Long-term safety of pembrolizumab monotherapy and relationship with clinical outcome: A landmark analysis in patients with advanced melanoma

Objective: Long-term safety of pembrolizumab in melanoma was analyzed in KEYNOTE-001, KEYNOTE-002, and KEYNOTE-006. Patients and methods: Analysis involved patients who received >= 1 pembrolizumab dose. Leadtime bias was addressed via landmark analyses in patients who were progression-free before day 147. Results: Adverse events (AEs) were analyzed for 1567 patients (median follow-up, 42.4 months). Most AEs were mild/moderate; grade 3/4 treatment-related AEs occurred in 17.7% of patients. Two pembrolizumab-related deaths occurred. Any-grade immune-mediated AEs (imAEs) occurred in 23.0%, most commonly hypothyroidism (9.1%), pneumonitis (3.3%), and hyperthyroidism (3.0%); grade 3/4 imAEs occurred in 6.9% of patients. Most imAEs occurred within 16 weeks of treatment. In landmark analysis, patients who did (n=79) versus did not (nZ384) develop imAEs had similar objective response rates (ORRs) (64.6% versus 63.0%); median time to response (TTR), 5.6 months for both; median duration of response (DOR), 20.0 versus 25.3 months; median progression-free survival (PFS), 17.0 versus 17.7 months; median overall survival (OS), not reached (NR) versus 43 months ( p=0.1104). Patients who did (n=17) versus did not (n=62) receive systemic corticosteroids had similar ORRs(70.6% vs. 62.9%) and median TTR(6.4 vs. 5.6months) butnumerically shortermedianPFS(9.9 vs. 17.0 months); median DOR, 14.2 months versus NR; median OS, NR for both. Conclusions: These results enhance the knowledge base for pembrolizumab in advanced melanoma, with no new toxicity signals after lengthy follow-up of a large population. In landmark analyses, pembrolizumab efficacy was similar regardless of imAEs or systemic corticosteroid use. (C) 2020 The Author(s). Published by Elsevier Ltd.