The Final Results of KEYNOTE-361 - Ajjai Alva

October 26, 2020

Immune checkpoint blockade therapy by blocking PD-1/PD-L1 axis signaling is approved in advanced urothelial cancers in a few settings: (1) first-line therapy for platinum-ineligible patients, (2) first-line therapy in cisplatin-ineligible patients with high PD-L1 expression, (3) maintenance therapy for patients who do not progress with first-line chemotherapy, and (4) as subsequent therapy for patients progressing after cisplatin-based first-line therapy. Regulatory approvals for settings #1 and #2 are based on single-arm phase 2 data, with setting #3 based on the JAVELIN-100 Bladder randomized control trial. To explore the efficacy of combining anti-PD-1 therapy with chemotherapy in platinum-eligible patients as first-line therapy, KEYNOTE-361 was designed as a randomized trial in which pembrolizumab maintenance therapy was given up to two years. Dual primary endpoints were progression-free and overall survival. Alicia Morgans, MD, MPH, and Ajjai Alva, MBBS, MD  discuss the final results from the KEYNOTE-361 trial which he presented the final results from at the 2020 European Society for Medical Oncology Virtual Congress.  


Ajjai S. Alva, MBBS, MD Associate Professor of Medicine and a Medical Oncologist at the University of Michigan Medicine Urology Oncology Clinic, Rogel Cancer Center, Ann Arbor, MI

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 at Northwestern University, where I'm also an Associate Professor of Medicine. I am so excited to have here with me today, Dr. Ajjai Alva, who is a GU medical oncologist and Associate Professor of Medicine at the University of Michigan. Thank you so much for being here with me today, Dr. Alva.

Ajjai Alva: It's a real pleasure to be here, Alicia. Thanks for having me.

Alicia Morgans: Wonderful. I wanted to talk with you a little bit about KEYNOTE-361 in the presentation and abstract that was recently presented. Can you give us a little bit of a background on why this study was important and then tell us a little bit about the study design?

Ajjai Alva: Yeah, of course. The study was a Phase III open-label randomized three-arm trial, global, of comparing standard of care platinum-based doublet chemotherapy versus chemotherapy plus pembrolizumab, which is as you know, an anti-PD-1 antibody, versus the third arm of pembrolizumab alone. This was done in first-line metastatic urothelial cancer, both upper and lower tract. The study background is that pembrolizumab is already being tested and approved in second-line therapy in urothelial cancer in KEYNOTE-052 and also has a label, FDA label in cisplatin-ineligible PD-L1 expressing urothelial cancer in the first-line setting as well as in those considered platinum ineligible. So, therefore, there was evidence before the study was done that there was an activity for the drug in this disease. We wanted to see if the addition of this agent to the upfront chemotherapy frontline would yield benefits to patients as well.

Alicia Morgans: I think at the time that this was designed, that was certainly the big question that we had in mind. And importantly, there was a chemotherapy alone arm and then a pembrolizumab alone arm just in case one of these, especially the pembrolizumab alone arm I think, was it of interest that perhaps it could be enough on its own? So, it seems that the primary endpoint... Was it progression-free survival or overall survival, what were you looking for?

Ajjai Alva: Yeah, so there were dual primary endpoints comparing the chemotherapy alone arm with the chemotherapy plus pembrolizumab arm. These two arms were the primary competitors for the dual primary endpoints of PFS, progression-free survival per RECIST 1.1 as determined by a blinded independent central review and the other co-primary endpoint was overall survival. The study included patients with either upper tract or lower tract urothelial cancer of the renal pelvis, ureter, bladder, or urethra. Those who had no prior systemic therapy for advanced disease. ECOG performance status range from zero, one or two, given that carboplatin also was allowed in the study. And there was, of course, a requirement for RECIST measurable disease of ineligible patients and a tissue sample requirement for PD-L1 assessment and other biomarker assessments. The stratification factors were charged for the platinum agent research by the treating investigator cisplatin versus carboplatin, as we know in urothelial cancer, that is an important first step and also, of course, the combined Foster score, which is the score used by mark for tumors of CPS 10 or greater or CPS less than 10, as stratifying factors as well. These two were the stratifying factors.

