I. ROC Analysis for Biomarker Discovery and Validation

Biomarkers play an important role in early detection of disease and clinical decision-making process. In particular, recent advances in genomics, molecular biology and imaging technologies allows us to identify tons of biomarkers simulataneously. Dr. Chungs research has been focused on using the receiver operating characteristic (ROC) curve for evaluating such biomarkers in various studies including:

• cross-sectional study with high-dimensional biomarkers;
• disease surveillance study with longitudinal biomarkers;
• two-phase biomarker discovery and validation study.

II. Shape-Restricted Hazard Analysis for Surviva Data Anslysis

Isotonic regression is a useful nonparametric technique for fitting a monotone increasing (or decreasing) function. It offers a flexible tool in estimating a monotone regression relationship between response and covariate. His research applies the isotonic regression techniques to Cox's proporiontal hazards model under a natural assumption that the hazard function is a monotone function of one of the covariates. His current research projects include estimations of unimodal (or U-shaped) hazard function where the hazard function is monotone increasing and decreasing over a mode.