Insights From: Ezra Cohen, MD, Moores Cancer Center; UC San Diego Health; Jared Weiss, MD, UNC Lineberger Comprehensive Cancer Center
Jared Weiss, MD: Pembrolizumab is now FDA-approved for second-line recurrent metastatic head and neck cancer. The approved dose is 10 mg/kg every 3 weeks. This came as a consequence of the KEYNOTE-012 A and B cohorts, which were subsequently supported by KEYNOTE-055. KEYNOTE-012A was a study of PD-L1-positive head and neck cancer. It was a positive study, with a response rate in the high teens and an impressive PFS, with a tail to the overall survival curve. Next, we asked the question, “What about the PD-L1 unselected patients?” That’s where KEYNOTE-012B came in, and the results were nearly identical.
The take-home message here is that pembrolizumab is a standard FDA-approved option for second-line recurrent metastatic head and neck cancer. In my practice, in my opinion, this is the preferred option for such patients, based on superior efficacy and superior tolerability compared to historic controls with competitor agents. The safety and efficacy signals seen in the pembrolizumab trials are rather comparable to what we have seen in other cancers: a response rate in the high teens, some immune-related adverse events—most commonly endocrine disruption in the thyroid—and a low, but real risk, of pneumonitis. But really, there are no great surprises compared to melanoma, lung, or other data sets.
Different cohorts evaluating different agents, including pembrolizumab, have shown different results for the HPV subgroups. When I look at these data as a whole—KEYNOTE-012A, KEYNOTE-012B, and KEYNOTE-055—the similar comparative data with nivolumab versus chemotherapy, I walk away from a feeling that while, theoretically, there could have been large differences in these 2 groups—the smoking-driven cancers versus the HPV-driven cancers—in practice, the data on this biomarker are not sufficient to use it to exclude patients with either characteristic. In both cases, PD-1 inhibitors appear to be less toxic and more efficacious than the cytotoxic alternatives in the second-line setting.
Ezra Cohen, MD: The KEYNOTE-055 trial was specifically designed to identify the activity in patients with recurrent metastatic head and neck cancer who had been treated with both platinum-based therapy and cetuximab, recognizing that that’s a setting where there truly is an unmet need—or at least there was an unmet need in patients who were refractory to both of those agents. Although we have cytotoxic chemotherapies that we can choose from to utilize in that setting, there clearly is no approved therapy, and there clearly are no randomized data, or even good nonrandomized data, in that cohort of patients. With that in mind, KEYNOTE-055 was designed to enroll those patients with a primary endpoint of overall response rate, additionally looking at progression-free survival, overall survival, and toxicity.
Interestingly, what we saw was consistent with the data that we had seen in other cohorts of squamous cell carcinoma of the head and neck patients. Specifically, we saw a response rate of 18%—very reminiscent of what we had seen in earlier studies with pembrolizumab. We saw a nearly identical toxicity profile, with about a 10% level of grade 3/4 toxicity characterized by immune-related adverse events—fatigue being the most common toxicity occurring in almost half of the patients, but rarely reaching a serious level. And again, at least from early data, there were encouraging median overall survival rates.
The other thing to note, of course, is that many of the responses are durable responses. At least at the time of the publication, almost all of the responses were still ongoing, suggesting that not only was there clear activity in this refractory set of patients, but also that some of those responses, if not most of those responses, may be lasting for a very long time. These are things that we don’t see, typically, with either single-agent cytotoxic chemotherapy or other nonimmune single agents in that setting.
When we think about the population that was enrolled in the KEYNOTE-055 study, we begin to recognize that this is a patient population that has never been previously systematically studied in recurrent metastatic head and neck cancer. That is to say, prior to cetuximab approval, we used to conduct clinical trials of patients who were platinum-refractory. Then, with the cetuximab approval, there really has not been a study that specifically enrolled patients who were both platinum and cetuximab-refractory—so, clearly, a paucity of data in that setting.
KEYNOTE-055 represents, truly, the first trial that specifically enrolled what, in effect, becomes a third-line patient population. We used to believe that those patients would not be responsive to any type of therapy. The response rate would be in the low single digits, and the overall survival would be in the very low single-digit months range. It turns out that, at least with pembrolizumab, that’s not the case. The response rate is 18%, with a very encouraging median overall survival—although those data are still fairly preliminary.
In terms of clinical practice and practical therapy for these patients, for what we’re doing in our clinics on a day-to-day basis now, it’s clear that we have a new standard of care for patients with recurrent metastatic disease in the second- and third-line settings. The standard of care is anti–PD-1 therapy, and it’s made a tremendous impact in how we treat patients. Now, patients who are platinum-refractory should, and do, receive an anti-PD-1 antibody, either pembrolizumab or nivolumab. Patients who are also cetuximab-refractory should be receiving an anti-PD-1 antibody, namely pembrolizumab. And so, we now have options that are driven by the data, by the practice-based evidence, to use anti-PD-1 therapy in patients with recurrent metastatic head and neck cancer.
Jared Weiss, MD: I’ve had a lot of personal experience with pembrolizumab in head and neck cancer because I’ve participated in a number of the trials. And subsequently, I clinically treat a very high volume of metastatic recurrent head and neck cancer in my practice. The experience on and off of the trials has really been very comparable. What I’ve seen in the real world echoes the results of the now published KEYNOTE-012A, KEYNOTE-012B, and KEYNOTE-055 data, where I found these agents to be efficacious. The response rate by RECIST in my practice has been about what it’s been in each of the series—in the high teens—with a clinical benefit of over 50%. Again, this is very consistent with the studies.
Similar to the studies, my average patient really feels quite well on them, with an acceptable to minimal toxicity profile in the real world. We certainly see elevations in the real world. They’re easily manageable with holding the drug, instituting steroids, and, usually, if you react quickly enough and aggressively enough, you are able to restart the drug with continued benefit in patients who were benefitting before. Similar to the trials, I have seen rare, weird side effects. I’ve seen pneumonitis in roughly 1 or maybe 2 of the last 100 patients I’ve treated. And I’ve seen 1 or 2 other weird autoimmune side effects—again, consistent with the trial literature—where the average patient does quite well, but you do have those rare, but severe, immune-related adverse events. Response rates are roughly on par with the trials. I think that that trial experience has translated well into real-world experience.
Ezra Cohen, MD: In our clinic, we’ve used a fair bit of pembrolizumab. What we’ve noticed is exactly what the clinical trials would suggest: not every patient responds, clearly. In fact, unfortunately, right now, the majority of patients do not respond. Having said that, in patients who do respond, the responses are often durable, especially if they’re deep responses. What I mean by that is a large partial response, 80% or 90%, or a complete response. Those patients tend to have very durable responses. Not only that, but the agents tend to be very well tolerated, with minimal side effects. Most patients are saying that they’re able to continue their day-to-day activities, their normal lives, with some of them not even noticing that they’re on an anti-cancer drug.
Right now, in terms of which patients we should be treating with these agents, we have minimal data. What we do have are patients who are platinum-refractory, again, in the second- or third-line setting, who should be treated with an anti-PD-1 antibody if there are no clear contraindications—autoimmune disease, a prior organ transplant, or any factor that may affect the immune system’s ability to respond to these agents.
Beyond that, we don’t really have any parameters to select patients for either pembrolizumab or nivolumab. Specifically, PD-L1 does not seem to be a factor that we should consider, although patients who are PD-L1-positive do have a higher response rate and do likely have a longer, or better, median overall survival. But it’s quite clear that even in patients who are so-called PD-L1-negative, there still is a chance for a response rate that’s higher than what chemotherapy would deliver or, at least, single-agent chemotherapy. When those patients do respond, the durability of response, again, appears to be quite long.
At this point, we don’t have any good biomarkers to suggest which patients should be treated with a PD-1 antibody versus chemotherapy—rather, more clinical parameters, or clinical characteristics, that we can use to select treatment for patients based on prior therapy.
Jared Weiss, MD:We do have progression-free survival data and overall survival data from multiple cohorts of pembrolizumab-treated patients. They do look favorable compared to historic controls. What we don’t have, thus far, are comparative data with that agent. But that should be forthcoming.