This is the program for the 2010 Joint Statistical Meetings in Vancouver, British Columbia.

Abstract Details

Activity Number: 32
Type: Contributed
Date/Time: Sunday, August 1, 2010 : 2:00 PM to 3:50 PM
Sponsor: Section on Bayesian Statistical Science
Abstract - #309414
Title: A Single Test Derived from Modeling Biological Heterogeneity of Multiple Serum Biomarkers via Mixture Models Using Bayesian Inference
Author(s): Eric A. Macklin*+ and Ying Zhou and Trenton C. Pulsipher and Steven James Skates
Companies: MGH Biostatistics Center and MGH Biostatistics Center and MGH Biostatistics Center and Massachusetts General Hospital
Address: 50 Staniford St, Suite 560, Boston, MA, 02114,
Keywords: Ovarian cancer ; classification ; mixture models ; screening tests ; diagnostics ; ROC
Abstract:

CA125 is a serum biomarker in early detection trials with high sensitivity and specificity for late stage ovarian cancer (OC) but low sensitivity for early stage OC when treatments are curative. Adding biomarkers should increase sensitivity, but OC is a mixture of tumors that do or do not shed particular biomarkers. We propose a multivariate mixture model for k biomarkers with 2^k classes of biomarker expression (shed/not shed) reflecting the biological heterogeneity of OC. Within a class, multivariate t-distributions ensure robustness of the Bayes factor. The posterior probability of OC given a panel of biomarkers is estimated via Gibbs sampling to yield a single diagnostic test. We analyzed CA125 and 15 other biomarkers from 536 patients to identify tests which best complement CA125 while retaining high specificity. Performance of our approach is compared to discriminant analysis.


The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2010 program




2010 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.