WARNING: the interactive features of this website use CSS3, which your browser does not support. To use the full features of this website, please update your browser.
The aim of this study was to compare and externally test the performance of risk scores developed to predict new cases of colorectal cancer that include common genetic variants (SNPs), with or without established lifestyle/environmental (questionnaire-based/classical/phenotypic) risk factors. Adding phenotypic risk factors without age improved the ability of models to predict colorectal cancer risk in men but not in women. Adding phenotypic risk factors and age improved risk score performance in all cases, with the best performing models including SNPs, phenotypic risk factors, and age. Among middle-aged people in the UK, existing risk scores including SNPs discriminate moderately well between those who do and do not develop colorectal cancer over 6 years. Consideration should be given to exploring the feasibility of incorporating genetic and lifestyle/environmental information in any future stratified colorectal cancer screening program.
Validating risk prediction models for common hormonal cancers
Genetic and other risk-factor data can potentially be used to predict a person?s risk of developing a particular type of cancer. These individual risks could be used to focus screening programs or risk-lowering interventions towards those at highest risk. We are developing risk prediction models for breast, ovarian, endometrial and prostate cancer which include information about common genetic variants and other risk factors. We would like to use Biobank data to evaluate the performance of our models.
Separately, we would like to estimate the effects of hormone levels and genetic variants on cancer risk and survival. This research is both health-related and in the public interest. The very large, prospective UK Biobank cohort is an ideal resource for validating our cancer risk prediction models ? it is essential that these models are thoroughly evaluated before they can be considered for clinical use. We will use our risk-prediction models to predict each participant?s risk of developing breast, ovarian, endometrial or prostate cancer within five or ten years of enrolment (excluding those with any prior cancer diagnosis), based on their genotypes, family history of cancer and other variables. We will divide participants into groups depending on their risk level, then compare the predicted numbers of cancers in each group with the number observed in that time period.
Separately, we will compare circulating hormone levels at enrolment between those who did or did not later develop cancer and perform GWASs using baseline cases. We are requesting access to data for the whole UKBB cohort.