Evaluating Long-Term Follow-Up Data for Treatment of mRCC

Daniel J. George, MD: Neeraj, these trials have different levels of follow-up. Some of these studies have just come out. You mentioned the CheckMate 214 study just presented 42-month follow-up data. How much of that factors into your comfort level in a recommendation for this frontline treatment, not necessarily having that 3- or 4-year follow-up data? Is that an issue for you? Is that something you feel comfortable with, still recommending some of these newer regimens?

Neeraj Agarwal, MD: Yes, a longer follow-up is always compelling for me, especially when I'm looking at the durability of complete response [CR], which is the first thing that comes to my mind when I think about ipilimumab/nivolumab. Yes, a 42-month follow-up is a big deal, and it continues to reinforce that this regimen induces those durable complete responses, at least until now. But not necessarily in the context of VEGF/TKI [tyrosine kinase inhibitor]-based regimens. When I'm not counting on complete responses, as Rana and Monty said, there are many patients who are going to be candidates for VEGF/TKI plus IO [immune-oncology] combinations. For those regimens, the durability of complete response may not matter as much as the overall response data, PFS [progression-free survival] data, or the absolute OS [overall survival] data. It’s relative. I don't expect, for example, cabozantinib, to induce durable complete responses for 5 years. However, I have patients, both in the METEOR trial and the original sunitinib trial, which randomized patients to 50 mg every day versus 37.5 mg, and I have a patient still responding to sunitinib from that trial. At the same time, I'm not expecting cabozantinib to show 5-year complete durability of response in most of my patients, so it's all relative. If I see data from a 42-month follow-up, it is definitely compelling.

Daniel J. George, MD: Neeraj, when you talk about a complete response, how complete is a complete response? Does it, in your mind, have to be pristine with not a millimeter of things left, or if there’s a 90%-plus reduction, then that’s good enough for completeness from your perspective?

Neeraj Agarwal, MD: I agree with you if it’s a near-complete resection. How will we know what that small spot in the lung is? We are not going to do a biopsy of that spot in the lung to see if there is still metastasis left. Many times, it’s scar tissue. Many times, it’s a benign lesion sitting in the liver, but we’d have to do a biopsy to figure out if it is really a benign lesion versus metastatic tissue. As long as I see the dominant lesions decreasing in size of more than 90%, I call it complete response.

Daniel J. George, MD: I do too. Rana, you brought up some of these other end points that patients are interested in and whatnot, and Neeraj has talked about these durable, near-complete responses. How do you view time off treatment as a goal for patients? That’s not something patients necessarily think about up front, but do you introduce that concept to patients when you’re starting on a frontline therapy? When do you consider a discontinuation of therapy in somebody who is having an ongoing response?

Rana R. McKay, MD: This is a continuing, evolving question in the field. With data from CheckMate 214, the way the trial was designed was that, if you had a significant toxicity while you were getting your nivolumab/ipilimumab, you had to discontinue treatment. This end point has evolved to treatment-free survival, which is basically the time that treatment discontinued to the time of initiation of the subsequent therapy. There are many different ways that a patient can spend their time off therapy. They can spend their time off therapy without any toxicity. They may have some minor toxicity, but that interval matters, and it’s becoming a new end point in the context of the IO/IO combinations, particularly people having to discontinue therapy for toxicity.

One of the biggest questions we get in the clinic is this: “How long am I going to be on this therapy, doctor? Am I going to do this forever?” The answer’s always been, “Well, indefinitely until it stops working or you have toxicity.” That’s always been our answer, but now the benchmark may be around 2 years. Patients are allowed to discontinue treatment at the 2-year mark if they were responding. We don’t yet have data on how those patients do. Time is going to tell what happens, but it’s a big question: can we safely discontinue therapy in responding patients?

To speak to that a bit, we looked at that in the context of a multicenter, phase 2 trial called OMNIVORE; it asked the question of can we stop therapy in responding patients? The trial was small; it enrolled 83 patients. For people who had a confirmed PR [partial response] or CR within 4 to 6 months of starting treatment with nivolumab, they could be discontinued from therapy. From the patients who were allocated to the arm to be discontinued, 42% at 1 year remained off treatment. The flip to that is that 60% weren’t, so how do we appropriately identify those individuals for whom we can safely stop therapy? That’s an area where we critically need a biomarker, a biomarker of extreme responders because some people can do quite well.

Speaking to the durability of response with the data coming out from the updated data from KEYNOTE-426 at ASCO [the American Society of Clinical Oncology annual meeting] and updated data from CheckMate 214 at GU [Genitourinary Cancers Symposium] ASCO, as we look at these progression-free survival curves over time, the nivolumab/ipilimumab combo seems to maintain a steady hazard ratio over time. There is not a big shift in that 0.71, 0.72 hazard ratio, and when we look at the KEYNOTE-426 data, we see a progressively worsening hazard ratio, with the most recent being 0.68 from the updated data with an intend-to-treat population. It doesn't look as good over time with the IO/VEGF, and again speaking to that, are we curing patients? For the durability of the IO/VEGF, time is going to tell what happens to those patients, but treatment-free survival matters.

Daniel J. George, MD: Absolutely. This is one of the nice things about the longer follow-up in these studies that we’ll learn. These data are all still relatively new, even 42 months is not that long in the natural history now of some of these patients. There's a longer story to these regimens.

Transcript Edited for Clarity