BRCA1 and BRCA2 mutations account for a large proportion of ovarian carcinoma cases
It is believed that BRCA1 and BRCA2 germline mutations account for the majority of hereditary ovarian carcinomas; however, to the authors' knowledge, there are scant data on the prevalence and spectrum of mutations, genotype/phenotype correlations, tumor histology, and family history characteristics. To address this gap, the authors conducted a population-based study of 232 incident epithelial ovarian carcinomas in the Tampa Bay area.
Genetic testing for the BRCA1 and BRCA2 genes was performed through full sequencing and BRCA1 rearrangement testing.
Of 209 women with invasive ovarian carcinoma, 32 women (15.3%) had mutations in BRCA1 or BRCA2, including 20 BRCA1 mutations and 12 BRCA2 mutations. Of the BRCA2 mutations, 58% were outside the “ovarian cancer cluster region” (OCCR). Variants of uncertain significance were detected in 8.2% of women with invasive ovarian carcinoma. No mutations were identified in women with borderline or invasive mucinous tumors. Among the BRCA mutation-positive women, 63% had serous tumors. A family history of breast and/or ovarian carcinoma was reported in 65%, 75%, and 43.5% of relatives of BRCA1 carriers, BRCA2 carriers, and non-BRCA1/BRCA2 carriers, respectively.
The data from this study suggested that 1) previous studies may have underestimated the frequency of BRCA1 and BRCA2 mutations in ovarian carcinomas, especially outside the OCCR; 2) it may be reasonable to offer genetic counseling to any woman with an invasive, nonmucinous epithelial ovarian tumor; and 3) among patients with invasive ovarian carcinoma, family history is not sufficiently accurate to predict mutation status. Cancer 2005. © 2005 American Cancer Society.
Ovarian carcinoma represents the seventh leading cancer among women in the U.S., accounting for 3% of all female cancers. In 2005, an estimated 22,220 new ovarian carcinomas will be diagnosed, and approximately 16,210 ovarian carcinoma-related deaths are expected.1 Established risk factors for ovarian carcinoma include age; family history of ovarian and other associated malignancies, including breast, uterine, and/or colorectal carcinomas; and inherited mutations in carcinoma-predisposition genes.2–4 The risk of ovarian carcinoma also may be increased by hormonal and reproductive factors, such as prolonged consumption of unopposed estrogen-replacement therapy, infertility and/or long-term use of fertility drugs, early age of menarche, late age at menopause, delayed childbirth, and low parity or nulliparity.1
Overall, it is believed that approximately 5–12% of invasive ovarian carcinomas are due to hereditary susceptibility.5, 6 Mutations in the BRCA1 and BRCA2 genes account for the majority of hereditary ovarian carcinomas,2, 3 followed by the MLH1 and MSH2 genes implicated in hereditary nonpolyposis colorectal carcinoma,4 yet few population-based studies, including those from the U.S., have been performed specifically to investigate the role of hereditary susceptibility genes in this deadly disease. Estimates of the frequency of BRCA1 and BRCA2 germline mutations in ovarian carcinomas have ranged between 2–12% and 2–6% for BRCA1 and BRCA2 mutations, respectively.6–21 The limitations of those studies included small sample sizes and/or case selection based on early age of diagnosis, a hospital-based series, prevalent cases, or self-referred patients. Furthermore, some of those studies were specific to certain ethnic groups with ethnicity-specific founder mutations.
Determining the proportion and characteristics of women who have epithelial ovarian carcinomas associated with mutations of the BRCA1 and BRCA2 genes is important, because it enables the appropriate provision of genetic counseling and testing services to these women and their families to aid in medical management. Few studies have addressed whether women who have ovarian carcinomas associated with BRCA1 and BRCA2 differ from each other and/or from women who have sporadic tumors with regard to the following characteristics: age at diagnosis, histologic subtype, stage at presentation, and personal and family history of cancer. To address these questions, we studied 232 unselected, population-based, incident ovarian carcinomas in 2 counties within the Tampa Bay area and examined the occurrence of mutations with respect to these characteristics.
MATERIALS AND METHODS
This study was approved by the Institutional Review Board of the University of South Florida. Written informed consent was obtained from each patient specifically for genetic testing.
The Tampa Bay Ovarian Cancer Study (TBOCS) is a population-based study in a heavily populated, 2-county region of west central Florida with a population in excess of 2 million. Because of the rapidly fatal nature of ovarian carcinoma among women who are diagnosed at late stages, we developed a rapid case-ascertainment mechanism that included 7 regional gynecologic and medical oncologists who treat 85% of women with ovarian carcinoma in the area. According to the most current Florida Cancer Data System (FCDS) statistics, 362 incident epithelial ovarian carcinomas in women ages 18–80 years were diagnosed between December 13, 2000 and September 30, 2003, and were eligible for participation in the current study.
The diagnosis of ovarian carcinoma was confirmed for all patients in a review of pathology records and tumor slides by a single pathologist with expertise in ovarian carcinoma. Clinical stage was determined according to International Federation of Gynecologists and Obstetricians (FIGO) criteria,22 and the histologic subtype was evaluated according to the World Health Organization (WHO) classification.23 When ambiguities in pathologic interpretation were present, the patient records were analyzed by a second ovarian carcinoma expert for a consensus.
The treating physicians introduced the study to eligible patients and informed the study team of patients who agreed to be contacted by a study member. A genetic counselor scheduled an in-person interview with the participant at the treating physician's office or at another location that was convenient for the patient. After obtaining informed consent, a questionnaire was completed, genetic counseling was provided, and a blood sample was collected. The questionnaire included information on multiple epidemiologic risk factors, including reproductive history. No TBOCS participant refused genetic testing, and all but 1 participant chose to obtain results of the BRCA1 and BRCA2 analysis.
