Frontline Metastatic RCC: Pembrolizumab Plus Axitinib

Martin H. Voss, MD: Why don’t we segue over to immuno-oncology-TKI [tyrosine kinase inhibitor] combination? Ipilimumab/nivolumab was FDA [Food and Drug Administration]-approved in 2018 based on the data that Amishi just told us about. Tom, maybe you can tell us a little about pembrolizumab/axitinib, which followed shortly thereafter. The original publication was in the New England Journal of Medicine in 2019. You were an investigator on the pivotal trial for this and have since published follow-up data to this. We also have more recent data presented at this year’s ASCO [American Society of Clinical Oncology] Annual Meeting with further survivor follow-up. Could you walk us a little through the design of the trial then maybe focus on the current data with extended follow-up?

Thomas Powles, MD, MBBS, MRCP: This was a large, randomized phase 3 study, and it’s a classic design. The design is similar for all of these trials, which is quite convenient in some respects. But it lends people to make indirect comparisons, and whether we should or shouldn’t be doing that is a discussion for a different day. But it’s exactly the same design. It’s important to remember there are some subtle differences in that this really was an ITT [intention-to-treat] design rather than focusing on the intermediate- and poor-risk categories. Although, the original CheckMate 214 study did include those good-risk patients as well.

It’s an ITT study focusing on good-, intermediate-, and poor-risk IDMC [International Metastatic Renal Cell Carcinoma Database Consortium] subgroups, with axitinib 5 mg BD [twice daily], 3-weekly pembrolizumab at standard doses. You could dose escalate the axitinib if you want to. Axitinib is a good drug in some respects. It has a short half-life and there was a degree of personalized dosing, but it is quite specific for VEGF [vascular endothelial growth factor] targeting. It’s a good combination. There was really good phase 2 data that Michael Atkins, MD, presented with 65% response rates that seemed unbelievably good at the time.

Then, of course, the randomized phase 3 read out at interim analysis with an OS [overall survival] hazard ratio of 0.53, which was spectacular. It’s important to remember the minimum follow-up was only 12.3 months at that time. That is an important piece of information because you’ll hear a series of hazard ratios this evening, but they’re all related to the timing of the analysis because all of them drift downwards a little as time goes by. That’s a relevant point. When someone says, “0.6? That’s not as low as 0.53,” you’re better off looking at the time in your follow-up and you’ll realize it’s more like 20 months rather than 12 months. That makes the indirect comparisons very difficult.

Nevertheless, the 0.53 hazard ratio was terrific. The hazard ratio was 0.6 for progression-free survival [PFS]. The response rate was 60% vs about 35%. It’s important to note that the progression-free survival was 15 months at that time point, which was long. Now lenvatinib and pembrolizumab will show you data that’s perhaps longer coming up in a minute, but it was groundbreaking at that time.

The tolerability profile was acceptable. There was some transaminitis, but it’s quite manageable because you can just stop the axitinib for large groups of those patients. I think there were 0 treatment-related deaths for transaminitis. The rest of the tolerability profile was as you would expect. The drugs are more difficult to give than single-agent sunitinib. It’s more challenging than single-agent VEGF-targeted therapy. There’s debate about whether it’s more or less complicated than immune-immune combinations.

Subsequent to that, we had a series of other cut-point analyses that were performed. There have actually been 3. One was EMA [European Medicines Agency]-requested. There was then a publication that we put in The Lancet Oncology. Then there was the most recent update. Essentially, it goes from 0.53 to 0.60. It’s now stabilized somewhere around 0.7. We’ve got 0.68 and 0.73 most recently, which is still a significant survival advantage. We now have much longer follow-up in that longer follow-up. If you look at the landmark OS and the landmark PFS, it looks very similar to what we saw with ipilimumab/nivolumab.

In summary, we have a positive randomized phase 3 clinical trial with a significant survival advantage. It’s a well-tolerated combination. There is a debate going on about the IMDC [International Metastatic RCC Database Consortium] good-risk population. There were better response rates and better PFS. The OS is not significantly better in that group. There were a small number of events. Some people say we should wait longer. Other people say that all the other data are pointing in the right direction. I’m sure there will be a tail on the curve in the end. I think pembrolizumab is a very attractive regimen with a good track record.

Martin H. Voss, MD: Thank you so much for the very comprehensive review here. I have a follow-up question. You speak to the series of reports we have seen going from just about a year of media and follow-up to just under 2 years and now 43 months median follow-up. What is unique about this regimen compared to ipilimumab/nivolumab is that the checkpoint inhibitor for all patients stops at 2 years. So the patient has not progressed on KEYNOTE-426 and reaches the 2-year point. They continue the TKI—the axitinib—but they discontinue the pembrolizumab infusions. I’m curious what your thoughts were seeing the follow-up data with the hazard ratios the way they’ve trended and looking where we are now beyond the 2 years with a median follow-up to see if you noted any change or trend that would speak to a biologic or therapeutic effect or lack thereof past the 2-year mark.

Thomas Powles, MD, MBBS, MRCP: Martin, the data’s not in the public domain around this specific issue, but it is an important issue. The common-sense approach is to say that after 2 years of immune therapy, it should probably have been enough. And I understand that. There were the other people who said, “If you’re responding to therapy, we should keep going, but how long is a piece of string, and how long is forever?” We don’t know the answer to that. We don’t see dips in the Kaplan-Meier curves. We don’t see a step in the progression-free survival at that time point, although clearly patients were enrolled at different times so you wouldn’t necessarily see that step which is the nature of Kaplan-Meier analysis.

On a Kaplan-Meier curve after 2 years, you would expect to see if there was a big problem, or a big step, and we haven’t seen that. That’s important. We are in the process. I’m very interested in looking at the individual patients who get to that time point and working out what happened subsequent to that. You’ll also be aware that there are some ipilimumab/nivolumab data of patients stopping combination therapy and having long treatment-free outcomes. The subtle difference there is in that group of patients, they’ve had toxicity that had resulted in cessation of therapy. Where in our study, in the KEYNOTE-426 trial, we stopped therapy at the 2-year time point.

It’s a really important issue. The reason it’s important is from a global perspective, reimbursement for an indefinite period of time comes with much higher potential cost because that could be 10 years of therapy. When you put those numbers into mathematical models, it makes the treatment much more expensive, even if that tail is relatively small. It’s perhaps an important issue which we need to address more carefully.

Martin H. Voss, MD: I couldn’t agree with you more, Tom. I’m anxious to see more data in that regard.

Transcript Edited for Clarity