The genetic attributable risk of breast and ovarian cancer

231 8

The Genetic Attributable Risk of Breast and Ovarian Cancer

Elizabeth B. Claus, Ph.D., M.D.”’

Joellen M. Schildkraut, Ph.D.3 w. Douglas Thompson, Ph.D.4 Neil J. Risch, Ph.D.’

’ Departments of Epidemiology and Public Health, Yale University School of Medicine, New Haven, Connecticut.

Department of Surgery, Yale University School of Medicine, New Haven, Connecticut.

Department of Family and Community Medi- cine, Duke University School of Medicine, Dur- ham, North Carolina.

Department of Applied Medical Sciences, Uni- versity of Southern Maine, Portland, Maine.

Department of Genetics, Stanford University, Stanford, California.

The Cancer and Steroid Hormone Study was supported by interagency agreement 3-YO1-8- 1037 between the Centers for Disease Control and the National Institute of Child Health and Human Development, with additional support from the National Cancer Institute.

This work was supported by a grant from the Patrick and Catherine Weldon Donaghue Medi- cal Research Foundation as well as by NIH grant l-RO3-CA59419 from the National Cancer Insti- tute and NIH grant HG00348 from the National Center for Human Genome Research.

W.D. Thompson and data collection in Connect- icut were supported by Contract # 200-80-0561 from the Centers for Disease Control.

The authors wish to acknowledge the contribu- tors to the Cancer and Steroid Hormone Study. Study Design and Coordination at the Division of Reproductive Health, Center for Chronic Dis- ease Prevention and Health Promotion, Centers for Disease Control: Principal Investigator: George L. Rubin; Project Director: Phyllis A.

BACKGROUND. The age-specific proportion of breast and ovarian cancer in the general population that is likely to be due to a breadovarian cancer susceptibility gene(s) is estimated. In addition, the age-specific penetrance of ovarian cancer for women predicted to be carriers of a susceptibility gene is calculated using population-based data. METHODS. Data are from the Cancer and Steroid Hormone Study, a population- based, case-control study conducted by the Centers for Disease Control, which includes 4730 breast cancer cases aged 20 to 54 years. Information regarding the occurrence of breast and ovarian cancer was collected for mothers and sisters of the cases during an in-home interview. The probability of being a breast cancer susceptibility gene carrier was calculated for each of the breast cancer cases using information on the family history of breast cancer. The calculated risk of ovarian cancer in the first-degree relatives of breast cancer cases with a high probability of being a gene carrier is compared with that seen in first-degree relatives of breast cancer cases with a low probability of being a gene carrier and used to calculate the proportion of ovarian cancer cases that are likely to be due to a breastlovarian susceptibility gene(s) as well as the age-specific risk of developing ovarian cancer for gene carriers. RESULTS. Approximately 10% of ovarian cancer cases and 7% of breast cancer cases in the general population are estimated to be carriers of a breastlovarian cancer susceptibility gene; these women are found primarily in families characterized by multiple cases of the early onset of breast cancer. The proportion of breast cancer cases predicted to be attributable to the gene decreases markedly with age; approximately 33% of cases age 20-29 years compared with approximately 2% of cases age 70-79 years. The proportion of ovarian cancer cases predicted to be due to the susceptibility gene ranges from 14% among patients diagnosed in their 30s to 7% among those diagnosed in their 50s. Carriers are predicted to have at least 15 times the age-specific risk of ovarian cancer of noncarriers. Among women predicted to carry the gene, the cumulative risk of developing ovarian cancer by the age of 59 years is approximately 10%.

Wingo; Project Associates: Nancy C. Lee, Mi- chele G. Mandel, and Herbert B. Peterson. Data Collection Centers Principal Investigators: At- lanta: Raymond Greenberg; Connecticut: J. Wis- ter Meigs and W. Douglas Thompson; Detroit: G. Marie Swanson; Iowa: Elaine Smith; New Mexico: Charles Key and Dorothy Pathak; San Francisco: Donald Austin; Seattle: David Thomas; Utah: Joseph Lyon and Dee West. Pa- thology Review Principal Investigators: Fred Gorstein, Robert McDivitt, and Stanley J. Rob- boy. Project Consultants: Lonnie Burnett, Rob- ert Hoover, Peter M. Layde, Howard W. Ory,

James J. Schlesselman, David Schottenfeld, Bruce Stadel, Linda Webster, and Colin White. Pathology Consultants: Walter Bauer. William Christopherson, Deborah Gersell, Robert Kur- man, Allen Paris, and Frank Vellios.

