Recent Trial Data on Tebentafusp in Metastatic Uveal Melanoma

Ryan Sullivan, MD: Now is a good time to talk a little about the clinical trial data, and then we can have a discussion about the toxicities and the management of the toxicities. One of the toxicities, as you mentioned, is cytokine release syndrome, which requires hospitalization. It’s important to note that the clinical trial that led to the approval of this agent was built on phase 1 and 2 testing.

The initial trial of this agent was in cutaneous melanoma in patients who had surgical resection. It was presented by one of our colleagues, Mark Middleton, at an AACR [American Association for Cancer Research] meeting about 6 years ago. I remember watching him give this talk and thinking, “This is a really neat thing, but that’s a weird population to start looking at a drug like this.” It fit in with the field of melanoma, where there were a lot of adjuvant studies of vaccines, and this was maybe a weird kind of vaccine. Obviously, it isn’t that. It’s a very different type of molecule. But soon after, it transitioned to being looked at in both cutaneous and uveal melanoma.

To your point, the gp100 peptide is much more consistently expressed in uveal melanoma than in cutaneous melanoma for reasons that probably are immunologic. Cutaneous melanoma is typically so genetically altered by UV or radiation that it has to come up with all kinds of mechanisms to get around the immune system, whereas uveal melanoma is much less genetically disturbed. It’s probably better able to remain differentiated as opposed to the dedifferentiated subtypes that we see in cutaneous melanoma. With that expression of the protein, it seemed like patients with uveal melanoma are doing better. But interestingly, doing better doesn’t mean high response rates.

We’ve collectively treated a number of these patients, Dr Hamid [Omid Hamid, MD]. I have a few responders and you probably have a few responders. But we have a lot of patients who have been treated for a while on this drug who didn’t respond but survived longer than we might have predicted. In fact, that was seen in those early studies. Patients seemed to live longer than we thought they would, based on their disease. Based on that, the design of the study wasn’t looking at response rate or progression-free survival, which are typical metrics for primary end points on definitive trials. The primary end point was overall survival, and it met its overall survival comparison of the Kaplan-Meier curves. In fact, the hazard ratio for overall survival was 0.51, which is about as good as you’re going to find in a clinical trial of a drug.

The trial design was 2:1 randomization. Patients were randomized to either tebentafusp or an investigator choice. We had 3 options in the investigator choice [arm]: single-agent pembrolizumab, single-agent ipilimumab, or dacarbazine. Most people recommended pembrolizumab as the investigator choice. Second to that was ipilimumab, and third to that was dacarbazine. Ultimately, this was a positive trial that led to the FDA approval of this agent.

Transcript edited for clarity.