Transcript:Robert A. Figlin, MD: At ASCO 2016, you’re going to present data looking at 3- and 4-year survival from some of these early nivolumab trials. Can you summarize what you think that data will have in terms of the impact of the conversation with the patient when you see one-third of patients possibly benefitting from a checkpoint inhibitor out 3 and 4 years?

David F. McDermott, MD: You don’t want to make too much out of nonrandomized data. But we have over 4 years of follow-up, minimum, on both studies now. And, as you mentioned, one-third of patients on both studies—over 200 patients—are alive at 4 years. In the phase I trial, where we have 5 years of follow-up, one-third of patients are alive at 5 years. That’s a pretty remarkable number when we consider back 5 or 10 years ago, the 5-year survival of all patients with kidney cancer was around 5% or 10%. We’re now seeing that with these salvage agents, so hopefully that survival benefit will be reflected again in the phase III trial, as well.

Other things we saw from that study were that patients benefitted regardless of risk group. So, we saw long-term survival in good-, intermediate-, and poor-risk groups. We’ve had a lot of conversations about what the Memorial Sloan Kettering criteria mean, but we saw benefits in all those groups regardless of performance status. If you had slightly reduced performance status, there were still some long-term survivors regardless of response, meaning half of our patients who survived 4 years on the phase II trial had either stable disease or progression of disease, which was Dan’s point earlier.

This turns a lot of things we thought on its head. The way I want to look at it is this is a good story. It makes it more difficult to talk to patients, but not in a bad way. It’s much more difficult to tell a patient we don’t have much for you. Here, we can come up with our own spiel, our own pathway with individual patients, but we’ll be talking about good options as opposed to no options.

Robert A. Figlin, MD: Dan, you had some thoughts?

Daniel J. George, MD: Yes. Look, I don’t mean to say we don’t treat in the third-line or we can’t do clinical trials in the third-line. We absolutely can, and there is definitely this subgroup that do well with these patients. Probably a lot of the patients on your phase I and phase II trials are patients that have gone through multiple lines of therapy. So, teasing out what’s a treatment effect and what is a natural history selection effect is challenging. And so we need randomized data at the end of the day for that. The other thing I would just say, in your randomized data in both METEOR, as well as in CheckMate-025, if you look at the patients in the Forest plots who have had two TKIs, those plots, those lines, are much closer to no benefit. I’m not saying there’s no benefit, but there’s probably diminished benefit when you get further and further out with any agent at the end of the day. It’s just the disease gets stronger, and, at some point, that’s going to diminish. So, it’s remarkable to me that your phase I and phase II data have this tail. How much of it is selection and how much of it is treatment effect? It’s got to be both. I think it’s a combination of both.

David F. McDermott, MD: Right. And it’s the sequence of drugs that followed.

Elizabeth R. Plimack, MD, MS: Right, not all of them are on drugs.

Daniel J. George, MD: And there may be a “secret sauce” of sequence that you have.

Robert A. Figlin, MD: Although I will say, again, not very different than the conversation that existed 30 years ago when we talked about a tail of the curve to IL-2. Those were not randomized trials. And I remember the conversation as it was yesterday, which was, how do you know the tail is just not the natural history of the disease? We now understand that that’s not the natural history of the disease. The natural history is people go on to die and they don’t live 10, 20, and 30 years later. But you have to start at some place, and that’s the key.

Transcript Edited for Clarity