Address for reprints: Elizabeth B. Claus, Ph.D., M.D., Yale University, Department of Epidemiol- ogy and Public Health, 60 College St., P.O. Box 208034 New Haven, CT 065204034,

Received October 18, 1995; revision received February 20,1996; accepted February 20,1996.

0 1996 American Cancer Society

Attributable Risk of Breast and Ovarian CancerKlaus et al. 2319

CONCLUSIONS. The estimates provided may prove helpful to clinicians until such time as large scale population-based screening for breast and ovarian cancer susceptibility genes is possible. Cancer 1996; 77:2318-24. 0 1996 American Cancer Society.

KEYWORDS: breast cancer, ovarian cancer, family history, genetic epidemiology, attributable risk.

esearchers have recently identified a gene associated R with the development of breast and ovarian cancer in some families.'-' This gene, named BRCAl and located on chromosome 17q, is thought to account for the majority of ovarian as well as a high percentage of breast cancer in families in which both early onset breast cancer and ovarian cancer O C C U ~ . ~ In addition, a second gene, BRCA2, has been located on chromosome 13q, which also confers an increased risk of breast and ovarian cancer within high risk families.*," Although these two genes appear to explain the majority of breast and ovarian cancers within these high risk families, the proportion of breast and ovarian cancer within the general population that may be due to such breadovarian susceptibility genes remains unknown.

The current study uses information on breast and ovarian cancer from the Cancer and Steroid Hormone Study, a large, population-based data set, to estimate the proportion of breast and ovarian cancer cases in the general population that are likely to be due to a breast/ovarian cancer susceptibility gene(s), presented as an overall estimate as well as by age at onset. In addition, the age- specific penetrance of ovarian cancer for women predicted to be carriers of a susceptibility gene is calculated. These estimates should prove helpful to clinicians inter- ested in providing cancer risk estimates to their patients until such time as large-scale population-based screening for BRCA1, BRCA2, and other mutations can take place.

STUDY POPULATION Data were obtained from the Cancer and Steroid Hor- mone Study, a multicenter, population-based, case-control study conducted by the Centers for Disease Control. The data set includes 4730 breast cancer cases as well as 4688 controls who were selected through random digit dialing and matched to the breast cancer cases by geo- graphic region and 5-year age intervals. All study partici- pants were between 20 and 54 years of age at the time of interview. A detailed description of the study may be found elsewhere."

Patients were interviewed about the occurrence of breast and ovarian cancer in specific female relatives. For all of the analyses in this study, only the mothers and sisters of cases are included. Second-degree relatives and nonwhites are excluded due to underreporting of both

breast and ovarian cancer as determined by comparison with annual age-adjusted cancer rates from the Surveil- lance, Epidemiology, and End Results (SEER) program. Daughters are also excluded from all analyses because the small numbers of daughters affected with breast or ovarian cancer precluded any meaningful analyses. A first-degree family history is therefore defined to include only mothers and sisters.

METHODS Genetic models previously fit to the breast cancer cases in these data have provided evidence for the existence of one or more rare autosomal dominant genes that lead to increased susceptibility to breast cancer.'' These analyses indicate that the two most important pieces of information to be used in determining whether a woman carries such a susceptibility gene are the number and ages at onset of any first-degree relatives diagnosed with breast cancer. We used this information in conjunction with the previously fit autosomal dominant genetic model to clas- sify the breast cancer cases into carrier probability classes. The model was fit, using the method of maximum likelihood, to the age-specific incidence data of breast cancer among first-degree relatives of the breast cancer cases and controls." Briefly described, the likelihoods for this model were computed as a joint analysis of mothers and sisters of the white cases and controls. The likelihood for the mother of a case was calculated conditional upon the age at onset of the case. For mothers of controls, the likelihood was calculated conditional upon the current age of the control. The likelihood for sisters of cases was calculated conditional upon the breast cancer status of the mother as well as the age at which the case was affected. The age at which a mother with breast cancer was affected, or, in the instance of an unaffected mother, her current age or age at death was also incorporated into the sister's likelihood. For sisters of controls, the likelihood was calculated in the same fashion with the excep- tion that current age of the control was substituted for age at onset of the case.

