The Future Treatment Landscape of RCC

Michael Atkins, MD: Looking at the future landscape, it’s going to get a little more complicated. There are some interesting studies going on. One is the phase 3 CLEAR study, which is looking at lenvatinib plus everolimus or lenvatinib plus pembrolizumab compared with sunitinib. We just saw a press release that the lenvatinib-plus-pembrolizumab arm is better than sunitinib. This is no surprise because lenvatinib, at least in my experience, is also a much better drug than sunitinib as a TKI [tyrosine kinase inhibitor]. That’s not a surprise, but we also saw in the press release that lenvatinib-everolimus was superior to sunitinib. I’m going to be curious to see that data to see how much better lenvatinib-pembrolizumab is than lenvatinib-everolimus and how much of the benefit that we’ve been seeing in these TKI–I/O [immuno-oncology] combinations vs sunitinib are related to the fact that axitinib, cabozantinib, and lenvatinib are all better VEGF inhibitors than sunitinib.

Another trial that is going on that will yield interesting results is the COSMIC-313 trial, which is comparing cabozantinib plus nivolumab, with nivolumab and ipilimumab, the triplet, vs nivolumab/-ipilimumab. It will be the first head-to-head comparison where we get a chance to look at the impact of a TKI on the response to an I/O agent. It will be interesting to see whether the triplet is synergistic, additive, or subadditive. There have been a lot of preclinical data that suggest that VEGF TKIs will enhance the immune response, but I always caution people that none of that data is in VHL-null immune-competent tumor models because those models don’t exist. It’s conceivable that, in kidney cancer, because of the VHL mutation, the TKI may be causing acute hypoxia, which may compromise the ability of the immune response to eliminate the last tumor cell. I’m paying particular attention to the COSMIC-313 study to see what the tail of that PFS [progression-free survival] curve looks like in the triplet arm to see whether it looks like the nivolumab-ipilimumab tail.

Finally, there is the PDIGREE study, which is going on in the Alliance group. Patients are getting 12 weeks of nivolumab-ipilimumab, and after they’ve completed 12 weeks of nivolumab-ipilimumab, if they have either a non–CR [complete response] or nonprogressive disease, which maybe will be about half the patient population, they’re being randomized to get either cabozantinib-nivolumab or nivolumab. That’s an interesting question, but it has 2 caveats in my view. No. 1, it’s a large number of patients being accrued to get to the randomized group, which is at the second-line setting. And it’s unclear how long that trial is going to take, how many people are going to be eligible for that randomization, and how many haven’t had such toxicity that they have to stop treatment or have progressed or refused the randomization at that point, and whether it’s going to be relevant if most of the patients are getting a TKI and an anti–PD-1 agent in the frontline setting.

Those 3 trials are coming, and they’re going to address some questions. Until we see those results, and even with those trials, it’s going to be a discussion about I/O–TKI versus I/O–I/O in the frontline setting.

Robert Alter, MD: That’s right. We then have the sequence of therapies. How do we use cabozantinib, which got it’s FDA approval in the METEOR clinical trial? CheckMate 025 brought nivolumab into our world. We mentioned the lenvatinib-everolimus combination before in the Study 111 trial, and it was the first time we saw a combination therapy that was well tolerated with significant numbers. In the second-line setting, initial presentation was a 43% response rate, but it went down to 37% as a second-line therapy with a progression-free survival of 14.6 months compared with everolimus by itself, which was the standard arm, which is 5.5 months.

We started realizing that not only do we have some excitement as a combination first-line therapy, but we may now be using them in second-line therapy. Can this lenvatinib-everolimus combination move to a first-line therapy to avoid I/O in the first line? It’s going to be very hard. We still have to believe that our patients have to have tolerable therapies, and it’s not just about giving them continuous oral therapies that will probably lead to great response rates and progression-free survivals. We’re avoiding it: The topic we’re thinking about is the CR and durable time off-therapy. When you tell a person who has metastatic cancer that you can have them on surveillance after they receive therapy, you are sometimes buying this patient years off-therapy. You have experience, and we have experience as well: For previous patients with high-dose IL-2, for as many courses as we got them through—which may have not been so many—you stop therapy. They’re on surveillance for decades. There’s nothing more fulfilling than that.

Michael Atkins, MD: When we’re talking about treatment sequencing, I agree: it’s always best to get the front first-line treatment that cures the patient so they never need second-line therapy, but whether I’m giving nivolumab-ipilimumab or a patient has gotten axitinib-pembrolizumab, I’m still leaning toward cabozantinib as my second-line therapy. I might occasionally give axitinib in that setting if the patient has a slow-growing disease but still wants therapy if they’re truly progressing. I’m then saving lenvatinib-everolimus for my third-line treatment because it’s 2 drugs, so it’s harder for the patients to take than cabozantinib. I find it to be very effective after cabozantinib, so I’m not surprised that lenvatinib is an active agent in the doublet with pembrolizumab in the CLEAR study, and I’m not surprised that lenvatinib-everolimus is better than sunitinib in the frontline setting. Where it may play a role in the frontline setting is in patients with non–clear cell kidney cancer, particularly chromophobe tumors or some of the aggressive papillary tumors that may not respond well to immunotherapy. That’s where the lenvatinib-everolimus combination looks like it might turn out to be the best option.

Transcript Edited for Clarity