Discovery of modifiable factors associated with Pan-Cancer Risk: An Exposome-Wide analysis.

INTRODUCTION: Hypothesis-driven studies have identified many modifiable cancer risk factors, but research focusing on single exposures overlooks their complex interactions. OBJECTIVES: This study applied an exposome-wide approach across multiple cancer types to systematically identify modifiable exposures and evaluate their combined effects with genetic susceptibility. METHODS: We analyzed data from over 460,000 UK Biobank participants with 15 years of follow-up, assessing 93 modifiable exposures in relation to 23 site-specific cancers using Cox models. Exposomic risk scores (ERS) were constructed to quantify the combined effects of identified factors. Population attributable fractions (PAFs) were calculated to estimate the potential population-level burden associated with these factors. Finally, polygenic risk scores (PRS) were incorporated to evaluate the relative contributions of genetic susceptibility and modifiable exposures to cancer risk. RESULTS: We identified 209 significant exposure-cancer risk estimates, with both their number and magnitude varying markedly across cancer types. Lung cancer exhibited the largest number of associations, whereas ovarian, testicular, and brain cancers showed no significant associations. Several exposures were associated with multiple cancers, such as basal metabolic rate, smoking, diabetes, household income, alcohol consumption, and body fat percentage, suggesting pleiotropic effects. ERSs summarized the combined influence of these modifiable exposures, and PAFs estimated their potential contribution to the population-level cancer burden across cancer types (4.5%-75.6%). PRS integration showed larger relative genetic contributions for prostate cancer (48.6%), melanoma (48.9%), and Hodgkin lymphoma (61%), while modifiable exposures were more influential in endometrial cancer (74.2%), lung cancer (71.1%), and liver cancer (57.8%). CONCLUSION: This study provides a pan-cancer, exposome-wide perspective on modifiable and genetic contributions to cancer risk, highlighting the potential value of targeting controllable exposures to reduce population-level cancer burden.