Read the Full Video Transcript

Rashid Sayyid: Good morning, everyone. This is Rashid Sayyid. I'm a urologic oncology fellow at the University of Toronto and along with Zach Klaassen, associate professor of the Department of Urology. We'll be discussing the latest publication looking at cabozantinib plus nivolumab and ipilimumab in renal cell carcinoma. This trial was recently published in the New England Journal of Medicine with Dr. Toni Choueiri as the first author. We know that over the last few years, there's been numerous studies in this disease space that have looked at combination therapies in the first-line setting for patients with locally advanced or metastatic clear cell carcinoma. One of the earlier trials in this setting was a CheckMate 214 trial that looked at the combination of ipilimumab plus nivolumab versus sunitinib, which was the standard of care at that time. This study included over a thousand patients with advanced clear cell RCC with IMDC intermediate or poor risk disease.

As we see in the Kaplan-Meier curve, there was a significant improvement in the overall survival with the combination of ipilimumab and nivolumab with 18-month overall survival improved from 60% to 75%. We also saw a benefit with the objective response rate from 27% to 42%. However, we note within this trial that about 20% of patients receiving the combination of ipi and nivo had disease progression as best response. So clearly, those patients did not benefit from this treatment. As such, is this a sign that we need to further intensify treatment in this disease space? What about cabozantinib? Cabozantinib is a tyrosine kinase inhibitor or TKI, which is part of the VEGF receptor MET, TAM family of kinases. This drug has been evaluated in the first-line RCC setting as a standalone drug in the CABOSUN trial, which included patients with IMDC intermediate and poor risk disease who previously had untreated advanced RCC. So a patient randomized to receive either cabo at the full dose of 60 milligrams versus sunitinib, 50 milligrams.

As we see in the curve here, the progression-free survival was significantly improved with cabozantinib at 8.6 versus 5.3 months, and also the objective response rate over twice as good at 20%. Have we looked at cabozantinib before in combination in the first-line setting? The answer to that is yes. The CheckMate 9ER trial evaluated the combination of cabo plus nivolumab versus sunitinib as well in patient with advanced RCC. Again, we saw an improvement in the survival outcomes with a median PFS of 16.6 versus 8.3 months. Overall survival at 12 months was improved by almost 10%. So the authors at that time, Dr. Toni Choueiri et al. also performed an exploratory analysis of the patients who received cabo with ipi and nivo later on and they found that that triplet regimen has excellent clinical activity with an acceptable safety profile. So clearly, quite a strong signal that there may be something there.

As such, they designed the COSMIC-313 trial to evaluate the combination of nivolumab plus ipilimumab with or without cabo in patients with previously untreated advanced RCC with IMDC intermediate or poor risk disease. So just by way of trial design, this was a phase III randomized, double-blind, placebo-controlled trial, was a one-to-one randomization to either cabo given at 40 milligrams. We note in the CABOSUN trial, 60 milligrams, so a bit of a reduced dose here plus the nivolumab and the ipilimumab as well. Conversely, the placebo-controlled arm received placebo instead of cabo. We also note that in both arms, the nivolumab was continued as maintenance therapy at a dose of 480 milligrams every four weeks for up to two years. We'll also see in the next slide that cabo was always continued afterwards. We note that randomization stratified by the IMDC risk and the geographic region. The reason this is done is to make sure that the two groups are well-balanced for these important risk stratification variables.

No crossover between treatment arms was not permitted. If there's any side effects or poor tolerance to cabo, dose reductions to 20 milligrams every day from 40 and then to every other day were also permitted. This schematic very nicely summarize the trial design. If we start here on the left side, we note that this trial include about 840 patients. It's very important these patients received no prior systemic therapy for metastatic RCC. They were allowed to receive systemic therapy in the adjuvant setting as long as it was not one of the drugs in the trial. We note also that this trial included clear cell renal cell carcinoma only, so others like papillary and chromophobe were not included. Inclusion's restricted to patient, the intermediate, poor risk per the IMDC criteria and they had to have measurable disease per RECIST to monitor the radiographic progression.

As is classic with these trials, patients had a good performance status with a Karnofsky performance status greater than 70%. We note, as we talked about before, the randomization's one-to-one to either cabo, nivo plus ipi. Then we also note here that cabo and nivo was continued as maintenance therapy afterwards. Conversely, in the control arm where they received placebo instead of the cabo and then only the nivo was continued as maintenance therapy. On this slide also summarize the eligibility criteria that we discussed. What about endpoints? The primary endpoint was progression-free survival. And as we'll see in the later slides, there was two cores. There was the PFS or the progression-free survival population, the OS population. These were powered differently to allow for earlier analysis. If we look at the primary endpoint of PFS, this was specifically in the first 550 randomized patients and this again was assessed per RECIST version 1.1. and this is classic with these trials.

They importantly used a blinded independent radiology committee to minimize the risk of reader bias or measurement bias. For secondary outcomes, this included overall survival, objective response rate, the duration of response and safety and adverse events outcomes. How were patients followed? Cross-section imaging was the main ones, either CT or MRI. So this was done at baseline and then week 10 and then every eight weeks through week 50 and then every 12 weeks thereafter. Next, we look at the target sample size, which was 840. The reason this was decided was that this will allow for enough power to evaluate overall survival. If we look at the PFS cohort, the first 550 patients, so this analysis required 249 events and as such, we allow for 90% power, which is higher than what we typically see to detect about an improvement of about 33%, 34% in experimental group.

