Abstract Background: The majority of individuals diagnosed with melanoma lack important clinical risk factors, particularly an increased numbers of moles. Identifying high-risk individuals among those with low-risk mole phenotypes could potentially improve public health through targeted surveillance and early detection. We previously demonstrated significant associations between melanoma risk, melanocortin receptor (MC1R) polymorphisms, and indoor ultraviolet light (UV) exposure. Existing melanoma risk prediction models do not include these factors so we investigated their potential to improve the performance of a risk model, especially among individuals with low-risk mole phenotypes. Methods: We determined MC1R genotypes and utilized self-reported phenotypic and UV (indoor and outdoor) exposure variables from 875 melanoma cases and 765 healthy controls from the population-based Minnesota Skin Health Study. We divided the subjects into discovery and validation cohorts balanced on case-control status, age, gender and mole phenotype. Mole phenotype was assessed using diagrams illustrating 4 categories of mole density (“none," “few," “some” or “many” moles). We used the discovery cohort (442 cases, 389 controls) to develop 2 risk models. Model A included conventional risk factors (i.e. age; gender; hair, skin, and eye color; mole phenotype; and freckling). Model B added outdoor UV, indoor UV, and MC1R genotype variables to those in Model A to determine the contribution of those variables to the model's predictive ability. Finally, we assessed the predictive ability of both models in the validation cohort (433 cases, 376 controls), including a subgroup analysis of subjects with a low-risk mole phenotype (i.e. “none” or “few” moles). Results: In the validation cohort the area under the receiver operating characteristic curve (AUC) for Model A was 0.721 (95% CI 0.684 - 0.757). The AUC was substantially decreased to 0.693 (0.652 - 0.733) when the analysis was confined to subjects with low-risk mole phenotypes (315 cases, 340 controls). Incorporating outdoor UV, indoor UV, and MC1R genotype into the enhanced model (Model B) significantly increased the AUC for the whole cohort to 0.740 (0.706 - 0.774; p=0.047). An improvement, albeit non-significant, was also observed for the smaller, low-risk group, AUC = 0.715 (0.676-0.755; p=0.09). Conclusions: We developed and validated 2 population-based melanoma risk prediction models. The enhanced model is the first in melanoma to demonstrate that a significant increase in predictive ability can be obtained by adding genotypic and UV exposure data to conventional phenotypic variables. These data raise the possibility that incorporating additional genetic risk markers into the model may further improve melanoma risk prediction, and highlight the need to better identify those who lack obvious high-risk clinical signs. Citation Format: Lauren Smith, Meng Qian, Yongzhao Shao, Marianne Berwick, DeAnn Lazovich, David Polsky. Improving melanoma risk prediction among individuals with low-risk mole phenotypes. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 2288. doi:10.1158/1538-7445.AM2013-2288