Towards personalized risk prediction of cardiovascular disease

Aim: The aim is to develop improved screening strategies to cost-effectively identify at-risk individuals before they suffer ischemic heart disease (IHD).Work-plan: We use comprehensive data from several large contemporary cohorts with extensive cardiac phenotyping to develop improved and personalized risk algorithms.

Machine learning will be used to optimize prediction models from the large datasets.Data for risk stratification will be selected from a wide range of personal data and biological sampling and include proteomics and genomics (polygenic risk scores).

Data is combined with imaging using computed tomography for risk stratification based on extent of coronary artery disease (CAD, coronary calcifications/atherosclerosis).

Simultaneous imaging of metabolic risk factors such as ectopic fat in liver and around the heart aid in prediction.Screening for CAD/IHD risk is done in two steps, the first will be aimed at the whole population creating a smaller population with high pre-test likelihood of CAD/IHD.

The second step will refine the risk further using imaging and blood sampling.Clinical importance: In Sweden there are 7,500 cases of first ever myocardial infarction in ages below 70-years (900 of which are fatal, Socialstyrelsen, 2022).

By developing cost-effective risk-prediction algorithms that identify at-risk individuals before they suffer a myocardial infarction we hope to reduce the incidence of IHD by focusing intense risk factor intervention towards this group.