Gender Disparity and Mutation Burden in Metastatic Melanoma

Abstract

Gender differences in melanoma incidence and outcome have been consistently observed but remain biologically unexplained. We hypothesized that tumors are genetically distinct between men and women and analyzed the mutation spectra in 266 metastatic melanomas (102 women and 164 men) from The Cancer Genome Atlas (TCGA). We found a statistically significantly greater burden of missense mutations among men (male median 298 vs female median = 211.5; male-to-female ratio [M:F] = 1.85, 95% confidence interval [CI] = 1.44 to 2.39). We validated these initial findings using available data from a separate melanoma exome cohort (n = 95) and found a similar increase in missense mutations among men (male median 393 vs female median 259; M:F = 1.59, 95% CI = 1.12 to 2.27). In addition, we found improved survival with increasing log-transformed missense mutation count (univariate hazard ratio = 0.82, 95% CI = 0.69 to 0.98) for TCGA samples. Our analyses demonstrate for the first time a gender difference in mutation burden in cutaneous melanoma.

Figure 1.

Tumor Mutation Burden Associated with Gender and Survival. A) To the left of the dotted line, boxplots depicting missense mutation distribution for 19 cancers on a log-scale are plotted in order of increasing median missense mutation count. To the right of the dotted line, boxplots depict missense mutation counts by gender from Hodis et al. (2012). TCGA and Hodis et al. (2012) samples are depicted on a normal scale to right of the main figure. Diamonds designate mean missense mutation count. Boxplot whiskers correspond to first and third quartiles of data. B) While the positive relationship between mutation burden and survival is continuous, we determined that a definition of “high” vs “low” groups of missense mutation burden based on the threshold of 130 optimally captured survival differences for visualization purposes (univariate hazard ratio [HR] = 0.46, 95% confidence interval [CI] = 0.32 to 0.67; multivariable HR = 0.43, 95% CI = 0.28 to 0.64). These plots show survival curves for high mutation tumors vs low mutation tumors among women on left and high mutation tumors vs low mutation tumors among men on right based on the 130 mutation threshold. Mutation counts will vary based on sequencing platform and bioinformatic pipeline and thus the ‘130’ is not intended to represent an exact threshold of predictive discrimination for survival. A statistically significant survival advantage is observed for the women (univariate HR = 0.31, 95% CI = 0.17 to 0.58; multivariable HR = 0.32, 95% CI = 0.16 to 0.62) and men (univariate HR = 0.57, 95% CI = 0.36 to 0.92; multivariable HR = 0.49, 95% CI = 0.30 to 0.81). Among patients with “high mutation” tumors, the median survival was 167.9 months for women and 112.6 months for men. Events per patient at risk are indicated for 2000 day intervals below survival curves. Asterisks indicate a statistically significant difference (P < .05) in missense mutation burden by gender (Supplementary Table 4, available online). P values were determined from the negative binomial regression model with use of two-sided Wald tests.