Balancing Immune-Related Adverse Events With Efficacy of Dual Checkpoint Inhibitor Therapy in Melanoma

Get Permission

Alexander N. Shoushtari, MD

Alexander N. Shoushtari, MD

MUCH PROGRESS has been made in the past decade in the treatment of patients with newly diagnosed metastatic melanoma. In the front-line setting, programmed cell death protein 1 (PD-1) monotherapy (nivolumab [Opdivo] and pembrolizumab [Keytruda]) and combined PD-1 plus cytotoxic T-lymphocyte–associated protein 4 (CTLA-4) inhibition with nivolumab plus ipilimumab (Yervoy) are approved by the U.S. Food and Drug Administration (FDA), and patients with BRAF V600–mutant melanoma can achieve a durable benefit with BRAF plus MEK inhibition. Physicians and patients now have the relatively new luxury of choosing between effective therapies, and to do so, it is imperative to better understand the relative burden of adverse events. 

The phase II and III trials of nivolumab plus ipilimumab demonstrated impressive objective response and progression-free survival improvements over ipilimumab and nivolumab monotherapy, albeit at the cost of higher rates of grade 3 and higher adverse events. For most patients and physicians, however, these data are difficult to grasp and place into routine clinical context. Common Terminology Criteria for Adverse Events (CTCAE) were not developed for immune-related adverse events, so some grade 3 events, such as lipase elevation, are infrequently significant,1 whereas some grade 1 or 2 events, such as neuropathy and pneumonitis, can be harbingers of progressive, even life-threatening, autoimmune events. 

Weighing Adverse Events and Clinical Benefit 

UNDERSTANDING THE BURDEN of immune-related adverse events is particularly relevant because the phase III trial of nivolumab plus ipilimumab was not powered to demonstrate an overall survival benefit over nivolumab alone. This leaves no simple answer to the patient who asks us, “What is the best treatment for me?” To balance the potential for profound systemic adverse events with durable clinical benefit, we must understand whether, for those with significant side effects, we are compromising efficacy to hold or stop therapy. 

Recently, as reviewed in the October 25 issue of The ASCO Post, Schadendorf et al2 analyzed the safety and survival data among approximately 400 patients with melanoma enrolled on the phase II and III trials of nivolumab at 1 mg/kg plus ipilimumab at 3 mg/ kg. They reported that one-quarter of patients discontinued therapy for toxicity during the induction phase of nivolumab plus ipilimumab and that diarrhea/colitis and hepatitis were the two most common reasons. The median number of doses was three in this subgroup. 


  • Overall, 24% of patients with advanced melanoma discontinued treatment during induction due to adverse events.
  • No significant differences in response rate, progression-free survival, or overall survival were observed among patients discontinuing treatment during induction due to adverse events vs those not discontinuing treatment.
  • The authors concluded that more evidence is needed to determine whether patients should discontinue therapy after they have demonstrated a clinical response and to determine the optimal duration of treatment required to reach a maximum response.

Reassuringly, these patients had similar rates of objective response and maintained similar progression-free and overall survival rates compared with those who did not discontinue therapy for adverse events, despite an approximately 8-month longer median time on therapy for the latter group. The vast majority of nonendocrine adverse events were reversible with corticosteroid therapy. This offers our strongest evidence to date that physicians should not blindly adhere to the schedule and dose number if patients develop severe adverse events. Physicians should adhere to systemic immune suppression guidelines and not risk worsening immune-related organ dysfunction in an effort to maintain dose intensity. In fact, the notion of maintaining “dose intensity” as it relates to the prespecified schedule may be misguided; instead, we may be better served asking how many doses each individual patient’s “immunostat” can tolerate before we have disinhibited the system too thoroughly. 

With particularly effective therapies, the eventual goal is durable disease control, preferably off active therapy, and longer follow-up will be needed to clarify any differences between progression-free and overall survival in these patients. Schadendorf et al reported a numerically lower 18-month progression-free survival rate for patients who discontinue treatment during induction vs those who never discontinue therapy (38% vs 49%) but a numerically higher 18-month overall survival rate (67% vs 62%), although neither is statistically significant. Notably, patients who discontinue therapy during induction may represent a group with lower-risk disease, having lower rates of M1c metastases and elevated lactate dehydrogenase levels. One may reasonably hypothesize that these patients’ immune systems may be too disinhibited with dual-agent therapy, disproportionately require corticosteroids, and may effectively re-engage the tumor upon repeat (single -gent anti–PD-1) therapy if and when disease progresses. Further study and follow-up are required to optimize patient selection. 

Watching Out for ‘Blind Spots’ 

THIS RETROSPECTIVE ANALYSIS of large randomized data is crucial in guiding our therapy, but randomized trials still have limitations in data collection and reporting that lead to “blind spots” when we consider how to care for patients. For example, patients on these trials who discontinued therapy were typically followed for adverse events for 16 weeks, after which time delayed immune-related adverse events would not be recorded. The pleiotropic nature of immune-related adverse events often makes outpatient triage difficult and requires the patient to present to the emergency room for medical attention far more often than do cytotoxic agents with more well-defined adverse events. Empiric corticosteroid use has significant interactions with cardiac and endocrine comorbidities, which are common in our patients. Most clinical trials, however, do not routinely report how often these scenarios arise. 

Broader Definition of Toxicity 

OUR GROUP recently studied a small prospective cohort of 64 patients treated with nivolumab plus ipilimumab as part of an expanded access protocol.3 Although this study had more immature follow-up than did the study by Schadendorf et al, we similarly found that among those free of disease progression after 12 weeks, patients who received therapeutic modifications for toxicity obtained similar 1-year freedom from treatment failure as those who did not have therapy modified for toxicity. 

“Our patients will go on living their lives, and we owe it to them to better understand how any perturbation of their immune system affects their ability to do so.”
— Alexander N. Shoushtari, MD

Tweet this quote

In our cohort, anti–PD-1 monotherapy off-trial was routinely available, and our treatment discontinuation rate during induction was higher than that of the phase II and III trials: 31 patients (48%) received fewer than 4 doses due to toxicity. Of them, 17 patients (54%) resumed anti–PD-1 monotherapy, which would not have been allowed on a phase II or III trial, likely reflecting the flexibility that physicians have in current clinical practice to “drop” the ipilimumab. Among the 31 patients who received fewer than 4 doses of induction therapy, 4 (13%) developed an immune-related adverse event > 16 weeks after treatment discontinuation, including cases of sarcoidosis and type 1 diabetes. These adverse events are not captured by the randomized trial data and suggest longer safety follow-up is required for this subset of patients with a severe reaction to initial induction therapy. The immune-related adverse event burden placed on patients is significant; 50% visited the emergency department, and 36% were hospitalized due to an immune-related adverse event. Immune suppression is the rule, rather than the exception: 73% required oral corticosteroids, and 25% required secondary immune suppression with infliximab (Remicade) or mycophenolate. 

Taken in this context, it is almost an understatement to simply quote the grade 3 to 4 adverse event rate, which was 59% both in our small cohort and in the phase III trial,4 when patients inquire about side effects. We thus argued that a broader, “clinically significant” adverse events definition was more appropriate. We defined clinically significant adverse events as grade 3 or higher or grade 1 to 2 for which oral corticosteroids or dose delays/ discontinuation are required. In our cohort, 91% experienced at least one clinically significant immune-related adverse event. This framework may improve our ability to communicate how immune-related adverse events may impact a patient’s life. 

More to Learn About Long-Term Effects 

OVERALL, we have made significant progress in developing checkpoint inhibitors for the treatment of advanced melanoma. We have discovered that a significant subgroup of patients will require dose modifications and immune suppression to protect them from autoimmune toxicity and that this will not compromise short-term and intermediate efficacy. Longer follow-up is required to understand long-term efficacy and safety outcomes in these groups of patients. 

In melanoma, important work is ongoing to investigate dosing modifications and ‘adaptive dosing’ strategies tailored to an individual’s early response to therapy. Future trials across oncology utilizing novel immune-modulating agents should prospectively report broader toxicity outcomes beyond CTCAE grade ≥ 3 immune-related adverse events, including the rates of emergency room visits, hospitalization, and use of corticosteroids and secondary immune suppression. 

Finally, we must acknowledge how little we understand the long-term effects of all immune-modulating agents. As the efficacy of these agents improves, our patients will receive additional antineoplastic and other medications and interventions; they will accumulate additional medical problems beyond their cancer. They may opt to have children, resume working, or travel the world. In short, they will go on living their lives, and we owe it to them to better understand how any perturbation of their immune system affects their ability to do so. ■

DISCLOSURE: Dr. Shoushtari is on advisory boards for Bristol-Myers Squibb, Castle Biosciences, Immunocore, and Vaccinex. 


1. Friedman CF, Clark V, Raikhel AV, et al: Thinking critically about classifying adverse events: Incidence of pancreatitis in patients treated with nivolumab + ipilimumab. J Natl Cancer Inst 109:djw260, 2016

2. Schadendorf D, Wolchok JD, Hodi FS, et al: Efficacy and safety outcomes in patients with advanced melanoma who discontinued treatment with nivolumab and ipilimumab because of adverse events: A pooled analysis of randomized phase II and III trials. J Clin Oncol. August 25, 2017 (early release online). 

3. Shoushtari AN, Friedman CF, Navid-Azarbaijani P, et al: Measuring toxic effects and time to treatment failure for nivolumab plus ipilimumab in melanoma. JAMA Oncol. August 17, 2017 (early release online). 

4. 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.