Transcript:Keith Flaherty, MD: When you watch cohorts of patients, and I’m sure you’ve been using these drugs in trials for a number of years now, Jason, there is this phenomenon of patients progressing over time. I wonder if you could give a summary view of the patterns of progression. How do you respond to that? If you’ve given a patient PD-1 antibody monotherapy as a lead-in, what’s your strategy? Is the ipilimumab still on the shelf? What’s your tolerance for amount of progression and those sorts of things?

Jason Luke, MD, FACP: This is definitely an area where the trials didn’t necessarily allow us to do a lot of sequencing, but in clinical practice we’re figuring it out anyway because we have to. Certainly, we do see more rapid disease control with anti-PD-1 antibodies. In my practice, I actually can’t think of a patient that I have, who had a response and has progressed. I know that there are those patients, but I don’t personally have one.

Patients who have had initial responses do very well for a long period of time. That being said, if patients have stable disease or not that much response—it’s not like the patient is cured. I think that that’s a misnomer. We think we have these drugs now, and we have these great response rates. We still have a problem with melanoma. Thinking about the sequencing issue, I think there is pretty good data now that suggests that anti-PD-1 antibody, absent other particular caveats, really ought to be the frontline therapy.

A study led by Dr. Weber, and he can comment on it further, the CheckMate-064 study, looked at sequential ipilimumab, then nivolumab, or nivolumab, then ipilimumab. And the outcomes, as I understood them, was that the toxicity, in aggregate, had ended up being very similar, but the response and disease control was better when nivolumab was given upfront. And given the toxicity profile of ipilimumab, I certainly would rather not give it upfront if I didn’t have to.

Getting into the question about sequencing with the combination, I think it’s even more complicated, and I have to say that my personal opinion right now is I prefer to use single agent anti-PD-1 antibody upfront because it’s robust, you get response rates, you get disease control, and you don’t have this massive increase in toxicity profile. That’s not a hard and fast rule. There are going to be individual patients here and there, but that’s my preference when I think about frontline patients right now.

Rene Gonzalez, MD: I would quibble a little bit with that for two reasons. One is we have the really long tail with the ipilimumab that nothing comes even close, except for IL-2, maybe. But the other thing is, we don’t know how long to treat people with anti-PD-1. One of the beauties of ipilimumab is you get your four doses, and you’re done. You’re not coming in every two or three weeks to get your injection. Mentally it’s better, and we know that that response is probably going to be long lasting.

Just as a little example, I had a patient a year or so ago when anti-PD-1s were approved, but you had to have prior ipilimumab to be able to get them. I had a patient with a big lung tumor, coughing up blood, the kind of patient you want a rapid response in, but I couldn’t get the PD-1. So I gave him ipilimumab with the idea of, okay, I’ll jump through that hoop, give him one dose, call it residual disease, and he responded with one dose. Now he’s a year out and almost a CR and a very rapid response, and no more treatment. So there’s the occasional.

Jason Luke, MD, FACP: I totally hear that argument. I think that’s where you get in with the patient-specific sort of issues, but I think the data is pretty robust to suggest frontline PD-1 antibody. If you look at progression-free survival, response rate, overall survival, all these things point to PD-1. So I totally take your point, and that’s where I think our practical research in the space needs to be in the future. How many doses of anti-PD-1 do you need?

Rene Gonzalez, MD: Right.

Keith Flaherty, MD: Also, the aggregate data is unarguable. But, of course, you have a 40% response rate drug in one hand and a 10% response rate drug in the other. And you could imagine a disease like melanoma, that we always used to consider to be almost uniformly aggressive, with the subpopulation that’s somewhat indolent. I guess I‘m wondering is there anybody in your practice now, across your spectrum of patients by disease burden (high disease burden, middle of the road, low disease burden, indolent, unresectable but small volume), for whom this prospect of four doses is something you’d still think about?

Jason Luke, MD, FACP: Yeah. I’ve never done it, actually. It’s not a hard and fast rule. It really gets into the situation for the patient. For a patient with small volume lung metastases who’s otherwise completely asymptomatic, I have this idea. But, I think most patients with metastatic disease end up ECOG 0/1 just to start. Then, I have less tolerance for wanting to do that initially.

Jeffrey Weber, MD, PhD: I think that, generally, the data would strongly lead us to want to use PD-1 frontline. Jason was alluding to the CheckMate-064 trial that Steve Hodi and I led. The response rate for those who got nivolumab and then had a forced crossover, forced in the sense that it was an immediate crossover to ipilimumab, had a 47% response rate with an excellent PFS.

Those who went the other way and got ipilimumab for four doses at week 12 and then had a forced crossover (meaning if they had failed, and 80% of them will have failed, and you immediately would treat them) had a 22% response rate and not as good an overall survival, which was very surprising. They all got the same therapy. Very few patients dropped out. It was a randomized phase II trial. The only difference was the sequence. That’s a huge difference, and we’re still trying to figure it out. But I think those data are sufficiently compelling that I would generally give PD-1 first and only hold ipilimumab in reserve later. I would not see a good rationale for giving ipilimumab first.

Keith Flaherty, MD: Well, let me torture you a little bit more on this point, Dr. Weber, because I think that the subsets in which we practice, no one clinical trial, particularly a randomized phase II trial, really gives you a lot of granularity there. You could imagine that some patients with metastatic melanoma will progress in a three-month, four-time timeframe to a point at which they may not be as responsive to whatever that second-line therapy is. In other words, the ipilimumab/nivolumab sequence for some patients just may not get them to PD-1 quickly enough.

For the low disease burden indolent patient, sub-centimeter multifocal lung only—you’ve been using ipilimumab in clinical trials and clinical practice for an awfully long time—what would you quote to be your ballpark response rate with what I said was a 10% response rate drug but in that population?

Jeffrey Weber, MD, PhD: I would tell that patient it’s a 15% response rate because those with low burden M1a and 1b disease tend to do better than the M1c visceral disease patients. And another 10% are going to be stable or will have progression followed by regression. So I give them a solid 1 in 4 chance of having significant clinical benefit. If, on the other hand, it was a patient with bulk, high-LDH disease, rapidly progressive—I think we’d all agree, those patients are less likely to respond to anything, particularly ipilimumab.

My concern is, I was at the chemotherapy symposium a couple of weeks ago, and it’s one of those scenarios where you have these little devices where the doctors render their opinion on questions. And one of the questions was, in a frontline BRAF wild-type patient, you would use what treatment? And they got a bunch of choices. And about 10% or 15% would still go with ipilimumab alone. And this was a relatively indolent patient with good performance status, low LDH. And I said, I don’t think so. I was up at the podium, and I said I think that the CheckMate-064 data would lead me to go with nivolumab or pembrolizumab first, and then hold ipilimumab in reserve. It’s got double the grade 3/4 immune-related adverse event rate. It’s got a 10% to 15% response rate, a clinical benefit rate of maybe 25% versus a solid 50 plus percent clinical benefit rate for nivolumab or pembrolizumab alone. So I’m hard-pressed to use ipilimumab in frontline. I think it’s generally going to be a very good second-line drug, especially if it’s in combination with drug X, Y, or Z.

Transcript Edited for Clarity