Melanoma Update: 3 Important Trials from ESMO 2021

Jeffrey S. Weber, MD, PhD


December 22, 2021

This transcript has been edited for clarity.

I'm Jeffrey Weber, a medical oncologist at the Laura and Isaac Perlmutter Cancer Center in New York City, and today I'd like to discuss three melanoma abstracts that were presented at the recent ESMO meeting held in Paris.

The first is the RELATIVITY-047 study. These were data presented by Steve Hodi at the meeting, and he discussed the updated results of a global, randomized, double-blind phase 2/3 study in which patients with previously untreated metastatic melanoma that was unresectable received either nivolumab at the standard 480 mg once every 4 weeks intravenously, or they got nivolumab ("nivo") plus the LAG-3 antibody relatlimab ("rela").

This was a very nice, large, 700-plus–person study with the primary endpoint of progression-free survival (PFS). Indeed, the updated PFS was positive, as previously presented at ASCO. Now, we see more updated data with further follow-up, where the hazard ratio for the superiority of the combination for PFS was 0.75. The median PFS was 10.1 months vs 4.6 for the control group that received nivolumab alone — again, more than a doubling of PFS with a break at 3 months, and the curves stay apart all the way. Very impressive data.

The forest plot also looked quite impressive for almost all of the prespecified subgroups. The hazard ratios were to the left of 1, with good 95% CIs by age, gender, lactate dehydrogenase (LDH) level, performance status, tumor burden, M stage, and programmed death ligand 1 (PD-L1). And, interestingly, even by LAG-3 — whether more than 1% or less than 1% — you still had benefit, meaning that LAG-3 staining, as previously seemed to be important, was really not that important in this study (which was kind of an interesting concept).

In terms of safety, yes, there was more toxicity, but in general, the combination of relatlimab plus nivolumab is well tolerated, with 18.9% grade 3/4 treatment-related adverse events. That's a pretty darn good number. Again, it was much less for nivolumab alone, at 9.7%.

In terms of other evidence of benefit, there is an idea that's been recently promulgated: the treatment-free interval that starts with discontinuation of study. It was longer — it was doubled from 1.4 to 3.2 months in the combo therapy arm vs nivolumab alone.

If you look at PFS2 — which is a somewhat unusual variable, calculated from the start of therapy until the second progression after there is progression on relatlimab plus nivolumab — even here, there was a significant break at 6 months and the curves stayed apart, with a hazard ratio of 0.77. All in all, these are very nice data demonstrating a clear benefit for the combo of relatlimab plus nivolumab. We look forward to seeing the response data and hopefully survival data in the next year or so.

The next study was presented by Jason Luke. Again, it was an update — this time on the KEYNOTE-716 study, which is an adjuvant study of pembrolizumab vs placebo in patients with resected high-risk stage IIb and IIc melanoma who were rendered free of disease. They were stratified by T substage: that is, T3b and T4a, which are stage IIb patients, or T4b, which are stage IIc patients.

The duration was 1 year. Relapse-free survival (RFS) was the primary endpoint. Secondary endpoints were distant metastasis–free survival, overall survival, and quality of life. It was a very well-balanced study. As you look at this study of over 900 patients by any prognostic category you're interested in, such as age, T category, or substage, it was very well balanced.

If you look at the RFS curves, they're beginning to break apart more than they were at the initial presentation. We're looking at very nice curves with a hazard ratio of 0.65. Again, the early presentation was only a press release, but it looks very nice. The curves break apart at 6 months, which is the first evaluation, not every 3 months; the patients are evaluated every 6 months in this trial. The curves stay apart and the hazard ratio, as I said, is 0.65. The P value is .006 for RFS. If you include new primary melanomas, which is an unusual thing to include in an RFS curve, even then the hazard ratio was 0.64. There were very few new primaries.

