Chicago Center for Jewish Genetics Disorders - Health Professionals

Risk Models for Breast and Ovarian Cancer

Risk models are mathematical formulas that use various factors and information to calculate a theoretical risk. In the case of breast cancer risk models, data such as family history of cancer, type of cancer, and age of diagnosis are entered for an individual, and a theoretical risk is then given for their likelihood to develop breast cancer within a specific time period or over their lifetime. Models also exist to predict the likelihood of a BRCA1 or BRCA2 mutation being present in an individual. There are several different breast and ovarian cancer risk models that have been developed. A review of three of the most common models, including what risk factors they use and their strengths and their weakness, are presented here.

Claus Model: The Claus model was developed from the information gathered from the Cancer and Steroid Hormone Study (Claus et al, Am J Hum Genet 1991; 48:232-42). This model makes that assumption that the inherited risk in an individual is due to an autosomal dominant gene mutation with high penetrance, such as is seen for BRCA1 and BRCA2. The risk estimate that is calculated is based upon a woman’s current age, the number of her 1st and 2nd degree relatives that have been diagnosed with breast cancer, and their respective ages of diagnosis. The risk is given as the likelihood of developing cancer over a woman’s lifetime or in 10 year intervals. This model is considered to be a good tool for estimating breast cancer risk based upon a family history of breast cancer. However, it does not take into account any of the other factors known to increase breast cancer risk such as atypical hyperplasia found on breast biopsies or hormonal history. It also only allows the calculation of risk for women who have at least one female relative with breast cancer. For extensive family histories of breast and ovarian cancer, the BRCAPRO model would be more appropriate.

Gail Model: The Gail model is based upon the Breast Cancer Detection Demonstration Project (Gail et al, J Nat Cancer Inst, 1989 Dec 20;81(24):1879-86) which identified major predictors of risk such as: current age, age at menarche, age of 1st live birth, number of previous breast biopsies, presence of atypical hyperplasia on breast biopsy, and the number of 1st degree relatives (mother or sisters) with breast cancer. It estimates the risk of developing invasive breast cancer over the next five years of the woman’s life or over the course of the her lifetime. It is considered the best current model for a woman without a strong family history of breast cancer and for those who adhere to regular mammography screening. However, it does not take into account 2nd degree relatives such as paternal relatives with breast cancer or the ages of onset of the cases of breast cancer. It also does not take into account any cases of ovarian cancer in a family. It overestimates the risk for women with mothers or sisters diagnosed with breast cancer after the age of 50 years and for premenopausal women who do not receive annual mammograms. It is considered an underestimate of women who have had 2nd degree relatives with breast cancer before the age of 50 or ovarian cancer at any age.

BRCAPRO: BRCAPRO is a model that calculates the probability that a particular family member carries a germ-line mutation in the BRCA1 or BRCA2 gene (Parmigiani et al, Am J Hum Genet 1998;62:145-158). The model has been validated and takes into account ethnicity (specifically whether an individual is of Ashkenazi Jewish descent or not), the individual’s personal cancer history, and their 1st and 2nd degree relatives’ medical history. An advantage of this model is the ability to add information about both affected and unaffected 1st and 2nd degree individuals in the calculation, and to use ages of cancer onset and ages of unaffected individuals. Many medical sites do BRCAPRO analysis before testing and have a calculated 10% cut-off risk of having either a BRCA1 or BRCA2 mutation before proceeding to genetic testing. Decisions to have testing are not driven exclusively by the risk number, but by medical decision making factors, and the impact of testing on the well-being of their relatives. Please click here for more information regarding updated guidelines for cancer genetic testing from the American Society of Clinical Oncology, 2003. (ASCO, 2003, J Clin Oncol 21(12): 2397-2406).

When providing risk assessment, it is important to give patients figures for the probability of carrying a mutation AND for the risk to develop breast cancer. In order to do so, one must incorporate the use of all models in order to balance the advantages and disadvantages of each model. Furthermore, it is crucial to choose the relative in the family history who will best capture all cases of cancer to determine risk figures and then to provide estimates of risk to your patient based on their relation to that relative. Because risk assessment is complex and technically difficult, expert opinion and clinical judgment are important to provide accurate information to patients, especially in families with pancreatic cancer and male breast cancer. Geneticists and genetic counselors are knowledgeable in these models and should be the medical professionals performing this risk estimation as part of a genetic counseling session. For more information and to find a genetic counselor in your area, please visit www.nsgc.org, www.cancer.gov, www.nci.nih.gov/cancerinfo/prevention-genetics-causes/breast.