Using risk advancement periods to derive starting ages of colorectal cancer screening according to sex and polygenic risk score: Results from the UK Biobank

ObjectivePolygenic risk scores (PRSs) derived from genome-wide association studies are strong predictors of colorectal cancer (CRC) risk. We applied the straightforward approach of risk advancement periods (RAPs) to derive risk-adapted starting ages of CRC screening according to sex and PRS in the UK Biobank. MethodsAmong 242,779 participants (40-69 years; no previous CRC screening; no family history of CRC), we assessed associations of sex and a PRS with CRC risk and mortality using Cox regression models. Hazard ratios (HRs) were translated to RAPs to quantify how many years of age earlier men and women in defined PRS deciles reach comparable risks as those in the reference group (5th and 6th PRS deciles). ResultsDuring a median follow-up of 11.2 and 12.8 years, 2,714 participants were diagnosed with CRC and 758 died from CRC, respectively. HRs (95% CIs) of CRC risk were 1.57 (1.46, 1.70) for men versus women and ranged from 0.51 (0.41, 0.62) to 2.29 (2.01, 2.62) across PRS deciles compared to the reference. RAPs (95% CI) were 5.6 (4.6, 6.6) years for men versus women, and ranged from -8.4 (-11.0, -5.9) to 10.3 (8.5, 12.1) years across PRS deciles compared to the reference. Risk-adapted starting ages would vary by 24 years between men in the highest PRS decile and women in the lowest PRS decile. Very similar results were obtained regarding CRC mortality. ConclusionConsideration of sex and a standard PRS alone could have far-reaching implications for starting ages of CRC screening in the "average risk population". SUMMARY BOXO_ST_ABSWhat is already known on this topicC_ST_ABSO_LIPolygenic risk scores (PRSs) are strong predictors of colorectal cancer (CRC) risk. C_LIO_LIMen have substantially higher CRC incidence and mortality than women. C_LIO_LICRC risk information from the combination of PRS and genetically determined sex, which are constant factors over lifetime, is so far not used for risk-adapted CRC screening in the "average risk population". C_LI What this study addsO_LIRisk advancement periods (RAPs) of CRC by PRS and sex were derived at high levels of precision from the large database of the UK Biobank. C_LIO_LIBy joint consideration of PRS and genetically defined sex, risk-adapted starting ages would vary by as much as 24 years between men in the highest PRS decile and women in the lowest PRS decile. C_LI How this study might affect research, practice or policyO_LIOur study demonstrates a straightforward way to translate CRC risk information from a large population-based cohort into risk-adapted starting ages of screening. C_LIO_LIPersonalized, risk-adapted starting ages of CRC screening could be derived from a single blood test performed in middle adulthood. C_LIO_LIThe RAP approach could be easily extended for defining personalized starting ages by incorporating additional risk factors in the regression models. C_LI