Long-Term Follow-Up From the KEYNOTE-052 Trial, Pembrolizumab in the First-Line Setting in Cisplatin-Ineligible Patients With Advanced Urothelial Cancer – Arjun Balar
June 8, 2021
In this conversation with Alicia Morgans, MD, MPH, Arjun Balar, MD discussed the updated results of the KEYNOTE-052 trial up to 5 years of follow-up. KEYNOTE-052 was a multicentre, single-arm, phase 2 study that launched several years ago tested pembrolizumab in the first-line setting in cisplatin-ineligible patients with advanced urothelial cancer. The goal of KEYNOTE-052 was to evaluate the role of pembrolizumab as monotherapy in this group of patients. Dr. Balar highlights that first-line pembrolizumab monotherapy continued to show durable antitumor activity up to 5 years after the last patient was enrolled, with an objective response rate of 28.9%, median duration of response of 33.4 months, and median overall survival of 11.3 months. Pembrolizumab was approved for cisplatin-ineligible patients with untreated advanced urothelial carcinoma based on the initial results of the phase 2 KEYNOTE-052 study.
Biographies:
Arjun V. Balar, MD, Associate Professor, Department of Medicine at NYU Grossman School of Medicine, Medical Director, Clinical Trials Office, Perlmutter Cancer Center, and Director of the genitourinary medical oncology program at NYU Langone’s Perlmutter Cancer Center. NYU Langone Health, New York, New York.
Alicia Morgans, MD, MPH Associate Professor of Medicine in the Division of Hematology/Oncology at the Northwestern University Feinberg School of Medicine in Chicago, Illinois.
Biographies:
Arjun V. Balar, MD, Associate Professor, Department of Medicine at NYU Grossman School of Medicine, Medical Director, Clinical Trials Office, Perlmutter Cancer Center, and Director of the genitourinary medical oncology program at NYU Langone’s Perlmutter Cancer Center. NYU Langone Health, New York, New York.
Alicia Morgans, MD, MPH Associate Professor of Medicine in the Division of Hematology/Oncology at the Northwestern University Feinberg School of Medicine in Chicago, Illinois.
Read the Full Video Transcript
Alicia Morgans: Hi, my name is Alicia Morgans and I'm a GU Medical Oncologist and Associate Professor of Medicine at Northwestern University in Chicago. I am so excited to have here with me today, a good friend and colleague Dr. Arjun Balar, who is the Director of the GU Medical Oncology Program at NYU Langone Perlmutter Cancer Center in New York. Thank you so much for being here with me today, Dr. Balar.
Arjun Balar: Thanks so much for having me.
Alicia Morgans: Wonderful. Well, Arjun I just wanted to talk with you a little bit about the long-term follow-up for the KEYNOTE-052 trial. Can you tell us a little bit about that study? And then we can explore some of the findings.
Arjun Balar: Absolutely. So KEYNOTE-052 was a single-arm phase two study that launched several years ago that tested pembrolizumab in the first-line setting in cisplatin-ineligible patients with advanced urothelial cancer. And to kind of give a context again, this was in an era where again, the accelerated approval was the aim, 370 patients were enrolled in the study. And the goal was really to evaluate, what is the role of pembro as monotherapy in this group of patients? The long-term data from this trial is now being presented at ASCO, a lot of maturities, over five years of follow-up from this trial, and really what we are seeing in this study is an objective response rate that hasn't changed at 29%, but really, what the real value in this study is looking at the duration of response in particular, in patients with high CPS scores versus the intent to treat analysis, looking at survival and in particular, the tail of the curve in terms of long-term survival. Looking at landmarks, 36 months, 48 months, even long-term, to really assess what is the tail of the curve in patients receiving first-line immunotherapy.
Alicia Morgans: So why do you think the tail of the curve is so important? And as a clinician, this is really important to me, but if you can just speak to that a little bit. Because we don't always see, a lot of our therapies do not have a tail. And we see a very kind of, I mean, they have a minimal tail, but with checkpoint inhibitors, we can actually see quite a durable tail, or response over time. This is really meaningful, I think, to patients and to clinicians. How do you think about it?
Arjun Balar: Yeah, I think that is the fundamental difference between, traditionally our cancer therapies, cytotoxics for instance. So whether it was molecularly targeted therapies or cytotoxic chemotherapies, which are basically the treatments we have used the longest in cancer, is you could achieve disease control, but because the treatment itself doesn't change, resistance is essentially expected, and eventually patients progress. So there really is no real tail to the curve with obviously the exception of cisplatin-based treatment in some patients that you might achieve a cure.
