Abstract:
|
Dynamic prediction is to use longitudinal biomarkers for real-time prediction of individual prognosis. This is critical for patients with non-curable disease such as cancer. Biomarker trajectories are usually not linear, or even not monotone, and vary greatly across individuals. Therefore, it is difficult to fit them by parametric models. With this consideration, we propose to use functional data analysis approaches to summarize the patterns of patients' longitudinal biomarker trajectory changings over time, and link these patterns to the outcome of disease progression or survival. Simulation studies show that the proposed approach achieves stable estimation of biomarker effects over time, has good predictive performance, and is robust against model misspecification. The proposed method is applied to patients with chronic myeloid leukemia. At any time following their treatment with tyrosine kinase inhibitors, longitudinally measured BCR-ABL gene expressions are used to predict risks of disease progression.
|