Optimal Duration of Checkpoint Inhibition in Melanoma Is No More Than 2 Years

Get Permission

For patients with advanced melanoma, the concept of treating to disease progression does not always apply. With many patients responding to checkpoint inhibition for years, when can treatment be safely discontinued? This important clinical question was addressed at the European Society for Medical Oncology (ESMO) 2018 Congress by Jeffrey S. Weber, MD, PhD, Deputy Director of the Laura and Isaac Perlmutter Cancer Center at NYU Langone Health, New York.1 Dr. Weber has spearheaded a number of the pivotal trials of checkpoint inhibition in melanoma.

“Treating to disease progression, as many of the original trials did, is now ridiculous. You could have patients on immunotherapy for years. Economically, it becomes unviable, it’s a stress on the infusion clinic, and it’s inconvenient—it’s absurd to ask patients to come in every 3 or 4 weeks for years. There’s also a significant risk of long-term toxicity. And, frankly, some patients just want to get on with their lives,” Dr. Weber said.

Indirect Data Guide Treatment Decision

Most of the data to guide this treatment decision are indirect, said Dr. Weber. Almost no trial has evaluated outcomes after stopping “at some arbitrary time,” but conclusions can be drawn from the following studies:

  • CheckMate 067 and KEYNOTE-001 in melanoma: Long-term follow-up of patients who discontinued checkpoint inhibitors early due to toxicity
  • CheckMate 037 in non–small cell lung cancer and KEYNOTE-001 and -006 in melanoma: Long-term follow-up of patients who finished on schedule after 2 years of checkpoint inhibition
  • CheckMate 153, which evaluated the duration of therapy: Long-term follow-up of patients with melanoma treated for 1 year versus those treated to disease progression
  • Data from position-emission tomography (PET) in partial or complete responders.

These trials provided no evidence that more than 2 years of checkpoint blockade is needed in metastatic melanoma, he said. “Patients who finish 2 years in complete response or even partial response have a very high long-term duration of remission,” he indicated.

“Also, patients with high-grade immune-related adverse events from single-agent checkpoint inhibition who stop treatment before 2 years do just as well as those who complete 2 years of therapy,” he added. “And patients receiving ipilimumab (Yervoy)/nivolumab (Opdivo) who stop treatment due to high-grade immune-related events and receive immune suppressants do as well as—or better than—those who complete 2 years of therapy.”

Show Me the Data

The definitive trial of front-line checkpoint inhibition was KEYNOTE-006, a three-armed trial of two doses and schedules of pembrolizumab (Keytruda) vs ipilimumab in immunotherapy-naive patients.2 The plan was for patients to be treated for approximately 2 years, which was achieved by about 20% of patients. In these patients completing 2 years of therapy, a complete response was observed in 27.2%, a partial response was seen in 63.1%, and stable disease was found in 9.7%. After 20 months of follow-up, 86% of the patients remained in remission. Patients whose disease progressed after stopping therapy and who received a second course of pembrolizumab also tended to enter remission again.

KEYNOTE-006 provides the best, though admittedly indirect, evidence that you don’t need to continue to treat patients forever who respond or are stable.
— Jeffrey S. Weber, MD, PhD

Tweet this quote

“The interesting thing is that of those 28 complete responders, 26 stayed in remission after almost 2 years of follow-up, as did 56 of the 65 partial responders. Even four of the seven patients with stable disease actually stayed in remission. To me, that’s remarkable,” Dr. Weber commented. KEYNOTE-006 provides “the best, though admittedly indirect, evidence that you don’t need to continue to treat patients forever who respond or are stable,” according to Dr. Weber.

Similar findings emerged from the KEYNOTE-001 trial, in which patients were treated an average of 2 years.3 Among the complete responders, 90% had prolonged remissions. This study also showed that the time to a complete response could be 1 year or longer.

CheckMate 067, which compared ipilimumab/nivolumab and the single agents in the front-line setting, also found that 90% of patients achieving a complete response after 2 years of treatment remained in remission. And among responders who stopped treatment early because of toxicity, 69% of the combination cohort, 70% of the nivolumab cohort, and 71% of the ipilimumab cohort also had ongoing responses.4 The average time for patients treated with ipilimumab/nivolumab to stay on therapy was about 2.5 months, but even among these patients, “two-thirds will stay in remission and do well,” he noted.

A pooled analysis of CheckMate 067 and CheckMate 069, involving 407 patients treated with ipilimumab/nivolumab, showed a 3-year progression-free survival of 39% and an overall survival rate of 58%.5 “I think everyone would agree you can have a nice plateau after getting no more than 2 years of ipilimumab/nivolumab. This curve is continuing to plateau at 4 years. You don’t need to continue ipilimumab/nivolumab forever,” he said.

Dr. Weber added that when his own patients on ipilimumab/nivolumab develop toxicity, which is usually in the first 12 weeks, he will continue to treat them with single-agent nivolumab. However, he acknowledged, this maintenance approach should be confirmed in a clinical trial.

