New Data for Melanoma Patients Who Progress After PD-1 Blockade

Jeffrey S. Weber, MD, PhD


December 08, 2022

This transcript has been edited for clarity.

I'm Jeffrey Weber and I'm a medical oncologist at the Laura and Isaac Perlmutter Cancer Center here in New York City. Today, I'll be commenting on a couple of abstracts presented at this year's AACR meeting, which was held in New Orleans, one of the first live meetings that we've had in quite a while.

I'll describe two abstracts that I thought had some importance. One was a report on the SWOG S1616 trial, which was a trial of ipilimumab (ipi) with nivolumab (nivo) at what we call standard doses vs ipi alone in patients with unresectable or metastatic melanoma who had not responded to anti-PD-1 therapy. Again, this is a big problem because the question for many oncologists is, what do you do when someone fails PD-1? How do you manage them?

This was a randomized phase 2 trial with an interesting 3-to-1 randomization. A total of 91 patients were ultimately treated who either got ipilimumab alone at 3 mg/kg, which is, I would say, a relatively standard dose, vs so-called standard-dose ipi/nivo with ipi at 3 mg/kg and nivo at 1 mg/kg during the induction, followed by up to 2 years of nivolumab at 480 mg every 4 weeks intravenously. The primary endpoint was progression-free survival (PFS).

The data suggest that there was a modest increment in PFS, which I thought was a bit of a surprising endpoint for an IO agent. Nonetheless, there was a 3-month median PFS for ipi/nivo vs 2.7 [for ipi alone]. That's a very small difference. The 6-month estimate was 34 vs 13 months, which seems pretty different. The hazard ratio was 0.63 with a one-sided P value of .037. It looks like the combo is the winner, as we might expect.

Again, if you look at the forest plot, it tended to favor the combo for virtually all the parameters, except for, as you might expect, prior adjuvant therapy, which usually would have been with PD-1 blockade alone.

If you look at the duration of response, interestingly, it was not reached with ipi — although a small number of responders — but it was 18 months for ipi/nivo. Response rates were 28% for the combo vs only 9%, as you might expect, for the ipi arm alone.

Toxicity, as you would also expect, was higher, at something like 56% grade 3/4 for the combo. It was only 31%, a little higher, but not too far off from what you would expect, with a reasonable dose of ipi alone given to patients who had already been through therapy. Again, if you look at the waterfall plot, it clearly looks like you had many more responses, with some complete responses (CRs). I think there were something like 6 CRs total in the folks who got the ipi plus nivo.

However, at the end of the day, when you look at the survival, the curves started out looking better for the combination, but they cross over and the medians were really not different, at 25.7 for ipi and 21.7 for the combination. The 12-month survival was a little better for the combination at 63 vs 57, but the hazard ratio for survival overall was 0.93 and the P value was .4, so there was no real difference.

Again, as some correlative marker studies showed that there were increases in CD8 T cells but pretreatment CD8+ T cells didn't help you, they were not predictive. The investigators — this was presented by Ari VanderWalde — conclude that ipi/nivo is inappropriate next-line therapy in those who have not responded to anti-PD-1 alone.

When Mario Sznol discussed it, he did have some concerns over the durable PFS difference, declining to less than 10% when you go much beyond a year. The question arose as to whether a higher dose of ipi would have been any better. The problem is that the higher dose of ipi is clearly very toxic.

The second abstract is something completely different. They tested yet another one of the intratumoral injectables with PD-1 blockade in those who were refractory to PD-1 blockade. This was the SPOTLIGHT203 phase 2 study presented by Iván Márquez-Rodas from the Hospital General Universitario Gregorio Marañon in Madrid.

This is an interesting drug, which is an intratumoral injection. It's a nanoplexed, double-stranded RNA that activates TLR3 and it also activates other inflammatory genes like RIG-1 and MDA5. It can theoretically improve antigen presentation and enhance T-cell infiltrate. Again, in the tumors that don't have a T-cell infiltrate, this might be very useful.

There was a first-in-human study in a small number of patients — I think fewer than 30 patients — where there was a 20% response rate. This was an attempt, in a slightly larger group of patients, to try to reproduce those numbers.

Baseline characteristics this time included acral lentiginous and mucosal melanomas in addition to patients with cutaneous melanoma. In terms of safety and toxicity, the percentage of patients with grade 3/4 side effects was relatively trivial at 5%. There were 40 patients treated on this trial, so those are, I think, reliable numbers. Overall, I think that this was a very well-tolerated regimen.

Best overall response rate was 25% with a CR rate of 10%, which on the surface of it looks pretty encouraging. The partial response (PR) rate was 15%, and 40% had stable disease at 24 weeks, so the disease control rate was 65%. When you look at duration of response, it's fairly favorable for those who have noncutaneous disease, as it has not been reached. It was something like, at 16 months, a 66% continuation of response.

When you look at the data closely, the vast majority of responses were in the injected lesions. The response rate was less than 10% in the noninjected lesions. I think there were three bona fide responders of the 40 in the noninjected lesions. To me, that's a really important piece of information.

If you look at PFS for the 42 patients in the intent-to-treat analysis, the median was only 3.8 months, which is fairly modest. If you look at the median PFS of the mucosal and acral patients, or those who had lactate dehydrogenase (LDH) more than three times the upper limit of normal, it's actually pretty modest at 2.2 months, but not reached in those who have nonacral/nonmucosal melanoma with what we think are probably normal LDH rates.

Again, the conclusion of the investigators is that the data warrant further investigation. I can't argue with a 25% response rate in PD-1–refractory patients. The big issue is whether this is really something that's going to be useful only for those where the injectable lesions are those that are predominant in those that you monitor. Or will this really be effective for distant disease, which is the parameter that's going to affect survival in the real world?

Please feel free to comment on these presentations and my suggestions. I'm Jeffrey Weber and I thank you for your attention.

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.