Longitudinal Biomarker Prediction for Precision Medicine


Shouhao Zhou, MD Anderson Cancer Center


2018.07.11 14:00-15:00


520 Pao Yue-Kong Library


This work was motivated by a study of chronic myeloid leukemia (CML), in which the limited-range continuous response variable, BCR-ABL transcript level, is a surrogate of residual disease following treatment with tyrosine kinase inhibitors. We propose a class of Bayesian hierarchical models to estimate and predict the distributions of longitudinal trajectory of BCR-ABL transcript levels for individual patients, to help physicians understand the relative extremeness of a recently measured BCR-ABL level at a certain time point, and/or to predict the future trajectory of BCR-ABL transcript levels for a specific CML patient given historical data. This information will enhance the personalized management of patients’ treatment for precision medicine.