Weber Reviews Data Supporting Checkpoint Inhibitor Discontinuation Following Stable Disease in Melanoma

Jeffrey Weber, MD, PhD

Jeffrey Weber, MD, PhD

With a number of checkpoint inhibitors available for patients with melanoma, the question that remains is how long a patient should receive this therapy. During a presentation at the 2018 ESMO Congress, data was reviewed to support discontinuation of checkpoint inhibition after 1 to 2 years following stable disease.

Remaining on treatment brings up a number of challenges, such as inconvenience, toxicity, and economical factors. For these reasons, it is impractical for patients to remain on treatment for many years after achieving stable disease, explained Jeffrey Weber, MD, PhD.

While current data is mostly retrospective, it still provides an idea of whether patients should remain on treatment after achieving stable disease. Weber said he is comfortable telling his patients that they can stop treatment and be monitored going forward.

“The likelihood of relapse is low and even if you do relapse, the rate of re-remission would be at least 50%,” Weber said. “Now how long those remissions last, we don’t know, but the bottom line is you’re safe to stop at 2 years.”

In an interview withTargeted Oncology,Weber, deputy director and co-director of the Melanoma Program, NYU Langone’s Perlmutter Cancer Center, discussed the data in support of discontinuing treatment in melanoma and how to properly follow these patients after they go off treatment.

TARGETED ONCOLOGY:What data did you discuss that supports stopping checkpoint inhibitor treatment?

Weber:I gave a talk directed toward the issue of how long you should treat patients with checkpoint inhibition. This is a major patient-related convenience factor, toxicity factor, and it’s also a pharma-economical factor, as you could imagine. In the original trials, for example, of some of the drugs and certain histologies, for example pembrolizumab (Keytruda), nivolumab (Opdivo), atezolizumab (Tecentriq), often patients were treated until progression. For example, in melanoma, 15% to 20% of patients would potentially stay on [treatment] in remission, so you could stay on treatment for 3, 4, 5, or 6 years. In terms of the toxicity, convenience, and economics, that’s just not practical. A number of trials were then done at 96 weeks or 104 weeks, so at 2 years, you were just arbitrarily done with PD-1 inhibition, you stopped and went off and were followed.

The question then becomes, could you treat for a year and then stop? Should you treat based on the results of scanning, and then stop? I reviewed the available data, which are mostly retrospective, which gives us some hints to how long we should treat patients at least with melanoma and also probably other cancers like lung, bladder, and head and neck cancer.

The most data, as you might expect, come from melanoma, and if we go to the KEYNOTE-006 randomized [trial of] pembrolizumab versus ipilimumab (Yervoy), which was actually a 3-arm study [it compared] 2 schedules of pembrolizumab versus the then standard ipilimumab in frontline melanoma. When patients achieved a complete response (CR) or partial response (PR) and went off treatment—and there were over 100 patients in that category and of course there were patients who were stable, which was a small minority—those patients would stop at 2 years on schedule and were then followed. Those patients did very well. The rate of patients remaining in remission was something like 90%, so the progression-free survival (PFS) upon stopping after achieving stability, PR, or CR in melanoma was outstanding, meaning you could probably reassure your patients that at year 2, you’re good to stop and be followed. The likelihood of relapse is low and even if you do relapse, the rate of re-remission would be at least 50%. Now how long those remissions last, we don’t know, but the bottom line is you’re safe to stop at 2 years.

In the KEYNOTE-001 study, a lot of patients stopped because of toxicity, and when those patients stopped, again, 80% to 90% of them stayed in remission after stopping for toxicity, refusal, or other reasons. Again, that’s a very favorable number and is reassuring to the patients.

Now, that being said, in other trials, there are bits and pieces of data in lung cancer and bladder cancer trials that if you stop for toxicity, at least half if not more of those patients will stay in remission. Not quite as encouraging as melanoma, but not all of those histologies are as sensitive to checkpoint inhibition as melanoma is.

The one direct trial that evaluated prospectively the issue of duration of treatment with checkpoint inhibition was the lung cancer CheckMate 153 study where, unfortunately, the outcome in terms of PFS was inferior for those who received 1 year of therapy and stopped compared to those who were treated until progression. The PFS curves really diverged. They look very different. The overall survival curves are immature, but the Kaplan—Meier plot looks really discouraging in that it favors continuous treatment versus 1 year. If you’re a lung cancer patient, you really shouldn’t arbitrarily stop at 1 year.

If you look at the small retrospective trials where patients with melanoma received pembrolizumab or nivolumab, and then at 1 year had a PET/CT scan or a PET/CT scan and a biopsy, those who were metabolically or histologically negative and stopped at 1 year, even if they didn’t have a CR by CT scanning, did very well and had a 90% [or higher] chance at remaining in remission. What that does for me is that in melanoma, I would be comfortable treating for no less than 1 year, no more than 2 years, and if at 1 year a patient has had a stable response or has been stable for at least 2 scans, I would be very comfortable telling them to get a PET/CT scan and if there is complete metabolic regression on the PET, no matter what you see anatomically on the CT, I would say you could safely stop. I might say we will go 1 more cycle and then stop. Those patients, I think, could safely be observed knowing their likelihood of relapse is really low. In addition, they can be re-induced with a single-agent PD-1 blocking drug and have a high degree of getting put back into remission, although a duration of unknown length, and they will also have in their back pocket the chance to go on ipilimumab/nivolumab as we know from data presented by Georgina Long, BSc, PhD, MBBS, FRACP, at ASCO that there is a 40% to 50% chance of subsequent remission even after failing a single-agent PD-1.

Those are the guidelines I follow. Will we have prospective studies to evaluate them? It’s not clear, but I think there are so many pieces of data out there that that’s the way a lot of practicing oncologists will act in the future.

TARGETED ONCOLOGY:What does the follow-up look like after stopping treatment?

Weber:In a patient who stops treatment with a stable PR and a potential complete metabolic response of say 12 months, they get scanned every 3 months for 2 years. I might get a baseline MRI of the brain at the time they stop treatment. I will not get PET/CTs; actually, insurance will not cover that sort of thing in the US for surveillance, but I will get CTs focused on the area they had their original disease. If they had disease in the upper neck or upper chest, I’ll get a neck CT, chest, abdomen, pelvis. If it was anywhere else, it would be chest, abdomen, and pelvis. It’s a little controversial if they had disease in the limbs, but it depends. If the majority of the disease was in their right leg, I might scan the right thigh, but, generally, I will let the physical exam determine that.

Every 3 months, scanning for 2 years, then I’ll start to go to every 4 to 6 months for a couple of years, every 6 months for a couple of years, at year 6 I’ll declare victory and tell them you can either go to 1-year scan or you can stop scanning. A lot of patients don’t like that. They’re very insecure. They’re very insecure about stopping. I’ve had patients who refuse to stop say pembrolizumab on the original MK-3475 KEYNOTE-001 trial. But again, most of them will come around to the logic that with a stable response after 2 years, there’s no reason to keep treating.

I suppose that if someone had a continued regression, I would continue them until stable response, meaning no further regression, go 1 cycle and then stop.

TARGETED ONCOLOGY:Does stopping treatment give patients a better quality of life?

Weber: