Abstract

Abstract Sex differences in melanoma are prominent, with females having a significant survival advantage. However, it is unclear why we see this survival advantage. Here we investigate the relationship between sex, clinicopathologic variables, and melanoma specific survival in 1,753 single primary melanomas from patients in the GEM study. Using Cox proportional hazard models and formal mediation analysis, the effect of sex on survival is explained largely by differences in the clinicopathologic features of tumors at diagnosis. Specifically, we find evidence that 86.5% of the effect of sex on melanoma survival is mediated by differences in age at diagnosis, Breslow thickness, ulceration, mitoses and site (HR 1.85, P < .001). This analysis indicates that the female survival advantage in melanoma is not due primarily to a direct effect of sex (HR 1.19, P = .42) but is largely a result of an indirect effect of sex mediated by clinicopathologic features.