For the underlying genetic model, we make the assumption that susceptibility to breast cancer and age at onset of breast cancer are both regulated by the same single diallelic locus and that transmission of the abnormal allele, A, follows an autosomal dominant inheritance pattern. The

2320 CANCER June 1,1996 / Volume 77 / Number 11

age at onset distribution is normally distributed. Two age at onset distributions were therefore estimated: one for carriers of the abnormal allele and one for noncarriers, each of which has an estimated genotype-specific mean and vari- ance, p g and crg (g = AA, Aa, aa), respectively. The model parameters, which also include the gene frequency, q, and genotype specific penetrances of breast cancer, kbg, were estimated using the maximum likelihood computer program MAXLIK.” The values of the model parameters were estimated as follows: pM (= pAa) = 55.435 years, pa = 68.990

= 0.100, and q = 0.0033. Under this model, the probability that a given breast

cancer case is a carrier of the abnormal gene is a function of her age at onset and her family history of breast cancer as well as the age at onset of any relatives affected with breast cancer. Breast cancer cases with multiple family members affected at early ages are most likely to be carriers. More specifically, the probability that a breast cancer patient has genotype i (i = Aa or aa) is calculated using the likelihood equations defined earlier, i.e., conditional upon the breast cancer status of her mother and sisters as well as upon the ages at onset of any affected relatives. If a breast cancer patient has no first-degree relatives affected with breast cancer, then the current age of these relatives is substituted for age at onset. The values of the model parameters used in these calculations are as listed above. The details of this model are presented elsewhere.”

Once the carrier probabilities for the breast cancer cases are calculated, they are stratified into 5 probability groups; 0.00-0.09, 0.10-0.19, 0.20-0.39, 0.40-0.79, and 0.80-0.99. The five groups are selected to contain suffi- cient numbers of relatives at risk. The observed age-specific Kaplan-Meier risks of ovarian cancer in mothers and sisters of the five stratified breast cancer case groups are then computed. The generalized Wilcoxon (Breslow) log rank test (BMDPIL)’3 is used to compare the Kaplan- Meier risk estimates by carrier probability grouping. A Cox regression model is used to test the assumption of proportional hazards by carrier probability grouping and to measure the relative increase in the hazard by probability grouping (BMDP2L).I3

These age-specific Kaplan-Meier risks are used to calculate the proportion of ovarian cancer cases in the general population that are due to the susceptibility gene by comparing the ovarian cancer risks in first-degree relatives of breast cancer patients who are likely to be carriers (i.e., women with a carrier probability of 0.80-0.99) with the ovarian cancer risks in first-degree relatives of breast cancer patients who are unlikely to be a carrier (i.e., women with a carrier probability of 0.00-0.09). Note that under an autosomal dominant model, first-degree relatives of breast cancer patients with a low (0.00-0.09) carrier probability are primarily normal homozygotes (i.e.,

years, CTM (= C T h = 0,) = 15.387, kbm (= X ~ A ~ ) = 0.928, hbaa

they also do not carry the breast cancer allele). The Kaplan-Meier estimates for these relatives will be used as ovarian cancer risk estimates for noncarriers. Among first-degree relatives of breast cancer cases with a high (0.80-0.99) carrier probability, approximately 50% are predicted to carry the allele, whereas the remaining 50% are normal homozygotes; the observed ovarian cancer risk estimates thus represent an average across the two genotypes. Therefore, the risk to relatives who are hetero- zygotes (carriers) is calculated as two times the observed Kaplan-Meier estimate for the relatives in the 0.80-0.99 group minus the value for the relatives in the 0.00-0.09 group. The proportion of ovarian cancer cases, Oi, who are likely to carry the breast/ovarian susceptibility allele within a given age at onset group (i) is estimated using the following equation:

in which q = 0.0033 represents the gene frequency of the abnormal allele A, p = 1 - q, and KMg,i is the age-specific Kaplan-Meier risk of ovarian cancer within the ith age interval for an individual with genotype g (g = Aa or aa). Due to the rarity of the gene, the proportion of individuals who are classified as abnormal homozygotes (g = AA) is extremely small (none have been described in the cohort of genotyped families to date) and hence not included in any of the calculations.

