Jeffrey S. Weber, MD, PhD: Michael, is there a place for ipilimumab alone in a patient who didn’t have a good response? Or would you add it to an existing PD-1 antibody?

Michael A. Postow, MD: That’s a really good question. Where are we with ipilimumab now? Because we’ve come so far with our use of immune therapies that, really, we’ve seen PD-1 antibodies, like pembrolizumab and nivolumab, as monotherapy replacing ipilimumab. We have superiority of both over ipilimumab, if we’re thinking about monotherapy with immune checkpoint inhibition. But a lot of patients, unfortunately, won’t respond to a PD-1 antibody by itself, and there is a role for ipilimumab in those patients. There are responses in patients who have progressed on pembrolizumab or nivolumab to ipilimumab monotherapy. One other question still remains: who should be getting ipilimumab with a PD-1 antibody as an initial treatment for patients starting with immune therapy? That question is still being worked out, as we see more and more data emerging from the CheckMate-067 study and longer and longer follow-up periods are being reported from that trial.

I think what we can conclude from that question—do we start with ipilimumab and a PD-1 antibody or do we start with a PD-1 antibody alone—is that the response rates are highest when we combine both drugs together. Long-term survival by numbers may favor the combination a little bit. However, it’s not a statistical difference, so we can’t say with significant confidence that the combination really improves long-term survival. So, if we’re looking at long-term survival, they are approximately similar, but there are higher responses with the combination. It’s a real question about what’s important for our patient—what kind of benefits do they need in the short run and what kind of patient could they be in the long run with the treatment—and trying to make that decision. Who should get the more aggressive combination, ipilimumab and a PD-1 antibody up front, or should we start with PD-1 and maybe use ipilimumab as a second-line approach? And there’s no right or wrong answer. It’s just a conversation with the patient and really trying to go on what your gut tells you might be the right thing in an individual circumstance.

Jeffrey S. Weber, MD, PhD: Is there ever a reason to sequence the drugs, meaning a forced sequence or an assumed sequence, or would you always just give the drugs concurrently if you know you’re going to use them?

Michael A. Postow, MD: The CheckMate-064 study that forced sequence, which is that patients either started with nivolumab and then force-switched to ipilimumab or vice versa, started with ipilimumab and force-switched to nivolumab. All the data seem to focus on the efficacy that when you started with a PD-1 antibody, it was better than starting with ipilimumab. So, if you’re thinking about sequencing, you still want to start with a PD-1 antibody. I don’t know if I would force-switch a patient who I was taking care of if they were having a good initial benefit with a PD-1 antibody alone. I would continue that for a patient if they’re tolerating it well. How long to continue that patient who’s doing well with a PD-1 antibody is an ongoing question. We’re getting some data from Dr. Robert and others that we really may be able to stop PD-1 in certain patients, either the complete responders or people who have been on for 2 years. But we’re still trying to work that out and know, what does it really mean? Maybe we should stop patients earlier. We just don’t have any prospective data, where we really test shorter durations of immune therapy to definitely know.

Jeffrey S. Weber, MD, PhD: Yes, that’s a question that patients ask. What do you tell patients when you have them on pembrolizumab and they’ve had a good partial response? When do you stop?

Axel Hauschild, MD, PhD: In Germany, it’s an ongoing discussion, and it’s very controversial, particularly if you have reached a complete response. There are centers now who made the decision to stop the treatment immediately. Others are leaving the patient on the drug for another 3 months and then they stop. My policy is completely different. I discuss it with the patient. In the absence of significant toxicities, if the patient is willing to be further treated, I don’t stop the treatment on my own. I continue, because in the German healthcare system, we don’t need to stop. I don’t want to be the guilty party at the end. So, if the patient is willing to be treated and I think I can continue, why should I stop? It’s tougher if you are talking about tyrosine kinase inhibitors. You have the impression that you have more relapses early after you stop, but we also have nice examples of continued responses. But the duration of response is extremely important, and I believe the more complete responses and partial responses you observe, the more likely the patients are to have long-term benefit.

Honestly, I would really appreciate clinical trials driven by communities, by working groups, who are comparing different durations of treatment. The best example for me is the treatment of ipilimumab and nivolumab because the median time on treatment is just 3.5 months. Why don’t you have a clinical trial with just 3 months of the combination immunotherapy compared to any other treatment with a longer duration? And it could well be that just 3 months of treatment is enough.

Jeffrey S. Weber, MD, PhD: Do you agree?

Caroline Robert, MD, PhD: I totally agree that today, we need to put the patient in the decision because we cannot say something that is based on the retrospective data. We need to inform the patient, and we know that we have some patients who are afraid to stop and we cannot oblige them to stop. We have some patients who just say, “Please, stop this drug. I’m doing well.” So, we can tell them that we are a little bit reassured now, because we have these retrospective data that are very reassuring. But I agree, we absolutely need to have this discontinuation randomized trial. And, in fact, the ipilimumab/nivolumab combination, because of its toxicity, is obliging us to stop even during the induction period for a lot of patients—close to 40% of the patients. We see that when they respond, they respond with the same duration as the patient for whom we didn’t stop. So, because of the toxicity, we have information that is very important for us.

Jeffrey S. Weber, MD, PhD: Although it seems to work pretty well in brain metastases, doesn’t it? There were some very impressive data at ASCO. Can you tell us about them?

Michael A. Postow, MD: What we learned from ASCO is that in phase II studies where ipilimumab and nivolumab were given to patients with melanoma brain metastases, the response rates were over 50%. So, we basically see almost the same response rates in the brain as we do with extracranial response, at least with the combination of ipilimumab and nivolumab. We still need some more information from that study, and I think what’s most important is, what is the rate of new lesion development in the brain? Because response rate is important in the brain. It’s great to see metastases shrink. But someone could also say, “Well, I could administer stereotactic radiosurgery and make the melanoma lesion in the brain shrink.” What’s really going to be important and informative from those studies is how effective this is at treating micrometastatic disease in the brain and the prevention of new lesions emerging over time. That’s going to be, hopefully, the most important benefit from this in treating the invisible enemy that is melanoma in the body that you just can’t see on scans. That’s really what the problem is for most patients.

Transcript Edited for Clarity