BRCA1 and BRCA2 mutations increase breast and ovarian cancer risks substantially enough to warrant risk reduction surgery, despite variable risk estimates. Underlying this variability are methodological issues, and also complex genetic and nongenetic effects. Although many modifying factors are unidentified, known factors can already be incorporated in individualised risk prediction. BRCA1 and BRCA2 were identified as genes mutated in hereditary breast/ovarian cancer by genetic analysis in families with multiple cases of these malignancies. In the following decade, as BRCA1 and BRCA2 testing became more common in patients with a personal or family history of cancer, numerous studies assessed cancer risks in carriers of these mutations. These studies have resulted in variable risk estimates, leading to difficulties in defining risk figures that are clinically applicable to the individual patient. However, they illuminated the complexities that underlie risk prediction even in the context of a single gene with major effects on cancer predisposition. The emerging picture is that breast and ovarian cancer risks in BRCA1 and BRCA2 carriers are substantially higher than in the general population, but that they are considerably affected by nongenetic, environmental factors, and by additional genetic modifiers.
BREAST AND OVARIAN CANCER RISK IN WOMENOriginal estimates of breast and ovarian cancer risks in carriers were based on the families used for positional cloning of the BRCA1 and BRCA2 genes, where all carriers had molecular testing, and all cancer diagnoses were validated. These families were chosen because they harboured multiple, early-onset cases, that is, they were essentially selected for high cancer risk. In the Breast Cancer Linkage Consortium (BCLC) families, by age 70 years, BRCA1 carriers had a breast cancer risk of 71% (95% CI 53 -82%) and an ovarian cancer risk of 47 -63% (Easton et al, 1995). BRCA2 carriers had a breast cancer risk of 84% (95% CI 43 -95%), and an ovarian cancer risk of 27% (95% CI, 0 -47%) (Ford et al, 1998). Many subsequent studies were case-based, that is, risk was estimated by analysis of family history in relatives of breast or ovarian cancer patients, often from hospital-based series (Antoniou et al, 2003). The design of these studies partially addressed the ascertainment bias inherent in studying highly selected, multiple-case families, but risk estimates were based on family history, rather than on testing carriers directly and validating their cancer status. With this design, misclassification can result from discrepancies between modelled vs true carrier status, and from errors in cancer reporting by relatives, a particularly common problem for ovarian cancer. In a meta-analysis of such case-based studies, by age 70 years, in BRCA1 carriers breast cancer risk was 65% (95% CI 51 -75%) and ovarian cancer risk was 39% (95% CI 22 -51%), and in BRCA2 carriers breast cancer risk was 45% (95% CI 33 -54%) and ovarian cancer risk was 11% (95% CI 4.1 -18%) (Antoniou et al, 2003). One s...