The age-specific and cumulative lifetime risks of breast cancer for carriers and noncarriers have been previously calculated for these data.” For carriers, the cumulative risk of breast cancer by age 29, 39, 49, 59, 69, 79, and 80+ years based on the assumption that age at onset is normally distributed is estimated to be 0.039, 0.132, 0.313, 0.546, 0.752, 0.869, and 0.928, respectively. For noncarriers, the values are 0.0005, 0.003, 0,010, 0.026, 0.050, 0.074, and 0.100, respectively. These numbers, along with q = 0.0033, may be used to calculate Bi , the proportion of breast cancer patients who are likely to carry the susceptibility gene within a given age at onset group (i).

Within a given age at onset group (i), the risk attributable to a breastlovarian susceptibility gene is calculated by noting that the following relationship exists between OA,,i (or BAa,i) and the attributable risk (ARJ:

RESULTS Table 1 presents the Kaplan-Meier estimates and standard errors of age-specific and cumulative risk of ovarian


TABLE 1 Cumulative Probability (%) (Standard Error) of Ovarian Cancer in Mothers and Sisters of Breast Cancer Patients by Case Probability of Being a Gene Carrier

Age of Relative (yrs) 0.0-0.09 0.10-0.19 0.20-0.39 0.40-0.79 0.80-0.99

0-24 0.02 (0.02) 0 0 0.13 (0.13) 0 25-29 0.06 (0.03) 0 0 0.13 (0.13) 0 30-34 0.06 (0.03) 0 0 0.28 (0.20) 0.51 (0.50)

1.84 (1.06) 35-39 0.13 (0.04) 0.19 (0.091 0 40-44 0.20 (0.06) 0.39 (0.14) 0.17 (0.17) 0.46 (0.27) 3.34 (1.49) 45-49 0.31 (0.07) 0.45 (0.16) 0.56 (0.33) 0.69 (0.35) 3.34 (1.49) 50-54 0.44 (0.09) 0.69 (0.21) 0.82 (0.41) 0.69 (0.35) 3.34 (1.49) 53-59 0.52 (0.11) 0.96 (0.26) 1.13 (0.52) 0.69 (0.35) 4.90 (2.13) 60-64 0.68 (0.13) 1.08 (0.29) 1.14 (0.57) 65-69 0.87 (0.16) 1.39 (0.42) 1.14 (0.57) 70-74 0.98 (0.18) 2.33 (1.31) 713-79 1.31 (0.29) 8 0 1 1.80 (0.46)

0.28 (0.20)

No. at risk Number affected

6747 40

3038 16

1023 5

772 6

242 6

Log rank test chi-square = 37.292; P < 0.0001; df (degrees of freedom) = 4.

cancer to mothers and sisters of the breast cancer cases calculated according to the breast cancer case's probability of being a carrier. Cases with the highest probability (0.80-0.99) of being a carrier reported significantly more first-degree relatives affected with ovarian cancer (log rank test = 37.29, P < 0.0001). Compared with relatives of cases with a carrier probability of 0.00-0.09, the hazard ratio for a first-degree relative of a case with carrier probability 0.80-0.99 was 8.18 (95% confidence interval [CI], 3.38-19.83) versus 1.82 (95% CI, 0.75-4.391, 1.39 (95% CI, 0.53-3.591, and 1.45 (95% CI, 0.80-2.66) for relatives of cases with a carrier probability of 0.40-0.79, 0.20-0.39, and 0.10-0.19, respectively. By age 59, the oldest age at which relatives of cases with a carrier probability of 0.80- 0.99 were observed to be diagnosed with ovarian cancer, the combined cumulative risk of ovarian cancer to first- degree relatives (averaged across carriers and noncarriers) was 4.90% versus 0.52% for first-degree relatives of cases with carrier probabilities of 0.00-0.09. Because approximately 50% of the relatives of cases with a high carrier probability are themselves estimated to be carriers, the ovarian cancer risk to age 59 years for these women is estimated to be (2 x [4.90%]) - 0.52% = 9.28%. Simi- larly, the hazard rate for carriers is thus estimated as ([2 x 8.181 - 1) = 15.4 times that of noncarriers. Using a Cox regression model, this ratio is found to be constant across age at onset groupings. The ovarian cancer risk for carriers may thus be estimated by multiplying by 15.4 the age- specific risks for the relatives of patients with a carrier probability of 0.00-0.09; this method yields a risk of approximately 27% by age 80 years. As a check on the extent

to which the cumulative ovarian cancer risk estimates are affected by error associated with the gene frequency estimate of 0.0033, the risk estimates were recalculated replacing the gene frequency with this parameter plus or minus twice the standard error (0.0012). The ovarian cancer risk estimates were not significantly altered by these changes.

