Comparing the value of mammographic features and genetic variants in breast cancer risk prediction.

Published online

Journal Article

The goal of this study was to compare the value of mammographic features and genetic variants for breast cancer risk prediction with Bayesian reasoning and information theory. We conducted a retrospective case-control study, collecting mammographic findings and high-frequency/low-penetrance genetic variants from an existing personalized medicine data repository. We trained and tested Bayesian networks for mammographic findings and genetic variants respectively. We found that mammographic findings had a higher discriminative ability than genetic variants for improving breast cancer risk prediction in terms of the area under the ROC curve. We compared the value of each mammographic feature and genetic variant for breast risk prediction in terms of mutual information, with and without consideration of interactions of those risk factors. We also identified the interactions between mammographic features and genetic variants in an attempt to prioritize mammographic features and genetic variants to efficiently predict the risk of breast cancer.

Full Text

Duke Authors

Cited Authors

  • Wu, Y; Liu, J; Page, D; Peissig, P; McCarty, C; Onitilo, AA; Burnside, ES

Published Date

  • 2014

Published In

Volume / Issue

  • 2014 /

Start / End Page

  • 1228 - 1237

PubMed ID

  • 25954434

Pubmed Central ID

  • 25954434

Electronic International Standard Serial Number (EISSN)

  • 1942-597X

Language

  • eng

Conference Location

  • United States