Alicia Morgans: Excellent. What did you end up finding?

Ajjai Alva: The study enrolled 1,010 patients between October 2016 and June 2018, among the three arms. You may remember that in the pembrolizumab-only arm, after 80% or so had been enrolled, the FDA came out with a guidance suggesting that pembrolizumab monotherapy should be reserved for PD-L1 high expressing tumors in-patients only, and therefore the final numbers in that pembrolizumab arm were slightly fewer than in the other two arms. But regardless, for the most part, this study was well balanced among the three arms in terms of all the prognostic factors, including visceral disease and lymph node metastasis, the distribution of cisplatin, and carboplatin as well. Overall, and this is an important point when we are trying to compare between Phase III trials, the cisplatin doublet was used in about 45% of the patients across the arms. And this varies a little bit from comparable studies as you know, this is one of three studies where monotherapy of immune checkpoint inhibitors was used in first-line urothelial cancer, the others being DANUBE and IMvigor.

They differ somewhat in these cisplatin versus carboplatin proportions, as well as some other biomarkers used to define high PD-L1 expression. Other than that the study, as in a large randomized study, was well defined and our findings were, for the dual primary endpoint, the PFS by blinded central review was 8.3 months median PFS in the pembrolizumab and chemotherapy combination compared to 7.1 months in the chemotherapy alone. This works out to a hazard ratio of 0.78, the p-value at first glance looks significant because 0.0033, however, there were two interim analyses that had been done before this final analysis, which was presented and therefore the alpha had failed at one of the primary interim analyses and therefore the mark we had to have, was much more stringent, which this primary endpoint did not meet. So, even though there was a slight increase in the PFS, it did not meet pre-defined statistical significance boundaries.

The other primary endpoint was the overall survival and here again, the pembrolizumab plus chemotherapy combination had a median overall survival of 17 months, which is as of now is better than historical numbers and in this study, slightly better than the chemotherapy only at 14.3 months. Again, a hazard ratio of 0.86, but the boundary crossed, the upper confidence interval crossed one. The p-value was 0.0407, again, did not meet the predefined endpoints, and therefore on both these primary endpoints, the study failed to meet its primary endpoint. So, as you said earlier, this was a negative trial and will not change the standard of care, unfortunately.

Alicia Morgans: But I do think that it is important for us to really dig into this because I think we can still learn, and there were many, many patients who participated, so certainly a lot to think about. How do you think this study, where we had a combination of chemotherapy and pembrolizumab or a checkpoint inhibitor, how do you think this approach really squares with the approach that we saw in JAVELIN 100, for example, where it was more of switch maintenance and actually we were selecting responders, right? So we were selecting people with stable disease off the get-go, so what are your thoughts?

Ajjai Alva: Yeah, that's a very good point. The take-home to me are a few; number one, adding cytotoxic chemotherapy and the anti-PD-1, in this case, our immune checkpoint inhibitors, is a different situation in urothelial cancer than in say non-small cell lung cancer, where triplet is proven to be superior and is now FDA approved. In urothelial cancer, I think the addition does provide a little benefit but the benefit is not big enough to be borne out on a Phase III trial. So if anything, we should not, of course, change the standard of care. On the other hand, the switch maintenance approach from the JAVELIN 100 Bladder seems to be the right step. However, as you pointed out, it takes those who get either stable disease or better responses. It completely ignores the people who progress as the best response to the chemotherapy.

So in a sense, it is selecting out the best performers and best prognostic patients. It is not a final solution, but obviously moving immunotherapy to the switch maintenance phase from the second-line phase is a step in the right direction. I think one of the other lessons from KEYNOTE-361 is those with visceral disease, especially liver mets, they do not seem to benefit either from cisplatin or platinum-based chemotherapy or from immune checkpoint inhibitors. So that's an unmet need where there is hope perhaps with antibody-drug conjugates that are coming out and the initial data on those trials are promising for these patients. There is some preclinical work providing a mechanistic rationale that liver mets specifically act as a sync for activated T lymphocytes and therefore decrease the efficacy of immunotherapy. That needs to be borne out, but there are two papers, one in Natural Medicine and one in Translational which seems to explain this finding we see on this and other studies. So this position on immunotherapy for now, as the switch maintenance, is the best place for it.

Alicia Morgans: I agree and I think we are looking forward certainly, as you said, to antibody-drug conjugates and those combinations of EV plus pembro, for example, that I'm really hopeful for, because even if the liver is in sync, perhaps EV can help clear the sync out and allow those T-cells to continue to be active or at least expose more new antigens and allow a more robust immune response, however it happens. I think that that is really exciting as you say. The other thing that you also mentioned that we have known actually for some time now, is that this study has demonstrated to us that pembrolizumab upfront as a single agent in patients who are chemotherapy eligible, is probably not the right way to go. I just want to know your thoughts on that, I mean, it was nice to learn that otherwise, we could be making mistakes over time without really even having the data. What are your thoughts?

Ajjai Alva: That is again a very important topic of clinical interest right now and discussion and I can understand why there could be two thoughts about it. For example, one way would be to start off even with carboplatin doublet and go to switch maintenance, but we still have this FDA label currently in high CPS patients with PD-L1 expression patients with pembrolizumab as frontline or atezolizumab, another drug that is approved as the frontline for CPS. In that case, IC-scores expressing. So between these two, there is no head-to-head data, there's plus-minus. One of the things about this trial as well as other immunotherapy trials is that the toxicity profile is vastly different between chemotherapy and immune checkpoint inhibitors. The monotherapy arm in this study as well, bore out that pembrolizumab monotherapy is exceedingly safe and the incidents of immune adverse events are what we know from other studies around 15% grade 3 and higher, and are for the most part manageable with steroids and other immunosuppressants, which we all now are familiar with.

Overall, I would say the favorability for a safety profile would favor a single-agent pembrolizumab in that situation that you post, but in another patient, perhaps where we want cytoreduction quickly, we have one shot at this patient, perhaps starting off with carboplatin doublet and then switch maintenance may be the better way. It really becomes again, individualized and patient-centered, depending on their prognostic factors as well as for example, if there are liver mets, I probably would not choose pembrolizumab alone or any immune checkpoint inhibitors, but if they have lymph node only disease, that would be somewhere where we perhaps can try immunotherapy in the hope that maybe that would be sufficient.

Alicia Morgans: I completely agree and certainly we have all seen upfront responses and we are always grateful when that happens. Thank you for really bringing that to the floor. As I think about this data, as I think about DANUBE and IMvigor, it's interesting to me that the patients who don't respond are going to cross over if they were in the chemotherapy arm to a checkpoint inhibitor, do you happen to know what the crossover rate was in the study?

Ajjai Alva: Yeah. So as you point out, because immune checkpoint inhibitors are widely available now, in our control arm of chemotherapy alone in KEYNOTE-361, 45% received subsequent anti-PD-1 or PD-L1 therapy. That obviously has an impact on the overall survival readout but on the other hand, it is a reality in today's world, so I think we need to factor that into interpreting these results.

Alicia Morgans: I completely agree. So, as you think about the results and I sincerely appreciate you sharing them and really digging into the data with us, do you have a take-home message for listeners?

Ajjai Alva: Yeah. The take-home would be, and what I have taken home is that adding immunotherapy to chemotherapy in the frontline setting directly together is not probably the way to go now. We wouldn't expect that to change without new data suggesting otherwise. There is one more study that has to readout, which is a CheckMate study, it's the 901, but that will not readout for a while. Until then, switch maintenance with the chemotherapy leading off four to six cycles and then going on to avelumab which is the current FDA approval would be the way to go, except for still those patients who we think could perhaps get away with just immunotherapy alone in PD-L1 expressing tumors. Perhaps there is still a role in there, especially for people without liver mets and favorable prognostic factors like lymph node only disease, et cetera.

Alicia Morgans: Great. Well, thank you so much for talking this through, certainly for your efforts on the study and for taking the time today. Thank you so much.

Ajjai Alva: It was a pleasure talking to you, Alicia. Thank you so much.

email news signup