Abstract
Background: Immune checkpoint inhibition (ICI) is the standard treatment for advanced melanoma patients. Despite its high efficacy compared to previous treatment options, immune-related adverse events (irAEs) occur frequently. While most of the patients experience mild to moderate irAEs, some patients develop severe to lethal irAEs under ICI treatment; hence, biomarkers are urgently required. Methods: In this retrospective single-center study, 157 advanced melanoma patients treated with ICI at the University Medical Center Hamburg-Eppendorf were included. IrAEs were correlated with clinico-pathological parameters, disease-related outcomes, and irAE-free survival. Results: In our cohort, 130 out of 157 patients receiving immunotherapy experienced irAE, of which more than half experienced irAE Grade ≥ 3. The most common irAE independent of its grade included cutaneous irAE, colitis, endocrine irAE, and hepatitis. Patients experiencing irAE had significantly longer progression-free survival (PFS) and overall survival (OS) compared to patients who did not experience irAE under ICI therapy. Stratification by irAE groups revealed that musculoskeletal irAEs are associated with the longest, whereas myocarditis is associated with the shortest OS and PFS. IrAE was a significant beneficial prognosticator for PFS in univariate, but not in multivariate Cox regression analysis. With respect to OS, the occurrence of irAE was an independent prognostic factor among ECOG status ≥ 2 and uveal melanoma. ROC analysis demonstrated that D-dimers have moderate predictive capability for irAE occurrence. Cox regression analysis demonstrated that elevated D-dimers and PD-1 monotherapy vs. CTLA-4 and PD-1 combination regimen are the only independent prospective prognostic markers for irAE-free survival. Conclusions: Our study demonstrates that different irAE across the irAE spectrum have a different impact on the PFS and OS of advanced melanoma patients. D-dimers may be used as a blood-based biomarker for irAE prediction, warranting future validation in multi-center studies.