The age-specific proportion of ovarian and breast cancer cases estimated to carry the breast/ovarian susceptibility gene as well as the age-specific proportion of cases that are likely to be attributable to the gene are presented in Table 2. The proportion of breast cancer cases that are likely to be due to the gene varies inversely with age at onset and ranges from a high of 33% among women diagnosed in their 20s to a low of less than 1% among women diagnosed in their 80s. Among ovarian cancer patients, a lower proportion of cases are attributed to the gene within each age group than was true for breast cancer patients. In addition, the proportion of cases that are likely to be due to the gene varies less markedly with age at onset. Approximately 14% of ovarian cancers among women diagnosed in their 30s are attributed to the gene versus 7% among women diagnosed in their 50s.

DISCUSSION Published statistical estimates of the proportion of breast cancer in the general population that is likely to be attributable to an inherited mechanism range from approximately 6- 19%,1','4-'6 depending upon the type of relative included in the calculation. When based solely upon information obtained from first-degree relatives, this risk


TABLE 2 analvsis of 33 breastlovarian cancer families of the Breast Distribution of Genotype (and Attributable Risk) among Breast and Ovarian Cancer Cases by Age at Onset

Genotype

Age at onset (yrs) Aaa aa

Breast cancer cases 20-29 30-39 40-49 50-59 60-69 70-79 80 t

Ovarian cancer cases 0-39 40-49 50-59

0.3311 (0.3267) 0.2216 (0.2164) 0.1406 (0.1349) 0.0860 (0.0799) 0.0514 (0.0451) 0.0302 (0,0233) 0.0144 (0.0078)

0.1495 (0.1438) 0.0916 (0.0856) 0.0782 (0.0720)

0.6683 0.7782 0.8592 0.9138 0.9485 0.9697 0.9855

0.8505 0.9084 0.9218

’A is the allele conferrine risk

is widely estimated to be approximately 6-7%, averaged across all ages at onset. However, many studies have found evidence for an inverse relationship between the risk of breast cancer and age at onset with a greater proportion of early onset breast cancer cases being likely to represent familial cases than late onset cases.”-” This result is reflected in the age-specific estimates of genetic attributable risk presented here. In these data, approximately one-third of all breast cancers diagnosed between the ages of 20-29 years are estimated to be associated with an autosomal dominant gene. Conversely, the great majority of all breast cancer diagnosed in women in their 60s, 70s, or 80s is not considered to be due to an inherited mechanism. Less is currently known regarding the extent to which ovarian cancer may be attributed to an inherited mechanism. In these data, we calculate this overall estimate to be approximately 10%. However, unlike for breast cancer, this risk does not appear to vary as markedly by the age at which a woman is diagnosed, a finding that seems to concur with that of other researcher^.^^^'

In addition to estimating the proportion of carriers in the general population, it is also of interest to estimate the extent to which carriers will become affected with either breast or ovarian cancer. Both segregation and linkage analyses predict a high lifetime risk of breast cancer for carriers of a breastlovarian cancer susceptibility gene.4.”,22 For the current data, the risks by ages 60 and 80 years of breast cancer for women who are carriers are estimated to be 55% and 87%, respectively.” For ovarian cancer, the risk to ages 60 and 80 years is estimated to be approximately 10% and 27%, respectively. These values can be compared with those obtained from a recent

Cancer Consortium, each of which contained at least 4 cases of ovarian cancer or breast cancer diagnosed before the age of 60 years that showed evidence of linkage to BRCAl.22-24 For these families, the cumulative risk to age 70 years of developing breast and ovarian cancer for BRCAl gene carriers was estimated to be 767bz5 and 63%,24 respectively. One can see that the values for cumulative breast cancer risk within the two studies are similar whereas the value for cumulative ovarian cancer risk is much greater in the Consortium data. One possible reason for the discrepancy is the fact that in the Consortium data, each analyzed pedigree contained at least four ovarian or early onset breast cancer cases whereas in the Can- cer and Steroid Hormone Study data there were essen- tially no families with this reported configuration of cancer, due in some part to the smaller pedigree sizes. Of interest, however, is the fact that the members of the Consortium found significant evidence of heterogeneity of risk between families when a model that assumed the existence of two BRCA genes was fit to the data.23 The results of that analysis indicate that although both genes confer a breast cancer risk of approximately 76%, the first gene is associated with an ovarian cancer risk of 26% by age 70 years (similar to the risk values presented here), whereas the second gene confers an ovarian cancer risk of 85% by age 70 years with the first gene representing the majority of all mutation^.'^ These results suggest that our data may allow us to estimate the effects of the first more common BRCAl gene and less so that of additional rare BRCA genes that may occur primarily in extremely high risk families and that are quite infrequent in the general population. In addition, some proportion of familial breast cancer not due to BRCAl is caused by other genes, such as BRCA2, that confer a high risk of breast cancer but a lesser risk of ovarian cancer.

