Health Systems & Services Research | Breast Diseases: Benign & Malignant | Methods of Clinical Decision-Making | Statistical Methodologies & Health Informatics
Clinical avatars for breast cancer risk prediction: creation and preliminary Bayesian network simulations
Rimma Pivovarov*, Matthew Crawford, Peter Kos, Prasad Patil, Peter Tonellato
*Corresponding author: Rimma Pivovarov
Center for Biomedical Informatics, Harvard Medical School, Boston, MA, USA
F1000Posters 2010, 1: 379 (poster) [ENGLISH]
Poster [2.34 MB]
Presented at
Koch 2010 Summer Symposium: Integrative Approaches to Cancer,
10 - 11 Jun 2010, 100
Current estimates calculate that one in eight women will develop breast cancer in her lifetime. This statistic has persuaded medical and academic professionals to produce a wide variety of algorithms to assess an individual’s risk for developing breast cancer.
Major advances have been made since the first published algorithm in 1983, including the incorporation of proven risk-associated genes (BRCA1 and BRCA2) and recently, a handful of implicated SNPs. We have examined each algorithm in detail, extracting the methods, inputs, outputs, study and target populations to perform comparative analyses. Currently, the 31 existing risk prediction algorithms have not been fully tested in all racial subpopulations.
We have created a framework for simulation of unlimited sets of patient populations to explore racial discrepancies in risk prediction algorithms. Utilizing statistical characterizations of multiple breast cancer patient cohorts we have begun preliminary population simulations, using Bayesian networks for variable dependency modeling.
No relevant conflicts of interest declared.
Please note that most posters on this site present work that is preliminary in nature and has not been peer reviewed.
This poster is open access subject to the CC BY-NC Creative Commons 3.0 License