However, with immunotherapy, the target of treatment is actually the host, the host elicits the treatment response. And the only thing that exists in nature that is actually able to keep up with cancer evolution is the immune system. And so you have these tails to the curve, which is not 10% to 15%, but it's actually the patients who actually achieve a response. And so it could easily be 20% or more depending on the upfront response rate. So that is what we really need to look at. Now, obviously, there is a lot of controversies. Is immunotherapy or chemotherapy upfront best? We won't argue that right now. Right. But what is not debated is that, if you are in a response to immunotherapy, there is true durability to this. And that's what the value of this trial is, is looking at that durability.
Alicia Morgans: Absolutely. And that is part of the value of these long-term follow-ups as the one that you and the team are presenting at ASCO 2021 this year. So can you tell us a little bit about the findings in your more correlative type analysis and your subgroup analysis?
Arjun Balar: Yeah, so essentially there were a couple of key findings from this study. Number one, patients with high CPS scores had a higher likelihood of response. So CPS 10 or greater using the 22C3 assay had a 47% response rate as compared to 29% in the intent to treat. Now, the problem was, that the same correlation with response was not duplicated in KEYNOTE-361. I think that is a shortcoming of PD-L1 expression. And again, I don't have a good answer for it, and I don't think anybody does. I think that it is just, PD-L1 expression is just not a good biomarker. I think we will just kind of have to leave it at that for the moment. But regardless, if patients are in response, there is clearly durability to it. So when we looked at a kind of long-term follow-up, the median duration of response in all comers, so all patients who achieved a response was 33.4 months, so basically just short of three years, that median ratio response was just short of three years. That's incredible. You don't really see that with cytotoxic therapies.
In patients who are CPS 10 or greater, the median duration of response has still not yet been reached. And then when you look at landmark survival rates, and this is the data that is going to be presented, it's about 19% of patients are alive at 48 months and the median follow-up for longer. Now, again, this is data that you would not expect with cytotoxic-based therapy. And if you look at the CPS 10 or greater, and you look at the 48 months, the landmark is nearly 32%. So that's the maturity that you see with the first-line pembrolizumab. Just as an aside, we're going to be presenting five-year and longer follow-up data for the first-line atezo. Now granted, that's a 119 patient study, much smaller, about a third the size. But we are going to be presenting that at ESMO, very similar data, by the way. Again, demonstrating that again, the role of first-line immunotherapy for patients who achieve and respond, the long-term outcomes are actually quite favorable.
Alicia Morgans: So I think, we just have to think about which population this is, of course, and really emphasize that as well. These are cisplatin-ineligible patients. And I'd love to hear your comments based on the findings in this study and your thoughts on PD-L1 as a biomarker. Agreed it has numerous challenges, but for cisplatin-ineligible patients who are carbo eligible at this point in time, our label still says that we really need to identify that CPS high patient population if we want to use pembrolizumab. What are your thoughts on that?
Arjun Balar: So I look at it as rather not so much that CPS high patients should get immunotherapy necessarily, it's really about CPS low patients should not get immunotherapy. That's how I look at it. And that's how the label really is meant to be read. So if you look at the evolution of the label in 2017, it was broad CIS ineligible patients. The restriction on the label just removed the CPS low patients. So all that was left were the CPS highs. So that's really the way to look at it.
The inconsistency in the biomarker reflects just the fact that PD-L1 expression is dynamic. We kind of knew this. So the fact that the same biomarker, same algorithm, same antibody, and so forth, did not perform consistently is the problem. And so we are going to have to use clinical judgment, frankly, to help our decision-making and what I call clinical biomarkers, prognostic models that we use. Ever since 1999, when the Steve Boorjian model was published, frankly, those are the same things we should be using in clinic today.
The issue that you pointed out, carbo eligible patients, should patients necessarily receive carboplatin as a preferred agent rather than immunotherapy. So let's say, for instance, we take the CPS score out of the picture for a minute because I don't think it's valuable in clinical decision-making. But let's say a patient presents to you and is necessarily carbo eligible. Is that necessarily the preferred treatment over immunotherapy? I don't think that is necessarily the right answer either.
Frankly, I think that patients who have small volume disease, don't have visceral metastasis, in particular liver metastasis, these patients can actually be safely treated with immunotherapy. It's really the consequence of guessing wrong. So patients with the small-volume disease, that's not progressing very quickly, if patients achieve a response to immunotherapy, they can actually avoid chemotherapy entirely. It's the question of patients who were randomized to chemo, or let me rephrase another way. Patients who were randomized to immunotherapy, where you were saying, gosh, I wish we had given them chemotherapy, those are the patients who actually did poorly in the randomized trials. And that is just something that can't be reflected in clinical practice. And so I think that is where again, judgment is really the most important factor that we should use in deciding whether patients should get immunotherapy or carboplatin. It's not a one-size-fits-all.
