Bivariate Bayesian Neural Networks for Prostate Cancer Study

• Activity Number: 1
• Type: Invited
• Date/Time: Sunday, August 7, 2005 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Malay Ghosh*+ and Sounak Chakraborty and Tapabrata Maiti
• Companies: University of Florida and University of Florida and Iowa State University
• Keywords: Bivariate Binary ; Margin and SV Positivity ; Neural Networks ; Histopathological Data ; Gene Expression Data ; Bivariate Logistic

Abstract

This paper considers a hierarchical Bayesian neural network approach for posterior prediction probabilities of certain features indicative of nonorgan confined prostate cancer. In particular, we find such probabilities for margin positivity (MP) and seminal vesicle (SV) positivity jointly. The available training set consists of bivariate binary outcomes indicating the presence or absence of the two. We consider both histopathological data and gene expression microarray data as covariates and compare their performance. We take a hierarchical Bayesian neural network approach to find the posterior prediction probabilities. The Bayesian procedure is implemented by an application of the Markov chain Monte Carlo numerical integration technique. For the problem at hand, our Bayesian bivariate neural network procedure is shown to be superior to the classical neural network, Radford Neal's Bayesian neural network, and bivariate logistic models to predict jointly the MP and SV in a patient in both the datasets as well as in the simulation study.