Although the findings presented here match reason- ably well to recent linkage data, caution should be exer- cised when interpreting our results for a number of rea- sons. Because the risks presented here are calculated solely on the basis of reported family history information, it is important to assess the quality of this information. In this study, data on ovarian and breast cancer in relatives was collected only through interviews of cases and controls. Although an examination of the observed versus expected rates of reported breast cancer among first-degree relatives of controls matched well to rates of the Surveillance, Epidemiology, and End Results Program (SEER), the number of first-degree relatives of controls reported to have ovarian cancer was 79% of the expecta- tion.26 However, the hazard ratio estimate of 15.4 for carriers versus noncarriers was based solely on family history information from 2 subsets of the same dataset and there-


fore should be more resistant to the effects of reporting bias.

By virtue of the fact that family history information as reported by the proband is used in this analysis, rather than information as confirmed by pathologic report, it is important to note that the definition of breast and ovarian cancer includes all histologic subtypes. For example, the term breast cancer includes diagnoses of both ductal and lobular carcinoma. Although at present histologic charac- teristics do not appear to define genetically distinct sub- groups of breast andlor ovarian cancer,2i these data are preliminary and further study will help to elucidate the nature of the relationship between histopathology and a family history of either breast or ovarian cancer. An additional limitation of these analyses is the fact that they include only data from white women. Generalization to nonwhite populations is thus not possible from these data; additional studies are needed to determine the prevalence of breast and ovarian cancer genes in a wide vari- ety of racial and ethnic groups.

The calculated estimate of the lifetime risk of ovarian cancer for carriers versus noncarriers is based upon the assumption that, among first-degree relatives of breast cancer cases with a carrier probability of 0.80-0.99, 50% are themselves expected to be carriers. Within this carrier probability group, approximately one-third of the breast cancer cases had a probability of 0.90-0.99 of being a carrier. This number indicates that somewhat less than 50% of the relatives of cases in the 0.80-0.99 group are likely to be carriers and that the true age-specific risks of ovarian cancer to carriers may be higher than those presented here. Although it is difficult to determine the exact probabilistic cutpoint above which women should be classified as carriers and below which women should be classified as noncarriers, the calculated estimate of the lifetime risk of ovarian cancer for carriers should be interpreted as a lower bound; women who are carriers are thus likely to have at least a 15-fold increase in ovarian cancer risk compared with noncarriers.

The estimates of attributable risk presented in this study are calculated using a gene frequency that most probably represents a population-based average of a number of genes involved in the development of breast and ovarian cancer, including BRCAl and BRCA2. Estimates of the attributable risk of specific germline breast cancer genes such as BRCAl are beginning to emerge?5,28-32 As in these data, evidence continues to be found for an inverse relationship between age at onset and genetic attributable risk. Ford et al. estimate that the proportion of breast cancer in the general population due to BRCAl is 5.3% in women younger than the age of 40 years, 2.2% in women between the ages of 40 and 49 years, and 1.1% in women between the ages of 50 and 70 years.25 Their corresponding estimates for ovarian cancer are 5.7%, 4.6%, and 2.1%,

respectively.” These estimates are markedly lower than those obtained in this analysis. This may be explained by the fact that BRCAl represents one gene likely to be repre- sented in our analysis. Furthermore, new laboratory data indicate that the proportion of breast cancer associated with BRCAl may be higher than initially predicted by the Breast Cancer Consortium A recent s t u d y of 80 women in whom breast cancer was diagnosed before the age of 35 years reported that approximately 10% of these women carried germline alterations in the BRCAl gene whereas a second study reported a mutation rate of 13% in a group of breast cancer patients diagnosed before the age of 30 year^.^" Germline BRCAl mutations, particularly the 185delAG mutation, have been identified at even higher rates among samples of Jewish women with a 1% overall rate and a 21% prevalence rate among Jewish women diagnosed with breast cancer before the age of 40 year^.^'.^'

