The prognostic impact of immune-related adverse events in real-world patients with metastatic melanoma treated with single-agent and combination immune checkpoint blockade.

9542 Background: Immune checkpoint blockade (ICB) has revolutionized the treatment of metastatic melanoma (MM). Immune-related Adverse Events (irAEs) associated with ICB have been shown to correlate positively with survival outcomes across solid tumours. In MM, conclusions on the impact of irAE severity have been conflicting, and combination ICB therapy experience is limited to smaller cohorts. We sought to clarify these relationships using the Alberta Immunotherapy Database (AID). Methods: The AID provides a multi-centre, province-wide observational cohort comprising consecutive patients treated with ICB. We included adult patients with MM, treated with ICB (single agent nivolumab or pembrolizumab, or combination ipilimumab and nivolumab) at any line of therapy, agnostic to site of origin, from August 2013 to May 2020, with analysis in December 2021. The primary endpoint of interest was the identification of a relationship between development of irAEs and subsequent overall survival (OS, defined from time of ICB initiation). To minimize immortal time bias from poor prognosis patients who may have died prior to the development of irAEs, patients who died before 12 weeks were excluded from OS and time-to-next-treatment (TTNT) analysis. Adjusted Cox regression analyses were performed to determine the association of variables with OS. Results: Of 492 MM patients receiving ICB, 124 received combination ICB, 198 developed an irAE and 67 required hospitalization for an irAE. irAEs were more common in patients < 50 years old (p = 0.02), with ECOG 0 (p < 0.001) and normal albumin (p = 0.002). Median time to irAE development (2.6 months) and frequency of individual irAEs were consistent with the published literature. In the overall population, patients who experienced an irAE had longer median OS (56.3 vs 18.5mo, p < 0.0001), and TTNT (49.6 vs 12.9mo, p < 0.0001). This remained consistent in combination ICB-treated patients (median OS 56.3 vs 19mo, p < 0.0001). Patients hospitalized for an irAE had improved OS and TTNT over patients requiring only outpatient treatment (median OS NR vs 27.9mo, p = 0.0039), while ICB re-challenge after an irAE also improved OS (56.3 vs 31.5mo, p = 0.0093). Development of an irAE retained independent association with OS after adjusted multivariable regression (HR 0.376, p < 0.001). Conclusions: These data support the association of irAEs and improved survival outcomes in MM, including those patients treated with combination ICB. Among patients with irAE, hospitalization for irAE, and ICB re-challenge post-irAE, were further associated with improved outcomes.