?Background: There is growing research on risk-based strategies to improve melanoma early detection. Understanding the contribution of genomic (polygenic) variants to melanoma risk over phenotypic (traditional) risk factors is important for assessing the clinical utility of incorporating polygenic information into prevention and risk-based early detection strategies.

Objectives: Our objective was to examine the degree of overlap between polygenic and traditional risk of melanoma in an Australian population sample.

Methods: As part of a randomised controlled trial previously reported, participants (n=1,025) from all States and Territories in Australia were recruited via Medicare (Australia’s publicly funded universal healthcare system). Eligible participants aged 18-69 years had European ancestry and no personal history of melanoma. The intervention group provided a saliva sample for genotyping and were given their personal melanoma polygenic risk information (n=509). The polygenic score was based on 40-variants (single nucleotide polymorphisms; SNPs) in 20 genes/gene regions with established melanoma risk associations. Traditional risk scores were based on a validated published risk prediction model that included variables for naevi, hair colour, sunbed use, first-degree family history of melanoma, and personal history of keratinocyte cancers (e.g. basal cell carcinoma, squamous cell carcinoma).

Results: When analysed as continuous scores, we found evidence of a moderate correlation between polygenic and traditional risk (Spearman correlation estimate: 0.18, 95% Confidence Interval (CI): 0.10-0.27; p<0.0001). When analysed as quintiles, 26% of participants were categorised in the same (concordant) quintile. Of participants in the lowest traditional risk quintile, 11% were in the highest polygenic risk quintile, indicating high genetic susceptibility despite a lack of traditional risk factors. Based on traditional risk alone, this discordant group would be considered at low melanoma risk and be a low priority for preventive interventions despite having elevated genomic susceptibility. We also found a similar proportion of participants who were in the highest traditional risk quintile but lowest polygenic risk quintile (11%). This risk discrepancy could be explained by behavioural factors such as sunbed use and high UV exposure causing previous keratinocyte cancers (contributing to high traditional risk) in a person with otherwise low genetic susceptibility, or due to a person with many naevi (a strong contributor to high traditional risk) but which is less well captured in the polygenic score since a naevus polygenic risk score has been shown to explain only 5% of variance in naevus count.

Conclusions/implications: Given the growing diversity in the Australian population, our findings support the need to consider comprehensive melanoma risk assessment including polygenic and traditional risk factors with tailored risk communication.