A0688
Title: Time-dependent ROC curve for multiple longitudinal biomarkers and its application in diagnosing cardiovascular events
Authors: Lizhe Sun - Shanxi University of Finance and Economics (China) [presenting]
Abstract: Since it can help people detect the early signs of diseases, accurate diagnostic techniques based on biomarkers are crucial in biomedical research. A novel bivariate time-varying coefficients Logistic regression model for addressing the combination of longitudinal biomarkers is proposed. By using the B-spline method to estimate the proposed model, multiple longitudinal biomarkers can be effectively combined, and the accuracy of disease onset prediction is improved. It is shown that the proposed method is theoretically consistent. It exhibits superior performance compared to the existing method, as presented through numerical results. The proposed method is verified in a study on predicting the probability of onset of future cardiovascular events for type 2 diabetic patients by combining longitudinal biomarkers: HbA1c and LDL-C. The combination of longitudinal biomarkers has been demonstrated to significantly improve disease diagnosis compared to a combination of updated biomarkers.