In looking at the forest plot, a couple of interesting things stand out. If you look at the RFS for the key subgroup of geographic region, there were relatively few US patients. But the patients outside the US did much better, with a vastly superior hazard ratio of 0.61 compared with 0.91 for the US patients, which is hard to explain. If you look at the T4b patients, who are the stage 2cs, the hazard ratio is only 0.94, whereas it's 0.44 and 0.43 for the T4a and T3b (stage 2b) patients.

It's very interesting and the opposite of what one might have expected, but nonetheless, clear benefit was seen for adjuvant pembrolizumab vs placebo in stage 2b/2c patients as a group.

Adverse events were typical and what you would expect. It's a 16% grade 3/4 treatment-related adverse event rate and a 15% rate of stopping, which is maybe a little higher than you might expect from a stage 3 trial. Again, I think the patients here were stopped pretty quickly by the investigators because of their generally good outcome. Interestingly, in terms of toxicity, something like 20% of patients had a hormonal adverse event and may have required the use of supplements, and that's a big deal.

However, if you look at the quality of life (QoL), it looks very good. Looking at the QoL assessment by the European Organisation for Research and Treatment of Cancer (EORTC), the QLQ-C30 never gets below 10% for either placebo or the treatment arm, which is highly favorable.

The final oral presentation at this year's ESMO was given by me, and it was an initial assessment of a phase 2 study. That's an interesting study of ipilimumab and nivolumab, so-called flipped doses with the interleukin (IL)-6 receptor–blocking antibody tocilizumab. This is the first time we've seen a significant therapeutic trial with tocilizumab included in it.

We know that IL-6 is associated with a poor outcome and is a negative prognostic factor in melanoma and many other cancers, like lung cancer. We know that tocilizumab has been shown to reverse steroid-refractory immune-related adverse events in checkpoint inhibitor (CPI)–treated patients. Therefore, we designed a study with flipped dose of ipilimumab plus nivolumab, where we expect about a one-third rate of grade 3/4 treatment-related adverse events, and we added tocilizumab for the first 24 weeks, figuring that's when the toxicity would occur, with the goal of cutting the toxicity from 35 down to 20.

It was a two-stage study. To progress to stage 2, you had to have a 30% or less grade 3, 4, or 5 adverse event rate, and you had to have a response rate at least of 60%. Amazingly, in the first 18 patients, we met those and reported at ESMO on a total of 29 patients who had a week 12 evaluation.

The median follow-up was 8 months, which is relatively brief, but there was a 58% best overall response rate: 5 additional patients with stable disease for 24 weeks or more. There were only three deaths among those 29 patients, all due to progression. One was COVID-positive, and only one of the stable patients and one responder progressed at the time of the presentation, which means 20 of 22 patients with benefit stayed in remission — stable, PR, or CR.

That's pretty good, given that in CheckMate 511, the response rate was 47%. If you look at the toxicity, all together 41 patients were evaluable for toxicity (because it was not time-dependent, but rather applied to anyone who had started). The grade 3, 4, 5 treatment-related adverse event rate was 17% compared with double that in CheckMate 511. We had two cases of grade 3/4 liver function elevations, which got better quickly and patients continued therapy without a problem. We had a couple of cases of elevated lipase and nephritis, both limiting, and of note both patients progressed.

All in all, we did very well. We cut down the toxicity. We had a response rate as good as, or even better than, previously shown. It's very impressive that we only had 14% of patients stop due to toxicity, which is very favorable.

The biomarker work was also interesting. High IL-8 at baseline or at week 7 was associated with progression, and high tumor necrosis factor (TNF) alpha at baseline was associated with toxicity. We're going to continue this study, finish it up, and hopefully report to you next year.

This is Jeffrey Weber, reporting on the most recent ESMO meeting in Paris. Feel free to call in or inquire about questions. Thank you very much.

Jeffrey Weber, MD, PhD, is deputy director of the Laura and Isaac Perlmutter Cancer Center at NYU Langone Medical Center and co-director of its Melanoma Research Program. His research, which has been continuously funded by the NCI for more than 20 years, focuses on experimental therapeutics and drug development, particularly in immunotherapy.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.