With Perioperative Pembrolizumab, Event-Free Survival for Early-Stage Lung Cancer Was 62% at the Interim Analysis of Phase 3 KEYNOTE-671

Heather A. Wakelee, MD, FASCO, the Professor of Medicine and Chief of the Division of Oncology at Stanford University School of Medicine and the Deputy Director of Stanford Cancer Institute in Stanford, California, and lead study investigator of KEYNOTE-671, joins Pharmacy Times to discuss positive data from KEYNOTE-671, which evaluated perioperative pembrolizumab (KEYTRUDA® ; Merck) with chemotherapy for adults with early-stage non-small cell lung cancer (NSCLC) at the 2023 American Society of Clinical Oncology (ASCO) Annual Meeting, held in Chicago, Illinois from June 2 to 6.

PT Staff: Could you discuss the phase 3 KEYNOTE-671 trial for early-stage NSCLC?

Heather A. Wakelee, MD, FASCO: KEYNOTE-671 is a perioperative trial that looked at giving neoadjuvant (presurgical) chemotherapy plus either pembrolizumab or placebo, followed by surgical resection, followed by adjuvant pembrolizumab or placebo for patients with stage 2 and 3 non-small cell lung cancer (NSCLC).

PT Staff: How did treatment impact primary efficacy and safety outcomes (i.e. overall survival, event-free survival, toxicity profile) compare to placebo?

Heather Wakelee, MD, FASCO: KEYNOTE-671 enrolled nearly 800 (797 to be exact) patients with stage 2 and 3 NSCLC to receive neoadjuvant chemotherapy with either pembrolizumab or placebo, followed by surgical resection followed by either the adjuvant pembrolizumab or adjuvant placebo. what we found from the trial was that there was a significantly higher rate of complete pathological response (CPR) and major pathological response (MPR) for the chemotherapy plus pembrolizumab arm. And the really exciting results were with the improvement in event-free survival (EFS). So the EFS hazard ratio was .58 at this first interim analysis, and that was highly statistically significant. The trial was powered with dual primary endpoints of both EFS and overall survival (OS). The OS is still immature but trending in a very positive direction. And this was all achieved with no unexpected toxicity. We did see— in a group of patients undergoing neoadjuvant chemotherapy and surgery— the expected toxicities, with the additional toxicity from the immune checkpoint inhibitor (ICI) pembrolizumab, as expected, and as in keeping with other trials in the perioperative setting using other checkpoint inhibitors. And so with this it looks like a really promising result. Again, that hazard ratio of .58 in favor of the pembrolizumab arm when we looked at EFS.

PT Staff: Did you observe any surprising findings during this trial?

Heather A. Wakelee, MD, FASCO: KEYNOTE-671 was positive which in and of itself not that surprising, considering that we've had 2 positive adjuvant trials; 1 with the Tezepelumab (Tezspire; Amgen, AstraZeneca) and 1 with pembrolizumab, and then now 1 very positive neoadjuvant trial, the CheckMate 816, with nivolumab. And very recently, 2 other perioperative trials that were positive.

So it was not surprising that KEYNOTE-671 was positive. What was really striking I think, was the depth of that EFS, the curve started to separate by 4 months and it stayed separated. And by 24 months—we had follow-up of around 25 months at this interim analysis—there was a 62% percent EFS rate for the pembrolizumab arm versus only 41% For placebo. And again, that hazard ratio is .58 and the curves continue to separate with time. We're also seeing with the OS. Again, it's too early to say anything definitively, but the curves are starting to separate even before the 24-month mark and are continuing to separate beyond that. We also see in the subgroups that every subgroup had benefit. When you look at the programmed death-ligand 1 (PD-L1) level, the group with the highest PD-L1, greater than 50%, the EFS hazard ratio was .42. But even in those with no PD-L1 expression, the EFS hazard ratio was .77. That didn't quite meet statistical significance, but it is looking very promising. And putting all the that together, it's a very positive study.

PT Staff: What is the value of pembrolizumab used as both neoadjuvant and adjuvant treatment in a perioperative treatment regimen? How does it compare to previous treatment regimens?

Heather A. Wakelee, MD, FASCO: When we think about using the checkpoint inhibitors in the setting of patients with earlier-stage lung cancer, we obviously have 2 options: we can give it before neoadjuvant or can give it after adjuvant. Or we can combine them all. And that's this perioperative approach. And what if we look at the data that exists with the pure neoadjuvant trial, the CheckMate 816? [It was a] positive study, and yet many patients still potentially having recurrence. When we look at the pure adjuvant trials, they were also positive.