The adjuvant therapy trials have also provided reassurance about stopping checkpoint blockade, he added: 1 year of adjuvant nivolumab in CheckMate 2386 and pembrolizumab in KEYNOTE-0547 seemed to be sufficient for improving outcomes. “It seems highly likely that you don’t need more than 1 year’s worth of adjuvant therapy,” he concluded. “After this, patients probably will plateau out at about 60%, and it’s unlikely that giving more therapy would do them any good.”

The bottom line, he said, is that patients who complete 2 years of pembrolizumab or nivolumab and respond to it should continue to have good outcomes for at least 2 more years. And patients who need to stop combination therapy early, for whatever reason, will probably also do “extremely well” over time.

“So, you can tell patients, ‘Yes, you can stop after 2 years; you’ll do very well, and even if your disease progresses, you can be re-treated (usually with the single agent) and still have a decent chance of being put back in remission,’” he said.

Role of PET-CT Scanning

“There is an emerging consensus that you should probably use PET scanning for your patients when they’ve achieved response or what you think is a maximal benefit,” he said.

Researchers at the Melanoma Institute of Australia evaluated whether PET scans could predict relapse among 114 patients with a complete response by Response Evaluation Criteria in Solid Tumors (28%) and/or a complete metabolic response by PET (75%; unpublished data). They found that patients achieving a PET metabolic complete response, but not an anatomic complete response, had good outcomes. Approximately 95% of patients were progression-free at 4 years, compared with a little over 50% who did not achieve this benchmark (hazard ratio [HR] = 0.07; P < .01). The median progression-free survival was not reached in either group. Patients with neither an anatomic nor a metabolic complete response did not have favorable outcomes.

I recommend at least 1 to 1.5 years of therapy or best response plus 1 cycle, whichever is longer, and not more than 2 years of checkpoint inhibition.
— Jeffrey S. Weber, MD, PhD

Tweet this quote

These data are echoed by a study from Georgetown University, which evaluated outcomes in patients with either a negative PET scan or a negative tumor biopsy after stopping treatment; they found good outcomes at 3 years.8 Also, an “impressive” 80% of patients who discontinued treatment because of toxicity stayed in remission.

“If a responding patient has plateaued after 1 year on nivolumab, pembrolizumab, or a combination and has a negative PET scan, I am comfortable encouraging the patient to stop treatment,” he said. “If you have a questionable PET, you do a biopsy. If it’s negative, you can probably safely stop treatment, and the patient will most likely stay in remission.”

Dr. Weber offered this ultimate piece of advice: “I recommend at least 1 to 1.5 years of therapy or best response plus 1 cycle, whichever is longer, and not more than 2 years of checkpoint inhibition.” 

DISCLOSURE: Dr. Weber has stock or other ownership in Altar Bioscience, Biond, and CytomX Therapeutics and has received honoraria or served as a consultant/advisor for Bristol-Myers Squibb, Merck, Genentech, AbbVie, AstraZeneca, Daiichi Sankyo, GlaxoSmithKline, Eisai, Altor Bioscience, Lion Biotechnologies, Amgen, Roche, Ichor Medical Systems, Celldex Therapeutics, CytomX Therapeutics, Nektar, Novartis, Medivation, and Sellas.


1. Weber JS: The right time for discontinuing immunotherapy. 2018 ESMO Congress. Presented October 21, 2018.

2. Long GV, Schachter J, Ribas A, et al: 4-Year survival and outcomes after cessation of pembrolizumab after 2 years in patients with ipilimumab-naive advanced melanoma in KEYNOTE-006. 2018 ASCO Annual Meeting. Abstract 9503. Presented June 4, 2018.

3. Hamid O, Robert C, Daud A, et al: 5-Year survival outcomes in patients with advanced melanoma treated with pembrolizumab in KEYNOTE-001. 2018 ASCO Annual Meeting. Abstract 9516. Presented June 4, 2018.

4. Larkin J, Chiarion-Sileni V, Gonzalez R, et al: Overall survival results from a phase III trial of nivolumab combined with ipilimumab in treatment-naive patients with advanced melanoma (CheckMate 067). 2017 AACR Annual Meeting. Abstract CT075. Presented April 3, 2017.

5. Wolchok JD, Chiarion-Sileni V, Gonzalez R, et al: Overall survival with combined nivolumab and ipilimumab in advanced melanoma. N Engl J Med 377:1345-1356, 2017.

6. Weber JS, Mandalà M, Del Vecchio M, et al: Adjuvant therapy with nivolumab versus ipilimumab after complete resection of stage III/IV melanoma: Updated results from a phase III trial (CheckMate 238). 2018 ASCO Annual Meeting. Abstract 9502. Presented June 4, 2018.

7. Eggermont AMM, Blank CU, Mandala M, et al: Adjuvant pembrolizumab versus placebo in resected stage III melanoma. N Engl J Med 378:1789-1801, 2018.

8. Christiansen SA, Swoboda D, Gardner K, et al: Off treatment survival in patients with advanced melanoma after anti-PD1 therapy. 2018 ASCO Annual Meeting. Abstract 9554. Presented June 4, 2018.