Immune checkpoint inhibitors (ICIs) have tremendously changed the therapeutic landscape of melanoma since they are associated with a durable response, allowing for intentional discontinuation of therapy after complete or partial remission. However, a subset of patients develops a relapse after cessation of ICI treatment and may not respond to reinduction of ICIs. The aim of the present study was to identify risk factors for relapse after intentional discontinuation of ICI therapy. Patients with intentional discontinuation of ICI therapy for metastatic or unresectable melanoma from 5 German university hospitals were analyzed retrospectively. Clinicopathologic and follow-up data of 87 patients were collected and analyzed by univariate and multivariate Cox proportional-hazards models. The following parameters were associated with relapse after cessation of ICI treatment in the univariate Cox regression analysis: concurrent radiotherapy and ICI, best overall response, and presence of brain metastases. Duration of treatment, type of primary tumor, body mass index, programmed-death ligand 1 expression, and lactate dehydrogenase levels did not significantly influence the risk for relapse. In the multivariate analysis, partial remission [hazard ratio 4.217 (95% confidence interval: 1.424−12.49), P = 0.009] and stable disease [3.327 (1.204−9.19), P = 0.02] were associated with a significant decrease in progression-free survival compared with complete remission. Concurrent radiotherapy and ), P = 0.015] are additional independent risk factors for decreased progression-free survival upon ICI discontinuation, whereas the presence of brain metastasis did not reach statistical significance on multivariate analysis.