Validation of existing risk prediction models for prostate cancer using data from UK BioBank and EPIC-Norfolk

In the UK, prostate cancer is the most common cancer in men with over 40,000 new cases diagnosed each year and approximately 10,000 dying as a direct result of the disease. It is a highly variable disease with some men at risk of serious illness that will harm them, whilst in others the disease is relatively harmless. At present, we don’t know how to get the balance right between 1) the early identification of those with serious disease to prevent the disease from causing harm and 2) minimising the detection of disease which is harmless, for which some may receive unnecessary treatment. Patient and clinician groups have identified this issue as research priority for the National Health Service. Systematic review evidence has identified many prostate cancer prediction models. However, few of the models that use variables in addition to PSA have been external validated, only one of which was validated in a UK population. We plan to test the these models using biological and medical data collected from large numbers of volunteers in the UK BioBank and / or the European Prospective Investigation of Cancer (EPIC) study. Findings from the review will be the first step towards developing an improved risk prediction model for prostate cancer designed for use in the UK male population.