Validating Bayesian Prediction Models: a Case Study in Genetic Susceptibility to Breast Cancer
A family history of breast cancer has long been recognized to be associated with predisposition to the disease, but only recently have susceptibility genes, BRCA1 and BRCA2, been identified. Though rare, mutation of a gene at either locus is associated with a much increased risk of developing breast as well as ovarian cancer. Understanding this risk is an important element of medical counseling in clinics that serve women who present with a family history. In this paper we discuss validation of a probability model for risk of mutation at BRCA1 or BRCA2. Genetic status is unknown, but of interest, for a sample of individuals. Family histories of breast and ovarian cancer in 1st and 2nd degree relatives are available and enable calculation, via the model of Berry et al. (1997) and Parmigiani et al. (1998b), of a carrier probability score. Results of genetic tests with unknown error rates are available with which to validate carrier probability scores. A model is developed which allows joint assessment of test sensitivity and specificity and carrier score error, treating genetic status as a latent variable. Estimating risk and using receiver operating characteristic (ROC) curves for communicating results to practitioners are discussed.
KeywordsOvarian Cancer Receiver Operating Characteristic Curve Mutation Carrier Genetic Test Result Probability Score
Unable to display preview. Download preview PDF.
- Berry, D.A. and Parmigiani, G. (1997). Assessing the benefits of testing for breast cancer susceptibility genes: A decision analysis. Breast Disease. Google Scholar
- Berry, D.A., Parmigiani, G., Sanchez, J., Schildkraut, J.M., and Winer, E.P. (1997). Probability of carrying a mutation of breast-ovarian cancer gene BRCA1 based on family history. J Natl Cancer Inst, 89:9–20.Google Scholar
- Best, N.G., Cowles, M.K., and Vines, K. (1995). CODA: Convergence diagnostics and output analysis software for Gibbs sampling output, version 0.30 Technical report, MRC Biostatistics Unit, University of Cambridge.Google Scholar
- Claus, E.B., Risch, N., and Thompson, W.D. (1991). Genetic analysis of breast cancer in the cancer and steroid hormone study. American Journal of Human Genetics, 48:232–242.Google Scholar
- Garber, J.E. (1997). Breast cancer markers: genetic markers of breast cancer predispositoin. In Proceedings of the Thirty-third nnual Meeting of ASCO, pp 213–216. ASCO.Google Scholar
- Joseph, L., Gyorkos, T.W., and Coupal, L. (1995). Bayesian estimation of disease prevalence and the parameters of diagnostic tests in the absence of a gold standard. American Journal of Epidemiology, 141:263–272.Google Scholar
- Oddoux, C., Struewing, J.P., Clayton, C.M., Neuhausen, S., Brody, L.C., Kaback, M., Haas, B., Norton, L., Borgen, P., Jhanwar, S., Goldgar, D.E., Ostrer, H., and Offit, K. (1996). The carrier frequency of the BRCA2 6174delT mutation among Ashkenazi Jewish individuals is approximately 1%. Nature Genetics, 14:188–190.CrossRefGoogle Scholar
- Parmigiani, G., Berry, D., Iversen, E.S., Jr., Müller, P., Schildkraut, J., and Winer, E.P. (1998a). Modeling risk of breast cancer and decisions about genetic testing. In Case Studies in Bayesian Statistics, Volume IV, pages 133–203, Springer-Verlag.Google Scholar
- Raftery, A.E. and Lewis, S.M. (1996). Implementing MCMC. In Gilks, W.R., Richardson, S., and Spiegelhalter, D.J., editors, Markov Chain Monte Carlo in Practice, pages 115–127, London. Chapman and Hall.Google Scholar