Alicia Morgans: I completely agree. And thank you so much for really pointing out those nuances. Though those are not specifically the topics that you are presenting at ASCO 2021, they are so important for our clinical practice, and really to emphasize that these CPS low patients should not be starting off with immunotherapy, like pembrolizumab, I think is a really, really good way to think about it. And it makes it a much cleaner decision for clinicians who are facing these patients. So final thoughts, a summary on KEYNOTE-052?
Arjun Balar: Long-term follow-up from pembrolizumab in the first-line setting clearly demonstrates that there is a subset who respond and those patients who respond do really, really well long-term.
Alicia Morgans: Well, there you have it. That was short, and sweet, and certainly to the point. So thank you so much for taking the time today, Arjun.
Arjun Balar: My pleasure.
Alicia Morgans: Hi, my name is Alicia Morgans and I'm a GU Medical Oncologist and Associate Professor of Medicine at Northwestern University in Chicago. I am so excited to have here with me today, a good friend and colleague Dr. Arjun Balar, who is the Director of the GU Medical Oncology Program at NYU Langone Perlmutter Cancer Center in New York. Thank you so much for being here with me today, Dr. Balar.
Arjun Balar: Thanks so much for having me.
Alicia Morgans: Wonderful. Well, Arjun I just wanted to talk with you a little bit about the long-term follow-up for the KEYNOTE-052 trial. Can you tell us a little bit about that study? And then we can explore some of the findings.
Arjun Balar: Absolutely. So KEYNOTE-052 was a single-arm phase two study that launched several years ago that tested pembrolizumab in the first-line setting in cisplatin-ineligible patients with advanced urothelial cancer. And to kind of give a context again, this was in an era where again, the accelerated approval was the aim, 370 patients were enrolled in the study. And the goal was really to evaluate, what is the role of pembro as monotherapy in this group of patients? The long-term data from this trial is now being presented at ASCO, a lot of maturities, over five years of follow-up from this trial, and really what we are seeing in this study is an objective response rate that hasn't changed at 29%, but really, what the real value in this study is looking at the duration of response in particular, in patients with high CPS scores versus the intent to treat analysis, looking at survival and in particular, the tail of the curve in terms of long-term survival. Looking at landmarks, 36 months, 48 months, even long-term, to really assess what is the tail of the curve in patients receiving first-line immunotherapy.
Alicia Morgans: So why do you think the tail of the curve is so important? And as a clinician, this is really important to me, but if you can just speak to that a little bit. Because we don't always see, a lot of our therapies do not have a tail. And we see a very kind of, I mean, they have a minimal tail, but with checkpoint inhibitors, we can actually see quite a durable tail, or response over time. This is really meaningful, I think, to patients and to clinicians. How do you think about it?
Arjun Balar: Yeah, I think that is the fundamental difference between, traditionally our cancer therapies, cytotoxics for instance. So whether it was molecularly targeted therapies or cytotoxic chemotherapies, which are basically the treatments we have used the longest in cancer, is you could achieve disease control, but because the treatment itself doesn't change, resistance is essentially expected, and eventually patients progress. So there really is no real tail to the curve with obviously the exception of cisplatin-based treatment in some patients that you might achieve a cure.
However, with immunotherapy, the target of treatment is actually the host, the host elicits the treatment response. And the only thing that exists in nature that is actually able to keep up with cancer evolution is the immune system. And so you have these tails to the curve, which is not 10% to 15%, but it's actually the patients who actually achieve a response. And so it could easily be 20% or more depending on the upfront response rate. So that is what we really need to look at. Now, obviously, there is a lot of controversies. Is immunotherapy or chemotherapy upfront best? We won't argue that right now. Right. But what is not debated is that, if you are in a response to immunotherapy, there is true durability to this. And that's what the value of this trial is, is looking at that durability.
Alicia Morgans: Absolutely. And that is part of the value of these long-term follow-ups as the one that you and the team are presenting at ASCO 2021 this year. So can you tell us a little bit about the findings in your more correlative type analysis and your subgroup analysis?
