B2009
Title: Long-term risk prediction utilizing mammography data
Authors: Shu Jiang - Washington university (United States) [presenting]
Abstract: Screening mammography aims to identify breast cancer early and secondarily measures breast density to classify women at higher or lower than average risk for future breast cancer in the general population. Our primary goal is to extract mammogram-based features that augment the well-established breast cancer risk factors to improve prediction accuracy. We will present a novel supervised functional principal component analysis to extract image-based features that are ordered by association with the failure times. A closed-form solution is provided through the proposed eigenvalue decomposition algorithm. Empirical comparisons are made to the conventional functional principal component analysis and the functional partial least squares method. The proposed method is applied to the motivating Joanne Knight Breast Health cohort at Siteman Cancer Center. Our study demonstrates superior prediction performance compared to the benchmark models and reveals insights into risk patterns within mammograms.