So how do we tell what the perioperative adds by doing both together? We don't have a clear way to show that. But what we do have is an exploratory analysis. When we look at the patients who either had a MPR or CPR, that implies that the patients had benefit from that neoadjuvant chemotherapy plus pembrolizumab. However, we see that patients who either did or did not have that MPR or CPR. When we then look at the separation of curves and whether or not there was benefit for overall, we see that clear benefit. So to explain that a little bit better.

When we separate the groups of patients who either did or did not have a MPR, the group that had a MPR (regardless of whether they got pembrolizumab or not) looked better than the group that did not. But when you look in those separate curves, you see that some patients who received pembrolizumab did not have MPR; you're not sure if they got any benefit from the new adjuvant approach. However, there's still a pretty robust separation of the EFS curves, and that hazard ratio is .73. So among the patients who did not clearly have any benefit from neoadjuvant (because they didn't have MPR), we still see a very robust EFS benefit in favor of the pembrolizumab. So that argues for a potential benefit (additional benefit) of the adjuvant treatment. While we can't say that for sure from this trial, we have hints of that. And therefore, I think to me that argues that the perioperative approach is probably better than either 1 purely alone. This, again, is for stage 2 and 3. I think we've got a lot of room for further discussions around the earlier stages, but this trial was just for stages 2 and 3A and we see this hint that maybe that additional extra treatment is actually beneficial.

I think what's also striking is when you look at that EFS hazard ratio, .58, it's better than what we've seen in some of the other trials, but we have to be careful. These are different patient populations and you don't want to attempt to do an apples to apples comparison across trials. But that is robust. And to me what's striking is that we really do see benefit in all of the subgroups in a way that makes sense. So there's more benefit with higher programmed death-ligand 1 (PD-L1), but there's still benefit even without. And when we look at the stages, there's more benefit for stage 3 than stage 2, but there's still benefit. And so looking at the subsets, it all sort of makes sense as to what we've seen from other trials and what we would anticipate to see.

PT Staff: What further research is needed? What are other possible clinical implications/applications of pembrolizumab for early-stage NSCLC?

Heather A. Wakelee, MD, FASCO: I think when we are talking about patients who are potentially curable, we want to try to cure every patient. And though we have great results from this trial, we still have a lot of patients who are having recurrence or dying if their disease eventually. So we’ve solve a lot [but] we still have room for improvement in trying to figure out how do we help every patient in a way where we can cure them if their cancer is found when it's in early stages.

We also want to be able to cure people without over-treating them. And so the remaining questions we have are how do we bring this benefit to even more people? So how do we identify the patients who are truly benefiting? And for those who aren’t, how do we figure out who those people are? And figure out other approaches that are going to help them? And then also, how do we figure out the patients who are cured with surgery alone, or really just need that surgery? That's going to be more often the stage 1/2. But there are plenty of those patients who still do recur.

So how do we figure out who's who? I think the promise of a future are technologies like circulating tumor DNA (ctDNA), where you think [that] “Maybe this is someone who is going to be cured with surgery alone. Maybe they have a higher than whatever percentage chance of that happening.” Those are people who maybe don't need that preoperative approach and can go straight to surgery. And then we can do additional testing to figure out well, who needs additional therapy?

This study doesn't give us that information. And that isn't necessarily unique to pembrolizumab, but that's something we need to be thinking about, though. We don't have those assays at that level of sensitivity yet on a commercial scale. Then the other questions (figuring out if patients who have had a really robust benefit with a neoadjuvant approach) are how many of them need that additional adjuvant treatment? From this trial, we see that even for patients who did have a PCR, the ones who were on the pembrolizumab arm still did better. And so perhaps there is some benefit from that additional treatment. Though, you can also interpret that in different ways. So trying to sort that out is important. And then we in this trial included a small percentage of patients who had known driver mutations either in estimated glomerular filtration rate (eGFR) or anaplastic lymphoma kinase (ALK) in their tumors. However, the numbers were too small to say anything definitively about whether they helped or [did] not help. And that's a problem that we've had in the other trials that have looked at checkpoint inhibitors. So we don't know “What is the best thing to do for patients who have specific driver mutations, and how does it mean therapy work there?” So, these are some of the big questions that are remaining.

PT A. Staff: Is there anything you would like to emphasize about this treatment regimen?

Heather Wakelee, MD, FASCO: Just to add that the KEYNOTE-671 trial is particularly exciting because it's combining the neoadjuvant plus the adjuvant approach with very positive robust outcomes with that EFS hazard ratio of .58 and we also are not seeing unexpected toxicity. So that's what's I think particularly exciting about this trial.