Interaction Between ECG and Genetic Markers of Coronary Artery Disease

Abstract

Coronary artery disease (CAD) is the main contributor to cardiovascular mortality in developed countries, making accurate diagnosis of utmost importance. We developed risk scores to assess CAD risk in a population without known cardiovascular disease by combining ECG and a genetic risk score (GRS) for CAD. We analysed data in 52,260 individuals in the UK Biobank study. ECG indices included heart rate, PR, QRS, QT and T-peak-toT end intervals, while we built the GRS from publicly available genome-wide association results for CAD that were derived in an independent population. In a training set (N = 39,195), the indices with the strongest CAD prognostic impact were the PR and QT intervals, and the GRS. When combined together into a Multivariate model, both the ECG markers and the GRS were independently associated with CAD. In an independent test set (N = 13,065), we then built three risk scores based on (1) ECG markers, (2) genetic data, and (3) a combination of ECG and genetic data, respectively. The hazard ratio (95% confidence interval) for CAD comparing high versus low-risk individuals was 6.

CAD-CDN: Coronary Artery Disease Prediction Using Convolutional Neural Network with Modified Densenet

“…In addition, ruptured plaques and calcification can constrict the patient's small arteries, leading to angina pectoris and myocardial infarction. We use a deep learning approach to investigate IVUS images for calcification in this study [9][10][11][12][13][14][15][16][17].…”