|
|
 |
| |
|
|
|
Estimation of Number of Involved Lymph Nodes in Breast Cancer Patients using Bayesian Regression Approach |
|
|
|
PP: 17-25 |
|
|
|
doi:10.18576/jsapl/040103
|
|
|
|
Author(s) |
|
|
|
Prafulla Kumar Swain,
Gurprit Grover,
Sangeeta Chakravorty,
Komal Goel,
Vikas Singh,
|
|
|
|
Abstract |
|
|
| Prediction of number of involved lymph nodes in breast cancer patients is an important criterion to assess the severity and progression of disease. The number of involved nodes is count data which often displays over-dispersion, hence the Poisson and Negative Binomial distribution is ultimate choice for modeling. In this paper we have made an attempt to estimate the number of involved lymph nodes in breast cancer patients using Bayesian regression approach assuming multivariate normal prior for the parameters. The posterior estimates have been derived using MCMC pack and the best model has been selected based on Deviance Information Criterion (DIC) values. The Bayesian Negative Binomial regression over performed than the Poisson regression. The predictors’ viz., tumor size, tumor grade, CA 15-3 marker and progesterone receptor status are significantly associated with the involved lymph nodes of the breast cancer patients. |
|
|
|
|
|
|
|