Although BRCAl and BRCA2 appear to explain the majority of early onset breast cancer cases, a number of families with large numbers of early onset breast cancer cases have been shown to be unlinked to BRCAl and BRCA2.1,4t8 Additional genes have already been implicated in the development of breast ~ance?~-~‘ ; a recent study associated 1 in 11 cancers of the breast in the general population with rare alleles of a minisatellite locus adja- cent to the HRASl gene located on chromosome 1 l.74 In addition, the p53 gene has also been associated with the development of breast cancer in families characterized by the Li-Fraumini syndrome.” The extent to which germline mutations in other breast cancer susceptibility gene^^''^^ play a role in the expression of breast cancer will be determined once the genes have been cloned and population-based screening begun. The search for mutations in BRCAl is predicted to be extremely challenging’ and time-consuming due to the large size of the gene and the fragmented nature of the coding sequence, as well as early indications that no one mutation occurs with any great f r e q ~ e n c y . ’ ~ ~ ~ ” * ~ ~ BRCA2 has been found to be an even larger gene than BRCAl. It will most likely take months to years to define the population-based risks associated with different BRCAl, BRCA2, and other genetic mutations involved in breast and ovarian cancer. The data provided here give interim estimates until such time as these molecular data become available.

REFERENCES 1. Miki Y, Swensen J, Shattuck-Eidens D, Futreal A, Harshman

K, Tavtigian S, et al. A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science 1994;

Hall JM, Lee MK, Newman B, Morrow JE, Anderson LA, Huey B, et al. Linkage of early-onset familial breast cancer to chromosome 17q21. Science 1990;250:1684-9.

266:66-71. 2.


3.

4.

5.

6.

7.

8.

9.

Narod SA, Feunteun J, Lynch HT, Watson P, Conway T, Lynch J, et al. Familial breast-ovarian cancer locus on chromosome 17q12-q23. Lancet 1991;338:82-3. Easton DF, Bishop DT, Ford D, Crockford GP, and the Breast Cancer Linkage Consortium. Genetic linkage analysis in familial breast and ovarian cancer: results from 214 families. Am J Hum Genet 1993;52:678-701. Castilla LH, Couch FJ, Erdos MR, Hoskins KF, Calzone K, Garber JE, et al. Mutations in the BRCAl gene in families with early-onset breast and ovarian cancer. Nut Genet 1994;

Simard I, Tonin P, Durocher F, Morgan K, Rommens J, Gin- gras s, et al. Common origins of BRCAl mutations in Cana- dian breast and ovarian cancer families. Nut Genet 1994;

Friedman LS, Ostermeyer EA, Szabo CI, Dowd P, Lynch ED, Rowell SE, et al. Confirmation of BRCAl by analysis of germline mutations linked to breast and ovarian cancer in ten families. Nut Genet 1994; 8:399-404. Wooster R, Neuhausen SL, Mangion J, Quirk Y, Ford D, Col- lins N, et d. Localization of a breast cancer susceptibility gene, BRCA2, to chromosome 13q12- 13. Science 1994;265: 2088-90. King MC. Late-breaking research session. American Society of Human Genetics Annual Meeting. Montreal Quebec, Can- ada, 1994.

8:387-91.

8~392-8.

10. Wingo PA, Ory HW, Layde PM, Lee NC, and the Cancer and Steroid Hormone Study Group. The evaluation of the data collection process for a multicenter, population-based case- control design. Am J Epidemiol 1988; 128:206- 17.

11. Claus EB, Risch N, Thompson WD. Genetic analysis of breast cancer in the Cancer and Steroid Hormone Study. Am J Hum Genet 1991;48:232-42.

12. Kaplan EB, Elston RC. A subroutine package for maximum likelihood estimation (MAXLIK). Inst Stat Mimeo Ser 823. Chapel Hill: University of North Carolina, 1972.

13. Dixon WJ, (editor). BMDP statistical software. Los Angeles: Regents of the University of California, 1990.

14. Slattery ML, Kerber RA. A comprehensive evaluation of family history and breast cancer risk: the Utah Population Data- base. JAMA 1993;270:1563-8.

15. King MC, Rowell S, Love SM. Inherited breast and ovarian cancer. What are the risks? What are the choices? JAMA

16. Colditz GA, Willett WC, Hunter DJ, Stampfer MI, Manson JE, Hennekens CH, et al. Family history, age, and risk of breast cancer. JAMA 1993;270:338-43, 1548.

