Re-treating Melanoma After PD-1 Blockade Shows Good Response

This transcript has been edited for clarity.

Hello. This is Dr Jeffrey Weber. I'm a medical oncologist at the Laura and Isaac Perlmutter Cancer Center at New York University Langone Health in New York City.

I'd like to report on an interesting article in the Journal of Clinical Oncology by Allison Betof Warner and colleagues, describing a single-institution experience, from the outstanding Memorial Sloan Kettering Cancer Center, of patients with metastatic melanoma who received single-agent programmed death-1 (PD-1) blockade. The article describes the outcome in those mostly nonprotocol patients and, importantly, also talks about what happens when a significant proportion of them progressed and went on to receive subsequent therapy.

A very important question for academicians and practicing oncologists is: What do you do if a patient has had single-agent PD-1 blockade and does well for a period of time but then progresses and requires second-line therapy? What is the best choice? The natural history of this disease after PD-1 blockade is a very important issue, and that is what this article addressed.

There were 396 patients in total; 69% of them were not on a trial, so this was a reasonable, real-world experience. Twenty-five percent had a complete response (CR), of which about 70% were subsequently confirmed on independent radiologic review. (These were called by the investigator; many of them had trivial disease and, in the real world, were probably complete responders.) Something like 14% had CNS metastases and 40% had M1b disease. Overall survival was excellent. The median survival was 39 months, which is about what you would see in frontline therapy in clinical trials of PD-1 blockade alone. The majority had pembrolizumab, but a significant minority had nivolumab — very similar drugs. The likelihood of remaining in remission at 3 years was over 70% if you had a CR, and the [estimated] survival of the complete responders at 3 years was 83%. Something like 73% of the complete responders did not require any further therapy by year 3. The median duration of therapy [of the complete responders] was 9.4 months, so a lot of them stopped before then. The majority of patients who went on to other therapy stopped because of toxicity or progression.

Tumor mutational burden was definitely associated with CR in a multivariable analysis. Stage M1b compared with M1c was associated with complete response. (There were very few M1a's.) For negative associations, high lactate dehydrogenase was associated with a poor outcome and not achieving a CR, as was having mucosal or acral lentiginous melanoma. Interestingly, in contrast to prior reports of the importance of body mass index (BMI), there was no association in men or women between BMI and outcome.

Of the 102 complete responders, 23 did progress and presumably required further therapy.

A total of 78 patients were described who did require further therapy after PD-1 blockade failed. Of the 34 patients who [received further therapy with] single-agent PD-1 blockade alone, there were only five responders (15% response rate). Not very impressive. Of those who then went on to receive combination ipilimumab and nivolumab, there was only a 25% response rate (11 out of 44 patients). So again, disappointing but not a surprise. In the 78 patients who were re-treated, 13 patients had stable disease and 37 patients had progressive disease. This is different from prior reports discussing how patients do well after having a CR or a partial response and then get re-treated. Here, it's not just the complete responders who were re-treated; it was those who had stable disease or progressed. The average amount of time between treatments was 6.3 months.

What does this tell us? To recap, it suggests that those patients who get to CR (again, not on a protocol setting) do very well. Eighty-two percent of them are alive at 3 years, and about 73% of them at 3 years did not require other therapy. But again, if patients progress or are progressing on single-agent PD-1 blockade, the likelihood of re-responding to PD-1 blockade is not high. In my view, it sets us up to think that combination checkpoint blockade may be the best way to go subsequently.

This is Dr Jeffrey Weber. Thank you for your attention.

Jeffrey S. Weber, MD, PhD, is deputy director of the Laura and Isaac Perlmutter Cancer Center at NYU Langone Medical Center and co-director of its Melanoma Research Program.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube