Association of Anti–Programmed Cell Death 1 Antibody Treatment With Risk of Recurrence of Toxic Effects After Immune-Related Adverse Events of Ipilimumab in Patients With Metastatic Melanoma

Key Points

Question  Is anti–programmed cell death 1 antibody (PD-1) therapy associated with reduced risk of recurrence of toxic effects in patients with metastatic melanoma who experienced severe immune-related adverse events (AEs) after receipt of ipilimumab therapy?

Findings  In this cohort study of 56 patients with metastatic melanoma, all of whom experienced severe immune-related AEs after ipilimumab therapy, only 36% of patients experienced immune-related AEs after receiving anti–PD-1 therapy, including 21% of patients with grade 3 or 4 AEs.

Meaning  The findings suggest that anti–PD-1 therapy may be associated with reduced risk of toxic effects in this population.

Importance  Since 2011, many patients with metastatic melanoma have been treated with ipilimumab therapy and have developed severe immune-related adverse events (AEs). Because several immune therapies are now available to treat metastatic melanoma, a better knowledge of mechanisms and recurrence risks of immune-related AEs is needed before reintroduction of immunotherapies.

Objectives  To evaluate the risk of a recurrence of immune toxic effects associated with anti–programmed cell death 1 antibody (anti–PD-1) therapy after discontinuation of ipilimumab monotherapy because of severe AEs.

Design, Settings, and Participants  This cohort study conducted at 19 French melanoma referral centers included patients with metastatic melanoma who experienced severe immune-related AEs after ipilimumab therapy and then were treated with anti–PD-1 therapy between February 1, 2013, and December 31, 2016. The study cutoff was June 1, 2017. Statistical analysis was performed from June 1, 2016, to August 31, 2017.

Exposures  Monotherapy with at least 1 cycle of ipilimumab that was associated with a grade 3 or 4 immune-related AE and subsequent treatment with at least 1 cycle of an anti–PD-1 (nivolumab or pembrolizumab) therapy.

Main Outcomes and Measures  The primary outcome was the rate of immune-related AEs associated with anti–PD-1 therapy. Secondary outcomes were characteristics of ipilimumab-related and anti–PD-1 immune–related AEs and overall response rate and overall survival associated with anti–PD-1 therapy.

Results  Of 56 patients with metastatic melanoma included in the study, all of whom experienced severe immune-related AEs after ipilimumab therapy (31 [55%] male; mean [SD] age, 64 [14.9] years), 20 (36%) experienced at least 1 immune-related AE associated with pembrolizumab (6 of 20 [30%]) or nivolumab (14 of 20 [70%]) therapy. A total of 12 patients (21%) experienced grade 3 or 4 immune-related AEs, and among these patients, 4 (33%) presented with the same immune-related AE as with ipilimumab therapy. Severe immune-related AEs were resolved with use of systemic corticosteroids (7 [58%]) and/or anti–tumor necrosis factor (1 [8%]), and no grade 5 toxic effects were reported. Five patients discontinued anti–PD-1 therapy because of immune-related AEs. The overall response rate was 43%, with a median overall survival of 21 months (interquartile range, 18 to ongoing).

Conclusions and Relevance  The findings suggest that anti–PD-1 therapy may be associated with reduced risk of toxic effects and improved survival among patients who have experienced severe toxic effects after ipilimumab therapy.

Since 2011, the prognosis of metastatic melanoma has markedly changed with the discovery of immune checkpoint inhibitors. The anti–cytotoxic T-lymphocyte-associated protein 4 (anti–CTLA-4) antibody ipilimumab and anti–programmed cell death 1 (PD-1) antibodies have been associated with improved overall survival among patients with advanced melanoma.^1,2

Checkpoint inhibitors are associated with frequent specific adverse events (AEs), especially immune-related AEs. Severe immune-related AEs are less frequent after anti–PD-1 therapy than after ipilimumab therapy^3-5; however, patients with previous serious immune-related AEs during ipilimumab treatment are often excluded from phase 3 trials with anti–PD-1 therapy.^4,5 Thus, efficacy and safety of anti-PD-1 therapy were not evaluated in this population.

Herein, we collected data on AEs and survival from patients who experienced severe immune-related AEs during ipilimumab treatment and were treated secondarily with anti–PD-1 antibodies.

For this cohort study, eligible patients had an unresectable stage IIIC or IV melanoma, previously received ipilimumab monotherapy, experienced at least 1 grade 3 or 4 immune-related AE, and were then treated with anti–PD-1 antibodies between February 1, 2013, and December 31, 2016. This study was approved by the ethics committee of the University Hospital of Rennes, and for living patients, a written consent form was completed.

We retrospectively extracted data from January 2, 2013, to December 31, 2016, from 19 French melanoma centers. Baseline patient characteristics were collected, and immune-related AEs during ipilimumab and anti–PD-1 therapies were assessed according to the National Cancer Institute Common Toxicity Criteria for Adverse Events (CTCAE, version 4.03).⁶ Characteristics of immune-related AEs and immunosuppressive treatments were collected.

The response to ipilimumab and anti–PD-1 therapies when used in monotherapy was evaluated using the best objective response rate as defined by Response Evaluation Criteria In Solid Tumors (RECIST 1.1).⁶ Overall survival was defined as the time from anti–PD-1 therapy initiation to death. Patients lost to follow-up were censored at the date of the latest visit or study end point (June 1, 2017).

Data analysis was performed from June 1, 2016, to August 31, 2017. Survival curves were estimated using the Kaplan-Meier method. Frequencies were compared using the Fisher exact test and McNemar test, with 2-sided P < .05 as the significance level. Statistical analyses were performed using R, version 3.3.1 (R Foundation for Statistical Computing).

Of 56 patients included in the study, 53 (95%) had stage IV disease and a good Eastern Cooperative Oncologic Group performance status (<2), and 7 (13%) had brain metastases (Table 1). The BRAF mutation was found in 11 patients (20%), all of whom previously received anti-BRAF therapy. The median washout time between ipilimumab and anti–PD-1 therapies was 25 weeks (range, 2-194 weeks). Pembrolizumab and nivolumab were administered at standard doses (2 mg/kg and 3 mg/kg, respectively). At the time of analysis, median follow-up was 21 months (interquartile range [IQR], 18 to ongoing).

All patients had a previous severe immune-related AE during ipilimumab therapy (Table 2). Of 56 patients, 15 (27%) presented with 2 (23%) or 3 (4%) concomitant immune-related AEs. Among the 75 immune-related AEs reported, the most frequent AEs were gastrointestinal (35 [47%]; 34 with colitis) and endocrine (17 [23%]; 15 with hypophysitis). A total of 41 patients with immune-related AEs (73%) were treated with corticosteroids alone, 3 patients (5%) with corticosteroid and anti–tumor necrosis factor therapy, and 15 patients (27%) with substitutive endocrine therapy alone. All immune-related AEs were resolved at the time of anti–PD-1 initiation, including some endocrine immune-related AEs requiring long-term substitutive endocrine therapy (10 patients [18%]).

Twenty patients (36%) developed 23 immune-related AEs associated with anti–PD-1 therapy (Table 2). The greatest number of immune-related AEs were endocrine AEs (6 patients [11%]), followed by pulmonary (5 [9%]), gastrointestinal (4 [7%]), and cutaneous (4 [7%]) AEs. Five patients (9%) who previously experienced gastrointestinal (2 [4%]), endocrine (2 [4%]), and cutaneous (1 [2%]) immune-related AEs during ipilimumab therapy presented with interstitial pneumonia while undergoing anti–PD-1 therapy. Eight patients (14%) presented with the same immune-related AEs during anti–PD-1 therapy as during ipilimumab therapy.

Twelve patients (21%) experienced grade 3 (10 [18%]) or 4 (2 [4%]) immune-related AEs associated with anti–PD-1 therapy. Only 1 of 56 patients (2%) required anti–tumor necrosis factor therapy for grade 4 colitis. All other patients with severe immune-related AEs were successfully treated with systemic corticosteroids (7 [58%]) or endocrine therapies (6 [50%]) alone. Definitive treatment interruptions were required for 5 patients with immune-related AEs (9%). No treatment-related deaths occurred. Median time to occurrence of immune-related AEs was 49 days (IQR, 117 days), and median duration of follow-up was 49 days (1-407 days); all AEs were resolved at the time of data cutoff except 2 cutaneous (4%) and 2 endocrine (4%) immune-related AEs.

The overall response rate among patients receiving anti–PD-1 therapy was 43% (Table 3). Patients who responded to ipilimumab therapy were more likely to respond to anti–PD-1 therapy than were patients who did not (response rate, 88% vs 35%, P = .01). A total of 17 patients (35%) with no response to ipilimumab therapy responded to anti–PD-1 therapy (P < .001). No significant association was found between objective response and occurrence of immune-related AEs.

Median overall survival with anti-PD-1 was 21 months (IQR, 18 to ongoing). At the time of data analysis, 16 patients (29%) continued receiving anti–PD-1 therapy, with a median follow-up of 21 months (IQR, 0-37 months).

Management of systemic treatments for advanced melanoma varies among different countries of the world. However, worldwide, dermatologists are the first responders when patients develop immune-related AEs. With skin being one of the main sites of immune reactions, it is important that dermatologists understand the risk of recurrence of AEs after resumption of immunotherapy.

Prognosis of metastatic melanoma has markedly changed since the development of checkpoint inhibitors. Because of the exclusion of patients with previous severe immune-related AEs in pivotal clinical trials, there is a lack of data concerning risks of immune-related AE recurrence after a switch of checkpoint inhibitors.

A combination of nivolumab and ipilimumab therapies is currently the first-line treatment for patients with metastatic melanoma and is associated with a high rate of severe immune-related AEs.⁷ Data on immune-related AE recurrence when different immune therapies are reintroduced are needed, especially regarding the anti–PD-1 and anti–CTLA-4 combination therapy in the neoadjuvant context.⁸ As suggested in a recent study,⁹ anti–PD-1 therapy seems to be associated with a low rate of AEs, including severe immune-related AEs, during the combination treatment course.

This study suggests that anti–PD-1 therapy is associated with a risk of toxic effects in patients who have experienced severe AEs associated with anti-CTLA-4 therapy. To our knowledge, only 1 other study¹⁰ explored this association but with shorter follow-up. The authors of that study observed a high rate of immune-related AEs associated with anti–PD-1 therapy (34%), with a rate of recurrent immune-related AEs of only 3%. Overall, immune-related AEs were manageable, and no treatment-related deaths were reported.

Similar to results of Menzies et al,¹⁰ the rate of grade 3 and 4 immune-related AEs was higher in our study compared with 2 phase 3 clinical trials,^4,5 with severe AEs in 21% of patients in our study vs less than 15% of patients in the trials. Several explanations can be offered, such as a hyperactivation of the immune system after a previous activation by a checkpoint inhibitor or presence of a genetic variant associated with immune reactivity toward some organs.¹¹ However, in our study, we did not find an increase in the risk of immune-related AEs associated with anti–PD-1 therapy after occurrence of a severe immune-related AE during ipilimumab therapy. The proportion of immune-related AEs (any grade) in this study was 36%, which was similar to the rate of immune-related AEs observed in anti–PD-1 therapy clinical trials.^4,12 As described previously^4,5,13-15 and supported by the findings in our study, differences existed between ipilimumab and anti–PD-1 toxic effect profiles. Severe (grade 3 or 4) immune-related AEs associated with ipilimumab therapy were mainly colitis and hypophysitis and were frequently multiple (27%), whereas immune-related AEs associated with anti–PD-1 therapy were often unique and mainly cutaneous events and dysthyroidism. These results reflect the specific mode of action of anti–CTLA-4 and anti–PD-1 therapies.

Even severe immune-related AEs were manageable, with few consequences or treatment interruptions. The overall response rate and overall survival observed in our study were similar to those observed in anti–PD-1 therapy trials, with findings not suggesting an association with AEs.^4,12 Of interest, we observed a significant response to anti–PD-1 therapy among patients who had no response to ipilimumab therapy.

This study has limitations, including the retrospective collection of data, which could lead to missing information or misclassification of immune-related AEs. However, because of the previous severe toxic effects associated with ipilimumab therapy, patients were closely followed up after the introduction of anti–PD-1 therapy. Although our study population is one of the largest cohorts on the topic, the small size of the study population does not allow a high precision in estimation.

The findings suggest that anti–PD-1 therapy may be associated with reduced risk of toxic effects and improved survival among patients who have experienced severe toxic effects associated with ipilimumab therapy. In addition, response to anti–PD-1 therapy appears to be favorable.

Accepted for Publication: May 4, 2020.

Corresponding Author: Angélique Brunot, MD, Department of Medical Oncology, Comprehensive Cancer Center Eugène Marquis, Avenue de la Bataille Flandres Dunkerque, Rennes 35000, France (angeliquebrunot@hotmail.com).

Published Online: July 15, 2020. doi:10.1001/jamadermatol.2020.2149

Author Contributions: Dr Brunot had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Brunot, Prey, Campillo-Gimenez, Lesimple.

Acquisition, analysis, or interpretation of data: Brunot, Grob, Jeudy, Grange, Guillot, Kramkimel, Mortier, Le Corre, Aubin, Mansard, Lebbé, Blom, Montaudie, Giacchero, Prey, Legoupil, Guyot, Amini-adle, Granel-Brocard, Meyer, Dinulescu, Edeline, Campillo-Gimenez, Lesimple.

Drafting of the manuscript: Brunot, Aubin, Giacchero, Lesimple.

Critical revision of the manuscript for important intellectual content: Brunot, Grob, Jeudy, Grange, Guillot, Kramkimel, Mortier, Le Corre, Mansard, Lebbé, Blom, Montaudie, Prey, Legoupil, Guyot, Amini-adle, Granel-Brocard, Meyer, Dinulescu, Edeline, Campillo-Gimenez, Lesimple.

Statistical analysis: Brunot, Campillo-Gimenez.

Obtained funding: Granel-Brocard.

Administrative, technical, or material support: Brunot, Guillot, Aubin, Giacchero, Prey, Amini-adle.

Supervision: Brunot, Grob, Guillot, Mortier, Lebbé, Prey, Edeline, Lesimple.

Conflict of Interest Disclosures: Dr Lebbé reported serving on the speaker’s bureau for Roche, Novartis, Bristol-Myers Squibb, and Amgen; receiving research funding from Roche and Bristol-Myers Squibb; serving as an advisory board member for Roche, Amgen, Bristol-Myers Squibb, MSD, and Novartis; and receiving honoraria from Amgen, Bristol-Myers Squibb, Incyte, MSD, Novartis, Pfizer, Pierre Fabre, and Roche. Dr Edeline reported serving as an advisory board member for Bristol-Myers Squibb. Dr Lesimple reported serving on the speaker’s bureau for and receiving honoraria from MSD, Novartis, Pierre Fabre, and Roche, and receiving research funding from Roche. No other disclosures were reported.

Meeting Presentation: This work was presented in the poster abstract session at the 2016 American Society of Clinical Oncology Annual Meeting; June 4, 2016; Chicago, Illinois.

Additional Contributions: We thank the patients and their families and the French Society of Dermatology members. Odile Audrain, MD (Comprehensive Cancer Center Eugène Marquis, Rennes, France), provided assistance with reviewing the article without receiving financial compensation.