Pembrolizumab Alone or Combined With Chemotherapy vs Chemotherapy As First-Line Therapy for Advanced Urothelial Carcinoma in the KEYNOTE-361 Trial Journal Club – Christopher Wallis & Zachary Klaassen

September 27, 2022

Christopher Wallis and Zachary Klaassen discuss a Lancet Oncology publication on the KEYNOTE-361 trial, pembrolizumab alone, or combined with chemotherapy versus chemotherapy as first-line therapy for advanced urothelial carcinoma: A randomized open-label phase three trial. The authors of this trial assessed outcomes with first-line pembrolizumab alone or combined with chemotherapy versus chemotherapy for patients with previously untreated advanced urothelial carcinoma.


Christopher J.D. Wallis, MD, Ph.D., Assistant Professor in the Division of Urology at the University of Toronto.

Zachary Klaassen, MD, MSc, Urologic Oncologist, Assistant Professor Surgery/Urology at the Medical College of Georgia at Augusta University, Georgia Cancer Center

Read the Full Video Transcript

Christopher Wallis: Hello, and thank you for joining us for this UroToday Journal Club. Today, we are discussing the recently published KEYNOTE-361 trial, pembrolizumab alone, or combined with chemotherapy versus chemotherapy as first-line therapy for advanced urothelial carcinoma: A randomized, open-label phase three trial. I'm Chris Wallis, a Fellow in Urological Oncology at Vanderbilt, and with me today is Zach Klaassen, an Assistant Professor in the Division of Urology at the Medical College of Georgia. This is the citation for this recent publication in The Lancet Oncology led by Dr. Powles.

For patients who are eligible, the first-line treatment of advanced urothelial carcinoma is a combination platinum-based chemotherapy, typically gemcitabine, and cisplatin or MVAC. While this approach is the best we have to offer today, disease progression is common and median overall survival is still less than 18 months. Thus, there is a distinct ongoing need to improve our treatment options in this disease space.

So if we look at the NCCN guidelines, we will see that IO therapy features prominently, including avelumab maintenance regimes for patients who get initial chemotherapy as well as initial frontline therapy for patients who are cisplatin-ineligible. More recently, two trials assessed the benefit of anti-PD-L1 agents in the first-line therapy. These are the IMvigor130, and the JAVELIN Bladder 100 trials and these showed some benefit. However, they both relied on an initial chemotherapeutic backbone to which IO therapy was added. The question therefore still remained whether IO approaches alone may be beneficial in first-line therapy. This was assessed in the DANUBE trial, which is quite similar in principle to the KEYNOTE-361 trial we are discussing today. And here patients were randomized in a 1:1:1 fashion to durvalumab alone, durvalumab plus tremelimumab, or standard of care chemotherapy. This trial was unfortunately negative and did not demonstrate benefit in their primary comparisons.

KEYNOTE-361 is somewhat different in that there is a combination IO and chemotherapy arm that was not present in DANUBE. So when we look at this again, our first-line options, we see pembrolizumab appears in the first-line monotherapy option for cisplatin-ineligible patients. When we go to second-line therapy options, the preferred regime is pembrolizumab. Therefore, it would not be surprising that this present study assesses the role of pembrolizumab earlier in the disease space. This is the schema highlighting the design of KEYNOTE-361. This is an open-label, multi-center, phase III, randomized controlled trial that was performed in 21 countries and over 200 medical centers. We will walk through each of the key eligibility and randomization features in the coming slides.

So the study included adult patients with histologically or cytologically confirmed urothelial carcinoma, which could be of the renal pelvis, ureter, bladder, or urethra. They had to have unresectable, locally advanced disease or metastatic disease without prior systemic therapy. And while variants were allowed, the cytology had to be predominantly urothelial. Additionally, they had to have a measurable lesion, per RECIST criteria, ECOG of zero to two, as well as have tissue available for PD-L1 testing. Exclusion criteria included patients who had disease suitable for local therapy, who had previous systemic immunotherapy, CNS metastasis, auto-immune disease, immunodeficiency, immunosuppressive medications, infections that may lead to immunosuppression, and pneumonitis. These were all reasons why an IO therapy would potentially not be feasible.  Notably, you see here at the end, peri-op chemotherapy, where it was completed more than one year prior was allowed assuming that this was in the localized muscle-invasive disease space, allowing for this to still be a first-line therapy trial in the metastatic space.

So again, this was a 1:1:1 randomization to pembrolizumab with chemotherapy, pembrolizumab alone, or chemotherapy alone. This was stratified according to PD-L1 status, as well as the investigator's choice of cisplatin versus carboplatin in the chemotherapy regime. It was an open-label allocation and as of February 2018, patients who had CPS scores of less than ten were limited to receive either the combination of the chemotherapeutic regime as it became clear that pembrolizumab monotherapy was not beneficial in this group.

