Ezra Cohen, MD: That’s pembrolizumab, and you took us through the data there. What about nivolumab, the other agent that’s approved? Barbara, take us through the data for nivolumab.

Barbara A. Burtness, MD: The advantage with nivolumab is that it is actually randomized trial data. And so, this was a trial for, as I said, platinum-refractory patients with the 2 definitions. There were 3261 patients randomly assigned to nivolumab or to an investigator’s choice, which could be docetaxel, cetuximab, or methotrexate. That was obviously defined a little bit by how heavily pretreated the patients were, and you could imagine that there would be big differences between the response rate in, say, docetaxel or methotrexate in a patient who had or hadn’t had a lot of taxane before. But what it showed was an improvement in overall survival for nivolumab compared with the investigator’s choice standard, going from 5.1 to 7.5 months. I think the control arm—5 months for platinum-refractory patients—is squarely in the middle of other control arms that have been reported over the years as people have looked for new agents in this setting. And then, it was 7.5 months with the use of nivolumab.

One of the most striking things about this paper was that there was no difference in progression-free survival. So, it was 2 months for nivolumab and 2.3 months for the investigator’s choice of chemotherapy. I think this highlights the point that Jared was making earlier, that this is a dramatic effect for a minority of the patients. There were responses both to investigator’s choice chemotherapy or to nivolumab. It was higher for nivolumab, and the overall toxicity profile seemed more manageable with nivolumab.

Ezra Cohen, MD: Perhaps that’s the key with all these drugs—as you said, Barbara—that there does seem to be a group of patients who have this dramatic benefit that may extend into not only months, but also to years. We know now that we have really 2 populations in head and neck cancer: HPV-positive and HPV-negative. Are we seeing differences in those groups with respect to immunotherapy?

Jared Weiss, MD: I’m rather underimpressed. There is some conversation, depending on which data set you look at. There are either hints that perhaps the HPV-positive patient might have a little bit more of a response rate, a little bit more efficacy, but I’m pretty unconvinced that this treatment really is at all HPV-related. The biology may be different. The rationale for response may be different, but in terms of pure clinical efficacy, I’m underwhelmed that the drug actually has differential benefit for one subgroup or the other.

Barbara A. Burtness, MD: Although, within the pembrolizumab trials, the median survival is better for the HPV-positive patients.

Jared Weiss, MD: I take that, though, as prognostic more than predictive.

Barbara A. Burtness, MD: It may be that the group of patients who are going to fail really quickly is smaller.

Joshua M. Bauml, MD: But I think, if you look at KEYNOTE-055 at least, these are heavily pretreated patients. Most of them were p16-negative, because that’s who these patients are who tend to fail and have all of those issues. I think that, Jared, you touched on the differential biology, and it’s important to remember we’re using these agents. They’re really exciting. They’re well tolerated. We have this minority of patients who benefit, but we really don’t have a good way of knowing who that group of patients are. And I think that underscores the importance of the development of biomarkers of how to identify who is going to benefit. Because we can parse it out by these rather crude measures—HPV, we can look at smoking, we can look at other factors—but, really, there’s likely underlying biology happening within the tumor that can more effectively predict who will benefit from these agents.

And I think the importance of that goes beyond just being able to tell a patient, “Oh, you’ll benefit from this drug.” If we think about immunotherapies, we will often give the patient the drug. If the tumor grows and they’re tolerating it, we may say, “Oh, sometimes the tumor gets bigger before it gets smaller,” which has not been reported very much in head and neck cancer at all. But we’re oncologists, we’re optimists, so that’s what we do. What we have, then, is a patient who’s being treated for 9, 12, or 15 weeks on a therapy. They have very limited life expectancy on a therapy that has minimal efficacy. So, it’s really important that we identify those patients prospectively.

Barbara A. Burtness, MD: I think that that point about pseudoprogression is important to underline. Clearly, pseudoprogression has been described with immunotherapy agents, and it’s probably in the 6% to 10% range for a cancer like melanoma. But if you go through the spider plots of the publications with pembrolizumab and nivolumab, there are 1 or 2 patients in each trial. It’s going to be less than half a percent. And if you have a symptomatic patient who has gotten a couple of doses of one of these agents, our approach is: if they’re not feeling better and there’s evidence of progression, we take them off. Now we’re hearing all of these anecdotes—and I think I’ve seen it, too—of these really remarkable chemotherapy responses in patients who have been exposed to but did not respond to immunotherapy. I’m sure we’ll be talking about trial designs that might get at, how do you exploit that?

Jared Weiss, MD: That point is really important for clinical practice, just to underscore it. The average patient who appears to be progressing on pembrolizumab or nivolumab should be called a progressor and should be moved on to something else. Any time that something different is done, there should be a clear-cut, simple explanation that you can explain to your grandmother why you did it: the patient feels much better; the cancer is near nothing scary, it’s not near a major vessel or major airway, it’s not threatening how the patient feels, or it’s not growing rapidly.

Barbara A. Burtness, MD: Or this one big mass got smaller and there are a couple of small new lesions, which is the classic.

Jared Weiss, MD: But there should be a story. It shouldn’t be the generic approach. It should be the exception, far-and-away. We reported this early on in KEYNOTE-012A and then almost never saw it again after that. It’s exquisitely rare with PD-1.

Ezra Cohen, MD: Most patients experiencing progression with head and neck cancer, on anti—PD-1 therapy, are truly progressing.

Joshua M. Bauml, MD: Any patient who’s symptomatic and progressing, I think that’s within the definition of pseudoprogression. If they’re symptomatic as a result of their progression, we should stop. And regarding the thought that, “Oh, we’re withholding a benefit from this patient,” I think that the data from melanoma at ASCO this year are very reassuring. You gave these patients 2 years of therapy, you stopped it, and up to 90% of them retained their benefit. So, if you’re giving the drug for 2 doses, then you stop and all their disease melts away, OK. You’re not depriving them of a potential benefit, in all likelihood, by holding your horses and seeing what’s next.

Transcript Edited for Clarity