Predicting Time-to-Event Variables by Using High-Dimensional Imaging and Genetic Data

Abstract

With the rapid growth of modern technology, many large-scale biomedical studies have collected massive datasets with large volumes of complex information (e.g., imaging, genetics, or clinical) from increasingly large cohorts, while high-dimensional missing data are frequently encountered in various stages of the data collection process. Simultaneously extracting and integrating rich and diverse information from such big data with the presence of high-dimensional missing data is critical for making major advances important for diagnosis, prevention, and treatment of numerous complex disorders (e.g., Alzheimer or breast cancer). This talk is to introduce survival models with functional covariates (SMFC) to integrate imaging, genetic, and clinical data to predict the time-to-event outcome by: (a) using multiscale functional component analysis and kernel machine methods to achieve dimension reduction for imaging and genetic data; (b) explicitly accounting for spatial correlation in imaging and genetic data. The new methods are primarily motivated from a longitudinal database with large volumes of imaging, genetic, behavioral, and clinical data collected by the Alzheimer's Disease Neuroimaging Initiative (ADNI) study.?