Pembrolizumab was given in standard dosing and approach with 200 milligrams IV every three weeks, continued until disease progression, intolerable toxicity, the decision to withdraw, or a maximum of 35 cycles. Chemotherapy was given with gemcitabine plus either cisplatin or carboplatin. A biomarker assessment was performed to assess the PD-L1 expression, which is measured using the CPS, defined as the number of PD-L1 staining cells divided by the total number of cells, times 100. Patients were assessed throughout the trial with CT and MRI imaging every nine weeks for the first 54 weeks followed every 12 weeks thereafter. RECIST v1.1 was used to assess treatment response and progression, and blood and urine samples were collected at each dose of the study treatment.

There are two primary endpoints. First, progression-free survival by independent blinded review. This comparison was pembrolizumab plus chemotherapy versus chemotherapy in the whole intention-to-treat population. The secondary primary endpoint was overall survival. This was assessed in two populations. First, the pembrolizumab monotherapy versus chemo comparison for those who are PD-L1 positive, and then pembrolizumab versus chemotherapy among the entire ITT population. Secondary endpoints assessed overall response rate, disease control rate, duration of response, safety and tolerability, landmark-based PFS, and time to deterioration of patient-reported outcomes.

The authors planned to accrue 990 patients in a 1:1:1 ratio with approximately 330 allocated to each arm. This would provide an 80% power for the overall viable comparison of pembrolizumab versus chemotherapy in the PD-L1 population and in the ITT population. Additionally, it would provide a 94% power for the overall survival comparison of pembrolizumab plus chemo versus chemo in the ITT population with an alpha of 0.02, and a presumed hazard ratio of 0.7. Finally, a group would provide a 97% power for the progression-free survival analysis of pembrolizumab plus chemo versus chemo with an alpha of 0.005 and a presumed hazard ratio of 0.68. There were two interim analyses planned, the first after 347 progression-free survival events, and the second after 357 OS events. The final analysis was mandated to be at least 22 months from the last patient accrual and at 616 OS events.

The authors did do alpha splitting using the Bonferroni approach, and then subsequently used a sequential testing strategy which assessed the superiority analysis of both PFS and OS for the pembrolizumab plus chemo versus chemo comparison followed by the non-inferiority and superiority analysis of OS for the pembro versus chemo analysis in the ITT population and in the PD-L1 positive population. If neither of these primary outcomes was met, statistical significance testing was not performed for subsequent endpoints. The authors did use that alpha spending given this approach. And so you can see the division of the alpha here, according to both outcomes and multiple analyses, and so we have the one-sided significance threshold of 0.0019 for PFS and of 0.0142 for OS.

So all efficacy outcomes were assessed in the ITT analysis.  The duration of responses was assessed in patients who had either a complete or partial response, and safety was assessed among all patients who received at least one dose of the study medication. An exploratory sensitivity analysis was done to assess the effect of subsequent post-protocol immunotherapy in the patients who received chemotherapy as their randomized regime.

At this point in time, I am now going to hand it over to Zach to walk us through the results.

Zachary Klaassen: Thanks, Chris.

So this is the trial profile laid out on this slide and there were 1,360 patients screened for eligibility and ultimately 1,010 patients were enrolled and randomly assigned. This included 351 patients assigned to the pembro plus chemo group, 307 assigned to the pembro group, and 352 patients assigned to the chemotherapy group. Looking down at the bottom of the table you can see that in the pembro plus chemo group, 44 completed treatment. In the pembrolizumab-only group, 46 completed treatment. And in the chemotherapy group, 195 patients completed treatment.

Looking at the baseline characteristics in the intention-to-treat population. This is stratified by pembro plus chemotherapy, pembrolizumab alone, and chemotherapy alone. Looking at several of these highlights, the median age for these patients was generally in the late sixties, and about three-quarters of patients were male. The majority of patients were ECOG zero and one, but you can see that 6% to 8% of these patients were ECOG two. In terms of primary tumor site, lower urinary tract in just over three-quarters of the patients, metastatic staging, the majority of these were M1 patients at 93% to 95%.

In terms of the most common set of metastasis, visceral metastasis was at around three-quarters of the patients. There were about 78% to 79% of patients that did not have liver metastasis. Looking at the PD-L1 CPS score, about 50/50 between greater than, or equal to 10, and less than 10. Looking at how many patients had prior adjuvant or neoadjuvant platinum-based chemotherapy, 11% in the pembro plus chemo group, 9% in the pembro group, and 13% in the chemotherapy alone group. Looking at the cisplatin or carboplatin as chosen by the investigator, about a 50/50 split between carbo and cisplatin in these three groups.

This is the first Kaplan-Meier curve that we will go through over a series of slides. This is progression-free survival, and you can see here that the pembrolizumab plus chemotherapy group is in blue and chemotherapy is in red. The hazard ratio, looking at these two treatment arms was 0.78 with a 95% confidence interval of 0.65 to 0.93 and a P-value of 0.0033. Looking at overall survival, comparing pembrolizumab plus chemo versus chemo alone, the hazard ratio of 0.86 and a 95% confidence interval of 0.72 to 1.02.

This is a Kaplan-Meier curve looking at overall survival with pembrolizumab alone versus chemotherapy alone, stratified by patients with a CPS score greater than or equal to 10. These are essentially overlapping lines with chemotherapy in red and pembrolizumab in blue with a hazard ratio of 1.01 and a 95% confidence interval, 0.77 to 1.32.

This looks at pembrolizumab alone versus chemotherapy alone, again in the population irrespective of the PD-L1 CPS score. And again, a non-significant relationship between these two arms, with a hazard ratio of 0.92 and a 95% confidence interval of 0.77 to 1.11. Again, pembrolizumab alone versus chemotherapy alone in patients that had a PD-L1 CPS score greater than or equal to 10 who received carboplatin, again, a non-significant relationship between pembro and chemotherapy in this analysis with a hazard ratio of 0.82 and a 95% confidence interval of 0.57 to 1.17.

So this table here looks at the summary of the anti-tumor activity. It's a relatively busy table, so I'll highlight a couple of items in this table. On the right are the three columns looking at patients with PD-L1 CPS scores greater than or equal to ten with all three groups, as you can see. And to the left of that in the middle of the table is the total population again, stratified by these three groups.

So in terms of proportion of patients with a complete or partial response, in the total population pembrolizumab plus chemo, 54.7%, in the pembrolizumab alone, 30.3%, and in the chemotherapy group alone, 44.9%. Shifting to the right, again, looking at the proportion of patients with a complete or partial response. In the pembro plus chemo group for patients with a PD-L1 status greater than or equal to ten, 57.2%, in the pembrolizumab alone group, 32.5%, and in the chemo group, 46.2%. Moving down to the middle of the table, looking at the best overall response for the total population, pembrolizumab plus chemotherapy, 40% partial response, 15% complete response, for the pembrolizumab group, partial response, 19%, and for the complete response, 11%, and in the chemotherapy alone group, the partial response of 33%, and the complete response of 12%. Looking at these same metrics in the patients with a PD-L1 status greater than or equal to ten, for pembro plus chemo, a partial response of 42%, and the complete response of 16%, for the pembrolizumab alone group, a 19% partial response rate and a complete response rate of 13%, and for chemotherapy alone, a partial response rate of 30% and a complete response of 17%.

In the next two slides, we will look at adverse events. As you can see, another very busy table is stratified by the three treatment arms. We'll look at basically the grade four adverse events in these groups. In terms of the chemo plus pembro arm, 31% of patients had any grade four adverse events, for pembrolizumab alone, 13%, any grade four adverse events, and for the chemotherapy group, 28% any adverse events of grade four. Looking at specific adverse events: blood and lymphatic system, 14% grade four adverse events for the chemo plus pembro arm, 1% for pembrolizumab, and 14% for chemo alone. Other ones to highlight, looking at the infections and infestations for pembro plus chemo grade four, 5%, for pembrolizumab alone, 3%, and for chemotherapy alone 3%. And just for completeness, this is the rest of the table and as you can see here, renal and urinary disorders, pembro plus chemotherapy, 1% grade four events, 2% for pembrolizumab, and less than 1% for chemotherapy.

So in terms of discussion of the KEYNOTE-361 trial, this trial did not meet the primary endpoint of superior PFS and OS with first-line pembrolizumab plus chemotherapy versus chemotherapy alone in patients with advanced urothelial carcinoma. Importantly, neither of the PD-L1 CPS greater than or equal to ten nor physicians' choice of platinum chemotherapy was associated with improved benefit from the addition of pembrolizumab to chemotherapy. The PFS by central review endpoint was not met in KEYNOTE-361 whereas the PFS by investigator assessment with atezolizumab plus platinum-based chemotherapy versus chemotherapy was significantly better in the atezolizumab group of the IMvigor130 trial with a hazard ratio of 0.82 and a 95% confidence interval of 0.70 to 0.96.

So in conclusion, the KEYNOTE-361 trial adds to the growing body of evidence showing that immune checkpoint inhibitors given with chemotherapy are not associated with clear survival benefits for urothelial carcinoma. Based on the primary findings of KEYNOTE-361, platinum-based chemotherapy remains the current first-line standard of care for patients able to receive this treatment with avelumab maintenance therapy for those who have a clinical benefit.

Thank you very much and we hope you enjoyed this UroToday Journal Club discussion of the recently published KEYNOTE-361 trial.