Discontinuing Immunotherapy: Is the Infusion Bag Half Empty or Half Full?

H. Jack West, MD


November 21, 2022

It's a "champagne problem" many of us have encountered over the past few years in the clinic.

A patient with advanced non–small cell lung cancer (NSCLC) is fortunate enough to continue to do well for 2 years on ongoing pembrolizumab or perhaps pemetrexed and pembrolizumab as maintenance therapy. The latest CT shows a residual but far smaller primary tumor than what she started with.

In this instance, you may be considering stopping treatment but are concerned about doing so with evidence of disease still present.

Clinical trials of immunotherapy or chemoimmunotherapy have generally terminated treatment in nonprogressing patients after 2 years. We also know that some patients in early trials of immunotherapy stopped treatment after a fixed period of 1 or 2 years and continued to show no evidence of progression many years later.

The reason some patients experience this kind of success: Unlike the mechanism of action of conventional chemotherapy or targeted therapies, where ongoing treatment would be important to continue to exert an inhibitory effect, the active substrate of immunotherapy is the patient's immune system, which can potentially have a self-sustaining efficacy beyond the stimulatory effect of the checkpoint inhibitor.

Though we don't want to overtreat our patients, the question remains: Will patients do just as well off treatment, with the potential to resume as needed?

One trial directly addressed this question of stopping vs continuing treatment in patients on immunotherapy. The CheckMate 153 trial, published in 2020, randomly assigned 252 previously treated patients who hadn't demonstrated progression after 1 year on nivolumab to either discontinue nivolumab or continue nivolumab on an ongoing basis. The results were strongly in favor of ongoing therapy. Both progression-free survival (PFS) and overall survival (OS) were significantly longer in patients who continued therapy: PFS of 24.7 months vs 9.4 months and OS not reached vs 32.5 months.

This finding is important, but there's an important caveat. The study population included many heavily pretreated patients, but, in practice, immunotherapy has generally moved into the first-line setting, where we see dramatic responses in a significant subset of patients.

Even more recent data are emerging that may help us evaluate who will do well off therapy and who should continue treatment.

We now have a growing collection of long-term data on patients who are more likely to have good outcomes with immunotherapy, specifically those with high tumor programmed death-ligand 1 (PD-L1) expression (≥ 50%), from the KEYNOTE-024 trial. In this study, 39 of 151 (25.8%) patients assigned to pembrolizumab completed the planned maximum of 2 years of treatment, among whom, 82.1% achieved an objective response; but, only 10% (4 patients) achieved a complete response. The proportion of patients without progression and remaining off therapy wasn't reported, but the OS rate 3 years after completing treatment was 81.4%.

In addition, restarting immunotherapy after discontinuing appears to be a moderately effective strategy. In the KEYNOTE-024 trial, 12 patients received a second course of pembrolizumab because of disease progression a median of 15.2 months after discontinuing pembrolizumab. In this small cohort, eight of these patients (66.7%) were alive at the data cutoff, and six (50%) achieved stable disease.

Recently, we received additional insight in the follow-up from two chemoimmunotherapy trials that have most shaped my practice for patients with advanced NSCLC and any level of PD-L1 expression. These are the KEYNOTE-189 trial of platinum-pemetrexed with pembrolizumab vs placebo in those with nonsquamous NSCLC, and the KEYNOTE-407 trial of carboplatin-taxane with pembrolizumab vs placebo in patients with advanced squamous NSCLC. The National Comprehensive Cancer Network has designated each as a "preferred regimen" for patients with advanced NSCLC.

Both regimens have demonstrated sustained efficacy benefits with prolonged follow-up, including significantly superior objective response rate, PFS, and OS with the addition of pembrolizumab. These findings merely cemented the role of these regimens in our practice, but the trials also reported on the cohort of patients who completed 35 cycles of treatment over 2 years then discontinued therapy. In both, the majority of patients showed an objective response (86% in KEYNOTE-189 and 90% in KEYNOTE-407), with most patients alive at 3 years after 2 years of treatment (71.9% in KEYNOTE-189 and 69.5% in KEYNOTE-407). In addition, the proportion of patients alive without disease progression or subsequent therapy was notable — 40.4% in KEYNOTE-189 and 43.6% KEYNOTE-407.

How should we interpret these data for the patient who is in the exam room with us?

The short answer is that we don't know. I see this as a half-empty, half-full conundrum.

I'm disappointed that more patients who responded for 2 years will experience disease progression in the 1-3 years that follow. This signals that their immune systems have not perpetuated their initial response over the long-term. But these patients may have demonstrated disease progression even if they had continued therapy.

We also know that some patients can be rechallenged and will respond again. Some of these patients will show stable disease, whereas others will progress with repeat treatment. I would love to be able to better predict which patients are destined to do well without treatment vs those who benefit from treatment beyond 2 years.

Might the level of PD-L1 expression tell us? Can PET imaging discriminate those with residual hypermetabolism who may need continued treatment from those with no residual uptake who could be spared it? Would serial measurement of circulating tumor DNA (ctDNA) in responding patients identify when they have achieved a point of diminishing returns, potentially indicating that some can safely discontinue treatment after 2 years, whereas others need to continue to suppress on prolonged maintenance therapy?

These questions have yet to be studied systematically. In the meantime, I take an individualized approach with my patients facing this decision. Some have experienced escalating arthralgias and myalgias, cost concerns, or other issues related to immunotherapy that may dissuade us from continuing treatment. But several others have been grateful to continue with their treatment, hesitant to do anything that could change the path of their disease.

In my patients who tolerate therapy well, I'm more worried about potential undertreatment than overtreatment. I tend to favor having my patients continue therapy in the absence of problematic toxicity or practical challenges. There is certainly room for debate here while we await data to better guide these decisions. How do you approach these patients?

H. Jack West, MD, associate clinical professor and executive director of employer services at City of Hope Comprehensive Cancer Center in Duarte, California, regularly comments on lung cancer for Medscape. Dr West serves as web editor for JAMA Oncology, edits and writes several sections on lung cancer for UpToDate, and leads a wide range of continuing education programs and other educational programs, including hosting the audio podcast West Wind.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


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.