Conversely, in the overall survival analysis, you need 840 patients and 433 deaths. It's very important that it's not just the number of patients that dictates the power in this study, but it's also the number of events. Obviously, you would expect a progression event to occur earlier than mortality and as such, this plays in part a role in determining the sample size. This also was powered at 90% to detect the hazard ratio improvement at 0.73 at a type 1 error rate of 5%. A pre-specified interim analysis of overall survival was performed at the time of the analysis of the PFS, but was not yet reported. The reason is that's to reduce bias to allow overall survival data to remain concealed from investigators and patients. However, as is a classic with these trials, it was only available to a small group from the trial sponsor and independent data and safety monitoring board to make sure that this was not a futile trial.

One additional interim analysis of overall survival is currently planned. To make sure that with repeat testing, we don't inflate our type 1 error rate and the false positive rate, the authors will apply the O'Brien-Fleming alpha spending function to minimize this issue. They also use survival analysis with Kaplan-Meier curves and the log-rank test to allow for between-group comparisons. And they performed stratified analysis with hazard ratio 95% confidence intervals that were generating using the Cox proportional-hazards modeling. At this point, I'll turn it over to Zach to go over the results and the discussion of this paper.

Zach Klaassen: Thanks so much, Rashid, for that great introduction. This is the baseline patient characteristics broken down by two slides based on the size of this table. Since this is the PFS analysis, we'll focus on the PFS population for discussion of the baseline characteristics. We can see in the experimental and the control arm, the median age was early 60s. Roughly three quarters of these patients were male. The majority of these patients coming from US, Canada, Europe, Australia, or New Zealand. Majority patients, roughly 80% were white. The second most common demographic was Asian at just under 10%. In terms of IMDC risk category towards the bottom of this slide, intermediate risk was three quarters of these patients, poor risk was 24%. As Rashid alluded to, all of these patients had good to excellent Karnofsky performance status of 70 to 100. When we look at the previous nephrectomy rate, this was 64% in each of these arms.

Number of target organs with target and non-target lesions was most commonly greater than or equal to two at just under three quarters of these patients. The median sum diameter of target lesions was well-balanced at roughly just under nine centimeters. The most common site of target and non-target lesions was the lung at roughly two thirds of the patients and lymph node at 47% in the experimental group and 43% in the control group. This Kaplan-Meier curve is the final analysis of progression-free survival. You can see it early in clear separation of the curves between the triplet therapy and the doublet therapy with a median PFS for triplet therapy of not being reached and for the control group doublet therapy of 11.3 months with a hazard ratio for disease progression or death of 0.73 and a 95% confidence interval of 0.57 to 0.94.

This is PFS in the pre-specified subgroups. To the left of the line, which is hazard ratio of one favors the triplet therapy, to the right, favors the doublet therapy. Just looking at a general overview of this forest plot, we can see that majority of these patients in all of these subgroups had some sort of benefit. Now there's some with small sample sizes such as sarcomatoid features and poor risks that did not derive as big of a benefit as some of the other groups or the overall cohort, but for the most part across these subgroups, there was benefit to the triplet therapy. This is PFS by intermediate risk subgroup. We can see that triplet therapy, the median PFS was not reached and in the doublet therapy, the median PFS was 11.4 with a hazard ratio favoring the triplet therapy of 0.63 to 95%, confidence interval of 0.47 to 0.85.

Similar curve looking at poor risk subgroup. Here we do see that there's no benefit to the triplet therapy with a median PFS of triplet therapy of 9.5 months and a median PFS of 11.2 months in the doublet therapy with a hazard ratio not statistically significant of 1.04. Looking at tumor response. In terms of objective response, the triplet therapy, 43% compared to doublet of 36%. Complete responses were actually the same between these two groups at 3% in each group, but we do see improvement in partial response, 41% versus 32%, stable disease, 43% versus 36% in the triplet versus the doublet. Progressive disease was less than the triplet therapy at 8% compared to 20% in the control group. We do see a disease control rate of 86% in the triplet therapy versus 72% in the control group. In terms of median time to response and median duration of response, no significant difference between these two.

With regards to adverse events, there's several important toxicity considerations from this table. At the very top, I've highlighted in the box any event. If we look at grade 3 or grade 4 for the experimental group, it was 79% compared to only 50% for the doublet therapy. We see significant increases in AST and ALT for the experimental group in terms of grade 3, 4 adverse events versus the control group, 27% versus 6% and 20% versus 5%. We also see for hypertension grade 3 or 4 events 10% for the triplet therapy versus 3% for the doublet therapy. So by way of discussion, in COSMIC-313 patients with intermediate and poor risk advanced renal cell carcinoma treated with cabozantinib plus nivo, ipi had significantly longer PFS versus nivo, ipi alone. However and importantly, the addition of cabozantinib to the nivo, ipi regimen did not provide additional benefit to poor risk patients. However, this may be secondary to a relatively limited sample size of poor risk patients.

The percentage of patients with complete response was similar in the two groups, which was relatively low at 3% compared to the nivo, ipi arm in CheckMate 214, which had a complete response of 9%. We do note that in COSMIC-313, the proportion of patients that had not undergone nephrectomy at 36% was higher than reported in other phase III advanced RCC trials, which may have reduced the percentage of patients that were able to generate a complete response. Other ongoing phase III trials are evaluating triplet combinations, and this includes pembro plus belzutifan plus lenvatinib or pembro plus quavonlimab plus lenvatinib versus pembro versus lenvatinib. So we do have additional phase III trial data coming, which hopefully will continue to show benefit of the triplet therapy.

So in conclusion, among patients with previously untreated advanced RCC who had intermediate or poor prognostic risk, treatment with cabo plus nivo plus ipi resulted in significantly longer PFS than treatment with nivo plus ipi alone. As mentioned, adverse events and discontinuation were more frequent in the experimental group than in the control group and this trial will have continued analysis for overall survival in the future. We thank you very much for your attention. We hope you enjoyed this UroToday Journal Club discussion of the recently published COSMIC-313 trial in the New England Journal of Medicine.