair to do. You have to look at a randomized trial by itself and not do the cross-trial comparisons.
We have 2-year data from an abstract from the 2023 American Society of Clinical Oncology Annual Meeting.3 The median PFS differed [10.2 months with the combination vs 4.6 months with nivolumab (HR, 0.81; 95% CI, 0.67-0.97)]. At the 2 year mark, the OS rates favor nivolumab/relatlimab with the longer follow-up that we have [51.5% vs 42.5% for nivolumab at 48 months]. That was not statistically significantly different, but the numbers look good, and that ORR still differs by 10%: 43.7% vs 33.7%.
What stood out in terms of safety outcomes on this trial?
The rate of treatment-related adverse events [83.7% vs 72.4%, respectively] and those leading to discontinuation [15.2% vs 7.2%, respectively] were a little higher for the combination [vs nivolumab alone].1,2 Rates of grade 3 or 4 toxicity were exactly 10% higher: 21.1% vs 11.1%. So we have 10% higher grade 3/4 toxicity, 10% higher PFS, and 10% higher ORR. There was a little bit more toxicity than PD-1 inhibitor alone, but not a lot more, and a bit more benefit than PD-1 inhibitor alone, but not a ton more.
[Concerning] the details of the different types of toxicities, adrenal insufficiency is the one thing that seems to be more prevalent with nivolumab/relatlimab [5.4% vs 1.1% with nivolumab]. Rates of grade 3 or 4 adrenal insufficiency were very similar across these arms [1.7% vs 0%, respectively]. Even any-grade adrenal insufficiency can be bad, because grade 2 means you have to be on treatment for something so that [nearly] 6% is a meaningful group of patients.
Do you recommend performing the troponin monitoring that was done for the first 2 months of the study?
The study checked troponins because that was a concern from earlier development with relatlimab. There was a very low rate of myocarditis in this study [1.7% with nivolumab/relatlimab vs 0.6% with nivolumab],1 as there is in many experiences with checkpoint blockade. But you can't compare the rates of myocarditis with this combination compared with other combinations because this study looked for myocarditis so much more diligently and we found asymptomatic troponin [increase] in a lot of these patients and didn't know what to do with it. I wouldn't say that the rates of myocarditis are higher with this than other combinations of checkpoint blockade.
In practice, I check on baseline troponin and that's it. That’s because someone [may] come in with nonspecific symptoms, and the emergency department checks the troponin at some point and [may find] it's mildly elevated with the new high-sensitivity assays. But if it was mildly elevated at baseline for renal dysfunction or other reasons, then it [won’t] have to go down the myocarditis care pathway. I've checked the baseline troponin in patients with checkpoint blockade when I give them combination regimens, either ipilimumab/nivolumab or nivolumab/relatlimab.
We did some experiments where we would just check on treatment with each dose. It ended up being annoying because they would come up as a bit more positive, but they were asymptomatic and electrocardiograms were normal.… We didn't find that most of those patients who had the on-treatment changes actually had myocarditis upon exhaustive interventions, and it led to treatment delays and a lot of anxiety on all fronts. So I [test] baseline…I found it helpful when patients get sick with other adverse events and that troponin ends up getting checked later on. Then at least you can compare it with baseline.
