Abstract B16: Development of breast cancer risk prediction for the UK population using the UK Biobank dataset

Abstract

Breast cancer is the most common female cancer and is the second most common cause of cancer death among females. The UK has the highest age standardised incidence and mortality rates in the world, with two in every 1000 women aged 50 and above developing breast cancer annually. Several risk prediction models have been developed to estimate the likelihood for developing breast cancer based on specific risk factors in currently healthy individuals within a specific period of time. The available models are derived principally from either genetic or non-genetic factors. The majority of these models are however not user-friendly, do not focus on modifiable factors entirely and are not specifically designed for the general public. Our research group is developing an individualised risk prediction model for breast cancer focusing on the modifiable risk factors using the UK Biobank data. A nested case-control study within the 273,467 female participants is being used to develop the model. We have split the data into training and testing sets and will carry out all statistical tests to ensure our model calibrates well. For model validation, we will further seek external validation cohorts. The model will provide risk scores derived from the presence or absence of specific risk factors and will be compared to the general public score. The model will allow people to modify their risk profile with appropriate prevention measures. The main goal of the model is to be used in cancer education and prevention. The results from exploratory analyses suggested positive associations between breast cancer risk and age, breast cancer family history, menopause age, age at first child, BMI, height, null-parity, smoking, alcohol intake, and family history of other cancer. An algorithmic model will be developed based on these factors. We will also evaluate public perceptions using focus group technique. We will be presenting the results of the model development from the training set and the results of the internal validation from the testing set. In conclusion, we are developing an individualised breast cancer risk prediction model for the UK population based on the modifiable risk factors. The model will enable us to educate and to design appropriate interventions tailored to the individual with the aim of assisting them to make appropriate changes to modify their cancer risk profile. Citation Format: Kawthar Alajmi, Artitaya Lophatananon, Kenneth Muir. Development of breast cancer risk prediction for the UK population using the UK Biobank dataset. [abstract]. In: Proceedings of the AACR Special Conference: Improving Cancer Risk Prediction for Prevention and Early Detection; Nov 16-19, 2016; Orlando, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2017;26(5 Suppl):Abstract nr B16.