Research Grants Awarded
Improvement And Validation Of Brcapro
Investigator Initiated Research
Prediction models that identify individuals at high risk of breast cancer because of inherited genetic susceptibility play a critical role in genetic counseling and cancer prevention, as demonstrated by the widespread use and the proliferation of alternative models. Mendelian models have consistently shown the best performance in independent validation studies. In particular BRCAPRO (Berry et al JNCI 1997, Parmigiani et al AJHG 1998) remains the most accurate publicly available approach for identifying individuals likely to carry deleterious mutations of BRCA1 and BRCA2, across a range of clinical populations (Parmigiani et al AIM 2007). Though the counseling package CancerGene, BRCAPRO is licensed to about two thousand users and successfully contributes to improved screening, prevention, and genetic testing, and to the design and analysis of cancer studies. For example, BRCAPRO can be used to efficiently identify women eligible for MRI screening (Kauff and Offit JAMA 2007).
BRCAPRO exploits domain knowledge of Mendelian inheritance and other biological characteristics of susceptibility genes. It is built in a modular way that allows to update input parameters such as population allele frequencies and penetrance curves on an ongoing basis as new evidence accrues (Chen etal JCO 2006) and to expand the existing functionality to include additional predictive information. Recent extensions of BRCAPRO include incorporating information about oophorectomy in the proband and relatives (Katki et al BMC-MG 2007); allowing for consideration of pedigrees of arbitrary size (forthcoming upgrade) and information about ER, PR, Her2, CK5/6 and CK14 status of tumor in every relative (Tai, submitted). Opportunities for significant refinements continue to arise from both new knowledge and new technologies.
This constant process of updating and extension guarantees that BRCAPRO continues to serve as a translational bridge between new discoveries in genetics and clinical management of high risk families. The main goal of this application is to support and streamline this process via an infrastructure, several important improvements of BRCAPRO, and the validation of the enhanced model that will result from these improvements. Our aims are:
AIM 1. To implement the following improvements of the BRCAPRO model and software:
1.a Inclusion of pancreas cancer and consideration of undiscovered pancreas cancer genes.
1.b Inclusion of prostate cancer.
1.c Inclusion of Ductal Carcinoma in Situ.
1.d Inclusion of the Ovarian Cancer Cluster Region or OCCR.
1.e Inclusion of parity as an alternative source of familial effects in breast cancer.
1.f Inclusion of familial clustering due to unmeasured shared environmental exposures.
1.g Extension to allow for mixed ethnicity, that is for different branches on a family to belong to populations with different prevalence of mutations.
1.h Consideration of possible errors in self-reporting of family history data.
Each extension will be evaluated using the steps in Aim 2 before being released. Extension will be validated one at the time and in combination.
BRCAPRO will be designed to operate with or without each of the additional sources of information, to maximize flexibility in clinical applications.
AIM 2. To validate the newly released versions of BRCAPRO on external data from the CGN validation study (Parmigiani et al AIM 2007) and the National Familial Pancreatic Cancer Registry.
The CGN validation study includes a subset of over 1500 families for which we have available both the complete pedigree and genetic test results.
Under the auspices of the CGN will will also be able to collect followup information on a subset of these families, which will allow to validate predictors that were not originally collected.
Existing software, documentation, and related publications can be found at
All new development will continue to be free to the academic community and open source.
When a woman has defects (mutations) in the BRCA1 or the BRCA2 genes, she is very likely to have breast or ovarian cancer during her lifetime. Though these mutations are relatively rare in the general population, they are more common among women with a family history of these cancers. Testing for the presence of these mutation is commercially available, though it is expensive and may not always identify existing mutations. Many women who are worried that they have mutations of these genes because of their family history, seek advice about whether to be tested.
To help women with this and other related decisions, counselors estimate the probability that a woman has a BRCA1/BRCA2 mutation. Because this is a difficult estimation, most counselors use computer-based prediction tools, or "models". One of the most effective and widely used of these models is called BRCAPRO, and was developed by the team that is also proposing this study. BRCAPRO uses knowledge about the relationship between a woman's history, including which family members had breast and ovarian cancer, and of which specific kind, and defects in the BRCA1/BRCA2 genes. Recently, studies carried out on a large number of families have shown that prediction models, particularly BRCAPRO, are accurate and reliable counseling tools, but have also identified areas where significant improvements can be made.
The goals of this study are:
1) to carry out a set of nine major improvements of the BRCAPRO model;
2) to verify, using data that we mostly already collected from counseling centers across the nation, that these improvements do indeed do a better job, and 3) to make the upgraded software available for public use free of charge.
BRCAPRO is now used by up to two thousand clinics and research centers. This makes BRCAPRO a unique venue to translate new knowledge about inherited susceptibility to breast cancer into information that can be directly used by physicians, patients and insurers to make better decisions about how to prevent and treat breast and ovarian cancer. Specifically, the results of this study will be of benefit for
A) Women considering genetic testing for deleterious mutations of the BRCA1 and BRCA2 genes.
B) Women who are at high risk for breast and ovarian cancer and are considering preventative surgical procedures such as bilateral mastectomy (the removal of both breasts) or oophorectomy (the removal of the ovaries).
C) Women considering MRI screening for early detection of breast cancer.
Also, insurers will be able to more reliably make decisions about coverage of genetic tests and MRI screening.