Arjun Balar: Yeah, so essentially there were a couple of key findings from this study. Number one, patients with high CPS scores had a higher likelihood of response. So CPS 10 or greater using the 22C3 assay had a 47% response rate as compared to 29% in the intent to treat. Now, the problem was, that the same correlation with response was not duplicated in KEYNOTE-361. I think that is a shortcoming of PD-L1 expression. And again, I don't have a good answer for it, and I don't think anybody does. I think that it is just, PD-L1 expression is just not a good biomarker. I think we will just kind of have to leave it at that for the moment. But regardless, if patients are in response, there is clearly durability to it. So when we looked at a kind of long-term follow-up, the median duration of response in all comers, so all patients who achieved a response was 33.4 months, so basically just short of three years, that median ratio response was just short of three years. That's incredible. You don't really see that with cytotoxic therapies.
In patients who are CPS 10 or greater, the median duration of response has still not yet been reached. And then when you look at landmark survival rates, and this is the data that is going to be presented, it's about 19% of patients are alive at 48 months and the median follow-up for longer. Now, again, this is data that you would not expect with cytotoxic-based therapy. And if you look at the CPS 10 or greater, and you look at the 48 months, the landmark is nearly 32%. So that's the maturity that you see with the first-line pembrolizumab. Just as an aside, we're going to be presenting five-year and longer follow-up data for the first-line atezo. Now granted, that's a 119 patient study, much smaller, about a third the size. But we are going to be presenting that at ESMO, very similar data, by the way. Again, demonstrating that again, the role of first-line immunotherapy for patients who achieve and respond, the long-term outcomes are actually quite favorable.
Alicia Morgans: So I think, we just have to think about which population this is, of course, and really emphasize that as well. These are cisplatin-ineligible patients. And I'd love to hear your comments based on the findings in this study and your thoughts on PD-L1 as a biomarker. Agreed it has numerous challenges, but for cisplatin-ineligible patients who are carbo eligible at this point in time, our label still says that we really need to identify that CPS high patient population if we want to use pembrolizumab. What are your thoughts on that?
Arjun Balar: So I look at it as rather not so much that CPS high patients should get immunotherapy necessarily, it's really about CPS low patients should not get immunotherapy. That's how I look at it. And that's how the label really is meant to be read. So if you look at the evolution of the label in 2017, it was broad CIS ineligible patients. The restriction on the label just removed the CPS low patients. So all that was left were the CPS highs. So that's really the way to look at it.
The inconsistency in the biomarker reflects just the fact that PD-L1 expression is dynamic. We kind of knew this. So the fact that the same biomarker, same algorithm, same antibody, and so forth, did not perform consistently is the problem. And so we are going to have to use clinical judgment, frankly, to help our decision-making and what I call clinical biomarkers, prognostic models that we use. Ever since 1999, when the Steve Boorjian model was published, frankly, those are the same things we should be using in clinic today.
The issue that you pointed out, carbo eligible patients, should patients necessarily receive carboplatin as a preferred agent rather than immunotherapy. So let's say, for instance, we take the CPS score out of the picture for a minute because I don't think it's valuable in clinical decision-making. But let's say a patient presents to you and is necessarily carbo eligible. Is that necessarily the preferred treatment over immunotherapy? I don't think that is necessarily the right answer either.
Frankly, I think that patients who have small volume disease, don't have visceral metastasis, in particular liver metastasis, these patients can actually be safely treated with immunotherapy. It's really the consequence of guessing wrong. So patients with the small-volume disease, that's not progressing very quickly, if patients achieve a response to immunotherapy, they can actually avoid chemotherapy entirely. It's the question of patients who were randomized to chemo, or let me rephrase another way. Patients who were randomized to immunotherapy, where you were saying, gosh, I wish we had given them chemotherapy, those are the patients who actually did poorly in the randomized trials. And that is just something that can't be reflected in clinical practice. And so I think that is where again, judgment is really the most important factor that we should use in deciding whether patients should get immunotherapy or carboplatin. It's not a one-size-fits-all.
Alicia Morgans: I completely agree. And thank you so much for really pointing out those nuances. Though those are not specifically the topics that you are presenting at ASCO 2021, they are so important for our clinical practice, and really to emphasize that these CPS low patients should not be starting off with immunotherapy, like pembrolizumab, I think is a really, really good way to think about it. And it makes it a much cleaner decision for clinicians who are facing these patients. So final thoughts, a summary on KEYNOTE-052?
Arjun Balar: Long-term follow-up from pembrolizumab in the first-line setting clearly demonstrates that there is a subset who respond and those patients who respond do really, really well long-term.
Alicia Morgans: Well, there you have it. That was short, and sweet, and certainly to the point. So thank you so much for taking the time today, Arjun.
Arjun Balar: My pleasure.