Stopping Immunotherapy After CR in the
'Real World'

Alexander M. Castellino, PhD

March 04, 2020

Clinical trial data showing sustained long-term outcomes for some patients with metastatic melanoma after immunotherapy achieved a complete response (CR) have led to much optimism. Some have described the change in prognosis as 'truly amazing.'

Although new 'real world' data from a single-institution retrospective analysis shows less promising results, it will not dampen that enthusiasm, say melanoma experts.

The new data come from the Memorial Sloan Kettering (MSK) Cancer Center in New York City from patients treated with pembrolizumab (Keytruda, Merck) or nivolumab (Opdivo, Bristol-Myers Squibb). The results show that not all patients who stop immunotherapy after achieving a CR maintain that response, and most relapse. Treatment failure was 27% at 3 years and most patients (75% to 85%) did not respond when they were retreated with immunotherapy following a relapse.

These results were published February 13 in the Journal of Clinical Oncology.

"Contrary to other studies, retreatment success was disappointing in our study of real-world patients. It is not obvious that retreatment will provide patients with the same benefits as the first course of therapy," MSK's melanoma expert Michael Postow, MD, a coauthor on the study, told Medscape Medical News.

"These are important data for community oncologists because they are likely more accurate than what will happen in clinical trials," he added.

" 'How long do I have to continue on therapy?' and 'Can I be retreated successfully if disease returns?' are questions that often come up with my patients in the clinic," Postow commented.

Previous data that address these questions come from clinical trials. "At MSK, we were in a good position to test whether this would hold true for real-world patients. We have been using these drugs for a long time and have a large dataset for patients treated outside of a clinical trial," he added.

Medscape Medical News reached out to melanoma experts not involved with this study for comments.

"These data support the practice that many of us carry out, and in the absence of a randomized trial, they are practice impacting," said Jeffrey S. Weber, MD, PhD, melanoma expert at NYU Langone Health in New York City. He was a coauthor on KEYNOTE-001, which provided the first evidence of retreatment success in metastatic melanoma using this approach.

"This study affirms prior results that anti-PD-1 therapy can be stopped in advanced melanoma patients who achieve a CR, but also cautions us that some patients may experience disease progression despite this," said Nikhil I. Khushalani, MD, from the H. Lee Moffitt Cancer Center, Tampa, Florida.

How Long to Treat After a CR Before Stopping Treatment

"The data across these studies supports that the majority of patients who achieve a CR with anti-PD-1 therapy have good outcomes after treatment discontinuation," Michael A. Davies, MD, PhD, from the MD Anderson Cancer Center, Houston, Texas, writes in an accompanying editorial.

"Data from KEYNOTE-001 and -006 have provided oncologists with clinical trial evidence to support the clinical practice of stopping treatment following a CR on immunotherapy," he told Medscape Medical News.

Results from the KEYNOTE-001 trial show that 94% (63 of 67) of patients who achieved a CR on pembrolizumab remained disease free 30 months after stopping treatment; 50% (2 of 4) of patients who relapsed regained responses after being retreated with pembrolizumab, Davies explained.

These data "were compelling and clinicians have been stopping treatment of patients who achieved a CR on immunotherapy quite routinely," Davies said.

He pointed out that the MSK cohort had striking differences from cohorts in other studies, such as when treatment was discontinued and how CR was defined.

"What is atypical about the MSK cohort is how little treatment patients received after a CR was achieved [median: 0 months]," Davies said. Across all clinical studies, before stopping treatment, an initial CR was confirmed following additional treatment that patients received, he explained.

All experts approached for comment agreed that treatment should continue for at least 6 months following documentation of a CR.

"This MSK study has somewhat limited applicability based upon its retrospective nature and the fact that patients who developed CR stopped treatment rather early," commented Alexander Menzies, MBBS, PhD, medical oncologist at the Melanoma Institute Australia in Sydney.

To support treating beyond achieving a CR, both Menzies and Davies noted that in a separate real-world international cohort, an increased risk for relapse was seen in patients who achieved a CR with less than 6 months of treatment compared with those who received more than 6 months of anti-PD-1 therapy (Ann Oncol. 2019; 30:1154-1161).

"In clinic we usually treat for 2 full years, but in those who have a CR and we/they want to stop, we recommend 1 year of treatment or at least continuing for another 3 months from time of CR," Menzies said.

"This MSK study is very important for routine 'real life' clinical practice to define some objective parameters on when one can consider stopping anti-PD-1 therapy for melanoma," Khushalani said.

Khushalani said that in his clinical practice, he endorses stopping anti-PD1 therapy for patients achieving a CR as long as they had received at least 6 months of treatment, based on data from KEYNOTE-001 and -006 trials. "I continue to use that paradigm in the clinic and the current results [from MSK] affirm validity of this practice," he said.

