The reported prevalence of nevus-associated melanoma varies substantially. We performed a systematic review and meta-analysis to determine the incidence and prevalence of this disease; we also performed subanalyses considering age, tumor thickness, and nevus-type classification. In 38 observational cohort and case-control studies, 29.1% of melanomas likely arose from a preexisting nevus and 70.9% de novo. Any given melanoma was 64% less likely to be nevus-associated than de novo (risk ratio 0.36, 95% confidence interval [CI] 0.29-0.44; P < .001; I = 99%); nevus-associated melanomas had a lower mean Breslow thickness than de novo melanomas (mean difference -0.39 mm; 95% CI -0.60 to -0.18; P = .0003; I = 66%). No significant differences were noted regarding the association of nevus-associated melanomas with nondysplastic nevi or dysplastic nevi (risk ratio 0.77, 95% CI 0.49-1.20; P = .24; I = 98%).