Routine [18F]2-fluoro-2-deoxy-d-glucose positron emission tomography (FDG-PET) may help predict clinical outcomes after response to immunotherapy. With a European Medicines Agency-recommended treatment length until disease progression or unacceptable toxicity, the optimal duration of immunotherapy remains to be defined. In a retrospective study, we retrieved from the Danish Metastatic Melanoma Database (DAMMED), all patients that were annotated as a partial or complete response based on the computed tomography (CT) of serial FDG-PET-CT scans. Patients treated with an anti-Programmed Death (PD)-1-containing regimen for <18 months, and ≥4 months without disease progression after halting anti-PD-1 were included. Cases were divided into an “elective” and a “toxicity” group based on the reason for treatment discontinuation. A total of 140 patients were included. At 29.3 months of median follow-up, a higher proportion of patients remained alive in the “elective” group (93% vs 75%, P =.0031) with an improved melanoma-specific (HR 0.07, 95% CI 0.02-0.32, P =.0041) survival (MSS). Patients without FDG-avid lesions at the time of treatment discontinuation had an improved MSS (HR 0.03, 95% CI 0.01-0.17, P =.0002), and the absence of FDG-avid lesions was the only independent predictive feature of improved MSS in multivariate analysis. In conclusion, patients with metastatic melanoma who obtain an early response and early discontinue immunotherapy have an excellent prognosis, especially in the absence of FDG-PET avid lesions when discontinuing treatment. These data support the option of early discontinuation, limiting possible overtreatment and thereby toxicity, health and economic expenses and improving logistics.