Panelists: Robert A. Figlin, MD, Cedars-Sinai Medical Center; Thai H. Ho, MD, PhD, Mayo Clinic Arizona; Martin H. Voss, MD, Memorial Sloan Kettering Cancer Center; Michael B. Atkins, MD, Georgetown Lombardi Comprehensive Cancer Center; Sandy Srinivas, MD, Stanford University Medical Center
Robert A. Figlin, MD: We’ve offered patients frontline therapy. We’ve made our best decision about what that should be. We’ve tried to optimize the approach, the treatment, the duration of treatment, and the management. How do you use all of that information? Do you use all of that information when thinking about a situation where now the patient is clearly progressing?
Using Martin’s words—new sites of disease, symptomatic, clearly needing a change in therapy—clinical trials are obviously one option. But do you use information about their response in the frontline setting to guide what you do in the second-line setting? And, if so, how do you think about that?
Thai H. Ho, MD, PhD: If you look at some of the METEOR subgroup analyses, where the drug cabozantinib was compared against everolimus, the patients that tended to benefit from cabozantinib had visceral metastases—they had bone metastases. For a patient who has a large explosion of disease, bone metastases, I tend to reach for cabozantinib because I think that there may be some opportunity to shrink disease or stabilize the disease. Whereas, for patients who may be intrinsically resistant to VEGF, maybe they lasted on VEGF for just 3 months, I’m more inclined to pick immunotherapy.
Robert A. Figlin, MD: Martin, this is going to be the hardest part of our conversation. It’s hard because we have a burgeoning amount of data, but not the perfect data. In the second-line setting, we have a series of trials with agents like cabozantinib, nivolumab, lenvatinib, and everolimus, as well as axitinib, which is an older agent in that space. Some of them have been compared to sorafenib, and others have been compared to an mTOR. Many of them demonstrate overall survival benefits—some of them without progression-free survival benefits, and others with a higher objective response rate. So, we have a very mixed bag of drugs in a very important disease state without fair ways to compare them to each other. How do you navigate through that space?
Martin H. Voss, MD: It’s challenging. It’s very confusing to treating investigators that are not constantly looking at kidney cancer patients. Mike has mentioned, a couple times now, that the newer agents appear to be more effective than what we’ve used for some time, and I would agree with that sentiment. Although these are novel TKIs—and there was at some time a sense, in the field, that doing more and more and more TKI therapy wasn’t going to get us anywhere—I think we all get the sense that these newer VEGF-targeted agents that co-inhibit other molecular targets, that are likely driving acquired resistance to TKI therapy with the older drugs, do have stronger efficacy. We always discourage cross-trial comparison, but we do it all the time because, what else can one do when trying to make a decision between medications?
In the second-line setting, the first decision one has to make is whether you continue a VEGF-targeted therapy, or a molecular-targeted therapy, versus an immuno-oncology approach. I think one of the most useful pieces of information that I get out of treatment with a first-line therapy before making that decision is how the patient tolerated that treatment. It’s very helpful to see whether a patient had a really rough time on first-line TKI therapy, especially considering the toxicity profile for the newer agents that are, in my mind, more effective than some of the traditional TKIs that we’ve had for a long time.
But, they also can be challenging in terms of managing toxicity. It’s manageable, and we’ve done it for a long time. We can keep patients on these drugs. But, if you look at the toxicity data, for instance, for lenvatinib plus everolimus, you can see on that second-line trial, a quarter of patients had to discontinue due to toxicity. And these are toxicities that we’re used to seeing with the same mechanism of action in the first-line space. So, if I have a patient who comes off from first-line therapy and has had chronic diarrhea for many months, I’m not very likely to put a patient on this combination that has a high rate of grade 3 diarrhea.
On the flip side, we can look at the disease kinetics that Thai has alluded to before. We can try to understand whether something progressed rapidly on first-line therapy or not. I actually think a little bit differently about it. If I have a patient who comes off of first-line therapy with rapid progression, I wouldn’t necessarily cross out another VEGF-TKI at that point. I think it’s very helpful, when we look at the best objective response data for all of these trials, to also look at primary progression of disease.
We can see that for some of these new drugs, the rate of progression as the best response is significantly lower than what we see on other drugs. And that’s helpful information if you have a patient who has got very symptomatic, rapidly progressive disease. It may speak to the pace of responding to therapy. We know that immunotherapy, for some patients, can be rapid in response, but it can take longer. It’s actually very interesting to compare the PD-1 data across diseases.
I also treat bladder cancer patients and, if you look at the bladder cancer data for PD-1 and PD-L1 inhibitors, responses tend to happen quite early—mostly on the first scan. For kidney cancer, it’s different. We see that dispersed first, second, third scan. The median time to response on the CheckMate-025 study was 3.5 months. So, again, that patient who has rapidly progressive disease, symptomatic, say, lung metastases, may not be able to wait another 3 months to respond to their therapy. And even if they’ve progressed quickly on the first-line TKI, I might re-challenge them with another TKI—maybe with a lenvatinib/everolimus combination, which has a high response and a low programmed death rate, or maybe with cabozantinib if they have a high disease burden and bone metastases.
Robert A. Figlin, MD: What I’m hearing, which I think is really very astute, is we’re thinking about prior toxicities on treatment. We’re thinking about the need for the rate of response that we’d like to see from the patient, given their clinical situation. We’re obviously, as Thai pointed out, looking at the sites of metastatic disease and possibly choosing between those. Those seem to be different conversations than we had 5 or 10 years ago about what to do next. Sandy, what are your thoughts about that?
Sandy Srinivas, MD: There’s a whole spectrum. We have about 10 drugs and, ideally, I think you would want a patient to derive benefit from each of these drugs. But, in reality, only 20% end up really getting to a third-line setting. You have to be careful when deciding on the right time to bring your most powerful drug up front, given how heterogeneous the disease is, considering that there may be a group of patients for whom axitinib may be a perfect drug. You don’t need to bring in immunotherapy. Or, you don’t need to bring in cabozantinib or lenvatinib. It’s challenging. There’s really no 1 biomarker that’s going to help us, and, in many ways, it’s us seeing so many of these kidney cancer patients that you have a sense of what the pace of the disease is going to be. I think that’s something that’s hard to really write down on a piece of paper—to say how to best pick that drug.
I think some of the other considerations surround, also, oral versus intravenous therapy—how strong the patient’s engagement is, how far they want to drive in coming in to get an IV drug versus wanting an oral drug, or 1 drug versus 2 drugs. One thing we haven’t brought up is the cost and the copayment for patients. That’s also a factor in terms of what they would want to do. So, I think it’s a little bit more than just picking a drug now.