17. Claus EB, Risch N, Thompson WD. Age at onset as an indica- tor of familial risk of breast cancer. A m J Epidemiol 1990; 13 1:961- 72.

18. Harris JR, Lippman ME, Veronesi U, Willett W. Breast cancer (First of three parts) N Engl J Med 1992;327:319-28.

19. Lynch HT, Watson P, Conway T, Fitzsimmons SL, Lynch 1. Breast cancer family history as a risk factor for early onset breast cancer. Breast Cancer Res Treat 1988; 11:

20. Newman B, Austin MA, Lee M, King MC. Inheritance of human breast cancer: evidence for autosomal dominant

1993;269:1975-80.

263-7.

transmission in high risk families. Proc Nut1 Acad Sci USA

21. Schwartz AG, King MC, Belle SH, Satariano WA, Swanson GM. Risk of breast cancer to relatives of young breast cancer patients. J Natl Cancer Inst 1985;75:665-8.

22. Ford D, Easton DF, Bishop DT, Narod SA, and the Breast Cancer Consortium. The risks of cancer in BRCAl mutation carriers. Am J Hum Genet 1993; 53:A298.

23. Easton D, Ford D, Bishop DT, and the Breast Cancer Linkage Consortium. Breast and ovarian cancer incidence in BRCAl mutation carriers. A m J Hum Genet 1995;56:265-71.

24. Ford D, Easton DF, Bishop DT, Narod SA, Goldgar DE. Risk of cancer in BRCAl-mutation carriers. Lancet 1994;343:692-5.

25. Ford D, Easton DF, Pet0 I. Estimates of the gene frequency of BRCAl and its contribution to breast and ovarian cancer incidence. Am J Hum Genet 1995;57:1457-62.

26. Schildkraut lM, Thompson WD. Familial ovarian cancer: a population-based case-control study. Am J Epidemiol 1988;

27. Claus EB, Risch NR, Thompson WD, Carter D. Relationship between breast histopathology and family history of breast cancer. Cancer 1993;71:147-53.

28. Shattuck-Eidens D, McClure M, Simard J, Labrie F, Narod S, Couch F, et al. A collaborative study of 80 mutations in the BRCAl breast and ovarian cancer susceptibility gene. lmplications for presymptomatic testing and screening.

29. Langston AA, Malone KE, Thompson JD, Daling JR, Ostrander EA BRCAl mutations in a population-based sample of young women with breast cancer. N EnglJ Med 1996334:137-42.

30. Fitzgerald MG, MacDonald DJ, Krainer M, Hoover 1, O’Neil E, Unsal H, et al. Germ-line BRCAl mutations in Jewish and non-Jewish women with early onset breast cancer. N Engl J Med 1996; 334: 143-9.

31. Struewing JP, Abeliovich D, PeretzT, Avishai N, Kaback MM, Collins FS, et al. The carrier frequency of the BRCAl 185de- lAG mutation is approximately 1 percent in Ashkenazi Jew- ish individuals. Nut Genet 1995; 11:198-200.

32. Serova 0, Montagna M, Torchard D, Narod SA, Tonin P, Sylla B, et al. A high incidence of BRCAl mutations in 20 breast- ovarian cancer families. Am J Hum Genet 1996;58:42-51.

33. Krontiris TG, Devlin B, Karp DD, Robert NJ, Risch N. An association between the risk of cancer and mutations in the Hrasl minisatellite locus. N Engl J Med 1993;329:517-23.

34. Malkin D, Li FP, Strong LC, Fraumeni IF Jr, Nelson CE, Kim DH, et al. Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms. Science

35. Swift M, Reitnauer PJ, Morrell D, Chase CL. Breast and other cancers with ataxia-telangiectasia. N Engl J Med 1987;316: 1289-94.

36. King MC. Breast cancer genes: how many, where, and who are they? Nut Genet 1992;2:89-90.

37. Futreal PA, Liu Q, Shattuck-Eidens D, Cochran C, Harshman K, Tavtigian S, et al. BRCAl mutations in primary breast and ovarian carcinomas. Science 1994;266:120-2.

38. Davies K. Breaking down BRCAl [editorial]. Nut Genet 1994; 8:310.

1988; 851 -5.

128~456-66.

JAMA 1995;273:535-41.

1990; 250:1233-8.

Documents

The genetic attributable risk of breast and ovarian cancer