Logistic Regression with Markov Chains as Covariates with Application in Identifying Lung Cancer Risk Factors

• Activity Number: 257
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Chung-Han Ho*+ and Carol J. Etzel and Wenyaw Chan and Randa El-Zein and Alanna C. Morrison
• Companies: The University of Texas School of Public Health and MD Anderson Cancer Center and The University of Texas School of Public Health and MD Anderson Cancer Center and The University of Texas School of Public Health
• Keywords: Logistic regression ; Markov chain ; EM algorithm ; CBMN ; Lung cancer

Abstract

The tobacco-specific nitrosamine, NNK,is an obvious carcinogen for lung cancer. CBMN (Cytokinesis-blocked micronucleus) has been found to be extremely sensitive to NNK-induced genetic damage, and its potential as an important lung cancer risk factor could improve risk prediction. However, the association between lung cancer and two CBMN stages under NNK-induced genetic damage has not been rigorously examined. This research proposes to model the chromosomal changes under NNK-induced genetic damage using a Markov chain in a logistic regression framework in order to predict the occurrence of lung cancer. Since the transition probabilities of these Markov chains are not directly observable, the EM algorithm will applied to carry on the estimation process. The ability of this approach to increase discriminating power to predict lung cancer will be compared to a baseline "non-genetic" model.