Nizar Tannir, MD: Scott, give me your perspective on the CheckMate 214 study. The discontinuation rate of nivolumab [Opdivo]/ipilimumab [Yervoy] was 22%, higher than any other IO [immuno-oncology]-TKI [tyrosine kinase inhibitor]. We saw from our data with longer follow-up—a minimum of 5 years—of CheckMate 214 that patients who discontinued nivolumab/ipilimumab because of toxicity, whether during induction when they were getting the dual blockade or later on during the maintenance therapy with nivolumab—didn’t have their survival compromised. They did well. In fact, some of those patients who discontinued therapy didn’t require any further subsequent therapy and remained progression free. Some of them were in CR [complete remission] and maintained durable CR.

To come back to the point you mentioned, the value of an IO-IO combination is that durability and the opportunity to discontinue therapy, whether you discontinued it because of toxicity or patients decided to stop therapy. That’s the value of the durability and treatment-free survival or treatment-free interval that we see with IO-IO. What’s your take on that?

Scott Tykodi, MD, PhD: I have a few patients who have fairly limited exposure to IO therapy and had serious toxicity, stopped treatment, had ongoing sustained disease control, and never went back on systemic therapy. That’s a fascinating subset of patients. One of the more common issues we face—reference was made to the KEYNOTE-426 trial—is that the IO drug was capped at 2-year total duration, and that actually carries forward to the FDA approval. The package insert says a maximum of 2-year duration therapy of pembrolizumab [Keytruda]. I don’t believe the nivolumab development pathway, [CheckMate] 025, had a maximum duration of therapy, and I don’t believe [CheckMate] 214 had a formal end point for the IO. Nivolumab doesn’t have that same language. There’s this regulatory difference that cuts across the 2 drugs.

That being said, it’s so common now that studies limit IO exposure to 2 years. We’ve adopted that in our clinical practice. We usually have a conversation about intentionally stopping therapy with patients with small-volume disease who have been stable and hit the 2-year landmark. But we like to feel that it’s elective and not forced on us by the insurance provider. I have a lot of patients who electively stopped therapy at about a 2-year total duration of immunotherapy—mainly patients on the ipilimumab/nivolumab regimen—and they’ve done extremely well. We’ve had a couple patients restart treatment, but not soon. I’ve had nobody with explosive progression of disease. That’s a fascinating feature that’s closest linked to the ipilimumab/nivolumab regimen, but we have fewer data in patients coming off the IO-TKI regimens in a similar clinical circumstance in terms of the durability and what to expect.

One feature of these regimens that strikes me as a funny disconnect is that the CR rates across the board are all better than we’ve ever had for renal cell: 8%, 10%, 12%, 16% in clear. But where are these patients? I have very few true radiographic CRs in my clinic population. I have a lot of small-volume PRs [partial responses], and those are the people for whom we talk about suspending therapy on purpose, but they aren’t experiencing a true CR radiographically.

There’s a backhanded explanation about the behavior of ipilimumab/nivolumab. The non-nephrectomized patients have been analyzed in a publication, and the CR rates in the non-nephrectomized cohort was 0%. Conversely, that tells you that if you achieved CR, you had to have a nephrectomy, and you had to have the primary removed surgically. I haven’t seen the same analysis applied across the other IO-TKI regimens, but it opens this can of worms that after CARMENA, if we aren’t doing a lot of cytoreductive nephrectomies in the metastatic setting, and our nephrectomy incidence in our population is falling lower, isn’t that dragging down the CR potential of these treatments that are lower? CRs aren’t 10% to 12%. They’re quite rare, and most patients have their primary tumor on board.

A CR and stopping therapy is a lovely end point. You’ve won. The patient gets their life back. And the toxicity risk goes down because they’re off therapy. Everybody agrees that’s a wonderful end point. What’s more troubling is the long-term stable disease and control. What’s the end point? Where do you stop? Do you stop people? If you’re on an IO-TKI regimen, do you stop both drugs or just the IO to mirror the clinical trial experience? We don’t have a lot of data to guide us on what happens to the long-term successful patient, but there are more of those, so we’re desperate to know the behavior of the long-term disease control to guide us about best choices for those patients.

Nizar Tannir, MD: Thank you, Scott. To your point about CheckMate 214, we didn’t stop maintenance nivolumab after 2 years like in the pembrolizumab phase 3 trials. Nivolumab was continued at the discretion of the treating physician as well as patient consent. Some of those patients continued therapy beyond 2 years, and even beyond 5 years. That’s 1 point.

You asked what we do for those patients. Do we stop after 2 years? My practice is that if they achieve a CR on nivolumab/ipilimumab, I discontinue therapy after 2 years. However, with an IO-TKI, I have stopped after 2 years in some instances if they achieve a CR and have seen relapses after the 2 years. The question then becomes whether you can salvage. Can you induce another response by going back to the same therapy that the patient responded to initially? In the case of nivolumab/ipilimumab, if the patient relapsed after stopping maintenance nivolumab, do you go back to nivolumab again? Have you seen responses after you resume nivolumab from CR if they relapsed? Or should the patient receive subsequent therapy?

Transcript edited for clarity.