Weber suggested stopping therapy within a reasonable time after achieving a CR, perhaps 2 cycles of 12 weeks. "Many academics routinely do this, as have I," Weber said. "Any patient with a CR, especially those with a low disease burden, will be candidates for stopping therapy before 2 years if they have received therapy at least for 6 to 12 months," he added.

"I offer this approach to all patients who have achieved a CR and had at least 1 year of treatment, and less for select patients such as the elderly, patients with comorbidities, and those who experience toxicity," Menzies said.

"For those patients with a partial response, I typically continue therapy as long as subsequent serial imaging shows further improvement in disease burden as the time to achieve CR may be between 9 to 12 months from the start of therapy," Khushalani said.

Defining CR in Clinical Practice

In the MSK study, CR was defined by the absence of radiographic progression or biopsy that showed no viable melanoma after a radiographic pathologic response, or complete regression of clinically evaluable disease.

"I commend the MSK researchers for using definitions that are practical and applicable to community clinical practice," Khushalani said. It is unreasonable to expect that all imaging in cancer medicine will be interpreted according to RECIST (Response Evaluation Criteria in Solid Tumors) used in clinical trials, he added.

"In my clinical practice, I use resolution of visible radiographic disease (CT or PET [positron emission tomography]) as my definition for CR.  And if a patient has received at least 6 months of anti-PD-1 therapy while achieving this landmark, I discuss and recommend elective discontinuation of therapy," Khushalani elaborated. 

If the CT results are equivocal in defining CR, then adding a PET-CT to complement affirmation of CR is very appropriate, he said, adding that this approach would be an excellent use of resources with immense economic implications given the current cost of healthcare. 

"I always do a PET scan before stopping, to determine if in complete metabolic response," Menzies said.

Postow agreed and noted that a negative PET is a good prognostic indicator in this setting and may help confirm a CR in the absence of a biopsy.

Weber recommended CR should be evaluated based on RECIST criteria. "It clearly means disappearance clinically and radiologically of all disease," he said. "Resected lesions could also render someone free of disease."

Details of 'Real-World' MSK Retrospective Study

For their study, the MSK researchers reviewed 1325 records and focused on 396 evaluable patients who had discontinued anti-PD-1 therapy. These patients had been treated with single-agent pembrolizumab (85.9%) or nivolumab (14.1%).

Of these 396 patients, 102 (25.8%) were considered to have achieved a CR.

CR was based on complete radiographic resolution of disease (76 patients), and  confirmation by a formal RECIST was documented for 58 patients. The remainder were based on pathology (13 patients) and physical examination and radiology (13 patients).

The median duration of treatment to achieve a CR was 9.4 months; the median duration of treatment after CR was 0 months.

With a median follow-up of 21.1 months from time of CR, the median overall survival (OS) has not been reached. At 3 years at CR, 82.7% of patients are alive, including 72.1% who are alive without subsequent treatment.  

Of the 78 patients who discontinued anti-PD-1 therapy for any reason and then experienced treatment failure, 34 received further anti-PD-1 therapy and 44 received a combination of ipilimumab and nivolumab.

Retreatment success was 14.7% (5 of 34 patients) with anti-PD-1 therapy and 25% (11 of 44 patients) with the combination of ipilimumab and nivolumab.

"It is not obvious that retreatment will provide patients the same benefit as the first course of therapy," Postow concluded. He suggested that patients who relapse after stopping therapy should be enrolled into clinical trials that are evaluating various combinations.

"Overall, we do not have a great marker which can guide us on treatment decisions such as when to stop treatment, who to stop treatment in, and why to stop treatment," Davies said.

Do Not Extrapolate to Other Cancers

Khushalani cautioned that these data are specific to anti-PD-1 monotherapy in melanoma and should not be extrapolated to other classes of checkpoint inhibitors or to other disease types. 

"These data apply to pembrolizumab or nivolumab use in advanced melanoma," he noted. As the revolution of checkpoint inhibition first started with melanoma, the longest follow-up data are in this disease compared to other cancers. 

"I anticipate we will see similar analyses for other tumor types as well, but quite possibly with different results given the heterogeneity among tumors and differential sensitivity to immunotherapy," Khushalani told Medscape Medical News.

Postow receives honoraria from Bristol-Myers Squibb and Merck. He consults with or advises Bristol-Myers Squibb, Merck, Novartis, Array BioPharma, NewLink Genetics, Incyte, and Aduro Biotech. Some coauthors also report disclosures, as detailed in the original article. Davies reports relationships with GlaxoSmithKline, Genentech, Novartis, Sanofi, Vaccinex, Bristol-Myers Squibb, Syndax, NanoString Technologies, and Array BioPharma. Khushalani and Weber both report having relationships with numerous pharmaceutical companies . Menzies is on the ad visory board of Bristol-Myers Squibb, Merck Sharp & Dohme, Novartis, Roche, and Pierre-Fabre.

J Clin Oncol. Published February 13, 2020. Full text, Editorial

For more from Medscape Oncology, join us on Twitter and Facebook


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: