PracticeUpdate: There were some interesting preliminary data presented on pembrolizumab in locally advanced head and neck cancer at this year's ASCO. Could you tell us more about this study?
Dr. Burtness: Trisha Wise-Draper from University of Cincinnati presented a small multicenter trial, that looked at pathologic response and disease-free survival among patients with locally advanced head neck cancer, who received neoadjuvant pembrolizumab, and then received pembrolizumab as part of their post-operative therapy. So this was a single cohort trial, that is to say there was no randomization, but it mimics a little bit the design of the KEYNOTE-689 trial, which is currently ongoing. So it's extremely interesting to be able to see these data.
PracticeUpdate: What was the patient population included in this trial?
Dr. Burtness: It was a very advanced patient population, so a lot of T3, T4 disease, a lot of N2a through N3 disease and patients received pembrolizumab. Then they went to surgery and then they were offered post-operative therapy that was risk adjusted based on the findings at surgery.
So, if they had high-risk disease, positive margins, extranodal extension, they were assigned to a treatment where they received radiation, cisplatin, and pembrolizumab. And if they met the intermediate-risk definition, which is to say that they did not have any or a positive margin, they received radiation with pembrolizumab alone.
PracticeUpdate: What was the specific issue being evaluated in these patients?
Dr. Burtness: The hypothesis was that for radiation-cisplatin in the high-risk population, the one-year disease-free survival ought to be about 65 to 67%. And in the favorable risk group, getting radiation alone, it should be about 69 or 70%. And they were interested in seeing whether or not the incorporation of neoadjuvant pembrolizumab would change that.
And then, they also did a very elegant analysis of who was most likely to respond to the neoadjuvant pembrolizumab by using NanoString and PD-L1 quantitation. And they used a response definition that incorporated the immune response pathologic criteria that had been developed by Dr. Uppaluri in the pilot experience with neoadjuvant pembrolizumab.
So they found a treatment effect in about 40% of the patients. And there were 39 patients who ended up being assigned to the intermediate-risk group. They saw outcome in the high-risk group that was comparable to the historical controls that they had taken from the NRG trials. But in the favorable-risk group, the use of neoadjuvant pembrolizumab, followed by surgery, followed by radiation with pembrolizumab, led to very, very remarkable one-year disease-free survival of over 95% and an overall survival of 100%.
I think these are obviously stunning data. I think the complication in understanding them is that the outcome really may be a reflection of what biology is it that responds so well to pembrolizumab. Are these tumors that are less hypoxic, that are smaller, that have less immune exclusion?
And so, to change the post-operative therapy on the basis of response to neoadjuvant therapy is obviously extremely attractive and a pragmatic patient-centered view of things, because this is a nice thing, if it allows patients to avoid cisplatin and have a high certainty of 1-year survival.
The complication is, what are you really learning about what pembrolizumab is contributing to the overall post-operative therapy? Is that result really comparable to patients who have intermediate-risk disease, who are not previously selected by response to IO?
PracticeUpdate: Would you consider changing your practice based on these data?
Dr. Burtness: So, it's not practice-changing. It certainly provides a lot of enthusiasm for continued accrual to the KEYNOTE-689 trial. The correlative studies, as I mentioned, really were very elegant. There was a clear evidence that having PD-L1 expression with a CPS of 20 or higher was very predictive of a response to neoadjuvant pembrolizumab, as determined by these immune pathologic response criteria.
And they also developed a NanoString signature that incorporated IDO PD-L1 that seemed to be predictive of response. There were interesting findings as well about changes between baseline and post-treatment tissue. And so, an extremely valuable study, very exciting. Provides a lot of support for the ongoing phase 3.
I think given the complexities of the shift in who might be in that patient population relative to their historical control, I think it's complicated to understand from a single cohort study like this how it might impact practice.
One of the interesting questions that can be posed when you look at this trial design is what is the advantage of giving neoadjuvant pembrolizumab in a group of patients where you know that you're going to be taking the patient to surgery.
And I think we're still figuring it out, but one of the things is this suggestion, both from Dr. Uppaluri's work from the nivolumab study conducted by Dr. Ferris and now from today's presentation by Dr. Wise-Draper, that the tumors that had previously been unperturbed by radiation or surgery in head neck cancer appear to be a little bit more sensitive to immune checkpoint inhibition. So that is one, strong rationale to explore it very early on, is that you may have a larger proportion of patients who benefit from it.
If we had the ability to understand upfront in whom this treatment could spare the morbidity of our conventional definitive therapy, that is to say, allow you to do a less extensive operation, increase your confidence in negative margins, increase the size of the population that goes forward to less intensive post-operative therapy, that obviously also would be very appealing. And I think the randomized design in the ongoing phase 3 trial will be very helpful there.
Further consideration is whether sensitivity to chemotherapy and radiation in the post-operative setting could be enhanced by pre-exposure to immunotherapy. There's certainly been a description in head neck cancer in the advanced disease, in the metastatic recurrence setting, that patients who've been chemotherapy refractory or re-exposed to chemotherapy after an immune checkpoint inhibitor may be more sensitive.
So, I think we have just so much more to learn about how to employ these agents in the curative setting, but the data from this presentation certainly increased our enthusiasm for studying it in patients who are going to surgery.