Models for predicting the presence of  mutations in the BRCA1 and BRCA2 genes


Models for predicting the presence of  deleterious mutations in the BRCA1 and BRCA2 genes are generally  reliable, and can be used in clinical activities. The models that  incorporate knowledge of how the genes are transmitted tend to provide a better discrimination, although that comes at the cost of a  somewhat more extensive data collection. One important caveat is that none of the models should be used  mechanically by applying a threshold on the probability across all  individuals. The main reason is that carriage probability is only one  of the several considerations thas should enter the decision about  seeking genetic testing. For a general oncology practice, the BRCAPRO model shows a distinct  advantage in terms of discriminating people who will test positive  from people who will not in the population-based portion of our  study. However, other models such as FHAT and Myriad are easier to  implement and may be preferable for practical reasons, especially in  the context of referring patient from primary care clinics to  counseling clinics, where a more in-depth evaluation of the family  history can be performed. Looking ahead, some of the models that we studied in our Annals article are undergoing significant improvements. This applies  especially to BRCAPRO, which now is able to incorporate information  about oophorectomy in any family members as well as information on  molecular markers in  any affected member, and will soon be extended  to handle pedigrees of arbitrary size. The next step for us will be  to reevaluate the models in the light of these ongoing improvements.
 

 

Bibliographic reference:

Parmigiani G et al.: "Validity of models for predicting BRCA1 and BRCA2 mutations", Ann Intern Med. 2007 Oct 2;147(7):441-50

 

 

Giovanni Parmigiani

The Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University, Baltimore, MD, USA