Cookies on this website

We use cookies to ensure that we give you the best experience on our website. If you click 'Accept all cookies' we'll assume that you are happy to receive all cookies and you won't see this message again. If you click 'Reject all non-essential cookies' only necessary cookies providing core functionality such as security, network management, and accessibility will be enabled. Click 'Find out more' for information on how to change your cookie settings.

Polygenic risk scores only modestly improve the predictive power of the QCancer-10 non-genetic risk score for colorectal cancer

When colorectal cancer is diagnosed earlier, the 5-year survival increases from 10% (when diagnosed at stage 4) to over 90% (if diagnosed at stage 1). Screening plays an important role in detecting these cancers earlier and improving patient outcomes.

Many countries, including the UK, have a staged approach to screening. Colonoscopy is the gold-standard technique for colorectal cancer screening and diagnosis but is costly and invasive and so is reserved for those at highest risk. Colorectal cancer risk can be assessed using simpler, less invasive tests such as the faecal immunochemical test (FIT), currently offered to all UK adults aged 60-74 (50-74 in Scotland).

Other approaches for stratifying risk are also being developed. For example, the QCancer-10 risk model combines a variety of non-genetic risk factors routinely recorded in electronic health records such as age, ethnicity, family history, lifestyle factors and some medical conditions, to predict the 15-year risk of colorectal cancer. Separately, genetic risk can be assessed through polygenic risk scores (PRS), which combine 100-1000s of single genetic changes that associate with colorectal cancer risk.

In this paper published in the British Medical Journal, Dr Sarah Briggs (Nuffield Department of Medicine) and colleagues in Oxford, London and Edinburgh evaluated the benefit of combining PRS with the QCancer-10 (colorectal cancer) non-genetic risk prediction model to predict colorectal cancer risk.

The team developed six PRS using data from the UK Biobank. The top-performing PRS (derived using a programme called LDpred2) was combined with QCancer-10 and the performance was compared to QCancer-10 alone. The integrated QCancer-10 and PRS models out-performed QCancer-10 alone, however, this improvement was only modest. Given that QCancer-10 can be calculated relatively easily using data available in health records, the team concluded that there is currently no clear justification for using PRS in risk-stratified population screening for colorectal cancer.