Funding for the study covered all costs, including genetic counseling and genetic testing of the BRCA1 and BRCA2 genes. During the genetic counseling interview, family history details were obtained, including current age, age at diagnosis, and site (and, when possible, histologic type) of malignancy in three generations. Family history information was taken as that reported by the proband, because medical record verification from family members was not possible. Studies suggest that the diagnosis of carcinoma reported in relatives is likely to be accurate at least for first-degree relatives, especially in families with hereditary breast carcinomas.24–27
BRCA1 and BRCA2 Analysis
Lymphocyte DNA was extracted from whole blood by standard procedures. All samples were analyzed for BRCA1 and BRCA2 through full gene sequencing and rearrangement testing according to published protocol28 through Myriad Genetics Laboratories, Inc. This analysis identifies gene mutations in all of the known protein coding sequences and additional adjacent areas in the genes and detects approximately 90–95% of mutations in BRCA1 and BRCA2. The samples also were analyzed for the presence of 5 BRCA1 genomic rearrangements by using recombination-specific polymerase chain reaction analysis with primers that were specific for the normal gene as well as for the rearrangement, including: a 3.8-kb deletion of exon 13 and a 510-base pair deletion of exon 22 described in individuals of Dutch ancestry,29 a 6-kb duplication of exon 13 described in individuals of European (particularly British) ancestry,30, 31 a 7.1-kb deletion of exons 8 and 9 described in individuals of European ancestry,32 and a 26-kb deletion of exons 14–20.33
Confidence limits for mutation frequencies were calculated under the assumption of binomial distributions of the observed numbers of patients. To compare personal and family history characteristics between carriers and noncarriers, the contingency table chi-square test was performed, and the asymptotic P values were calculated. The Proc Freq procedure in SAS® statistical software (version 8.2) was used.
Of the 362 women with epithelial ovarian carcinoma in the 2-county region who were identified through FCDS data, we were given contact information on 267 women through our 7 regional gynecologic and medical oncologists. Of these women, 232 agreed to participate. The reasons for nonparticipation included patients were too ill (20 women) and patients refused to participate (15 women). Reasons for refusal were unwillingness to make an extra trip to participate (10 women) and patients indicated they had no female children or siblings for whom they believed the genetic testing would be useful (5 women). Thus, the overall participation rate was 64%.
The distribution of women in the TBOCS by race and ethnicity was similar to the distribution of women with ovarian carcinoma in the catchment area. The state registry data revealed that the ethnic distribution in the catchment area was 91% white and 9% nonwhite, similar to the 90.9% and 9.1% distribution observed in our study among white patients and nonwhite patients, respectively. The distribution according to histologic subtype and stage also was similar to that seen in the general population. Briefly, the distribution of histologic features expected in the general population is approximately 75% serous ovarian carcinomas and approximately 15% endometrioid ovarian carcinomas. Among women in the TBOCS, the distribution is similar, with 58.2% serous ovarian carcinomas and 12.9% endometrioid ovarian carcinomas. The distribution by stage was similar to that seen in the general population, in which approximately 30% of patients present with Stage I or II disease, and 70% present with Stages III or IV disease. A summary of data from the TBOCS is included in Table 1.
Table 1. Summary of Data on Patients in the Tampa Bay Ovarian Cancer Study
|Age at diagnosis|| || || |
|Mean age (yrs)||56.8||56.6||49.3|
|≤ 40 yrs||16 (6.9)||11 (5.3)||5 (21.7)|
|41–50 yrs||57 (24.6)||49 (23.4)||8 (34.8)|
|51–60 yrs||69 (29.7)||62 (29.7)||7 (30.4)|
|≥ 60 yrs||90 (38.8)||87 (41.6)||3 (13.0)|
|Race|| || || |
|White||211 (90.9)||191 (91.4)||20 (87.0)|
|Nonwhite||21 (9.1)||18 (8.6)||3 (13.0)|
|Prevalence of Ashkenazi Jewish ancestry||9 (3.9)||9 (4.3)||0 (0)|
|Histology|| || || |
|Serous||135 (58.2)||121 (57.9)||14 (60.9)|
|Mucinous||19 (18.2)||13 (6.2)||6 (26.1)|
|Endometrioid||30 (12.9)||29 (13.9)||1 (4.3)|
|Clear cell||9 (3.9)||9 (4.3)||0 (0.0)|
|Other||39 (16.8)||37 (17.7)||2 (8.7)|
|Prevalence of deleterious mutation|| || || |
|BRCA1||20 (8.6)||20 (9.6)||0 (0.0)|
|BRCA2||12 (5.2)||12 (5.7)||0 (0.0)|
The Prevalence of BRCA1 and BRCA2 Mutations
Of the 232 women enrolled in the study, 32 women (13.8%) had mutations in BRCA1 or BRCA2, including 20 BRCA1 mutations (8.6% of women) and 12 BRCA2 mutations (5.2% of women). Two of 32 women with BRCA1 mutations were related, and the remaining 30 women were unrelated. No mutations were identified in the 23 women who had borderline tumors. When considering only the women with invasive ovarian tumors (n = 209 women), 15.3% had mutations in BRCA1 or BRCA2. Among the 209 women with invasive tumors, no mutations were identified among the 13 women with mucinous tumors. Thus, the percentage of BRCA1 and BRCA2 mutations among the 196 women with invasive, nonmucinous tumors was 16.3%.
Table 2 shows that the BRCA1 mutations detected in this study were distributed across the length of the gene. Of the 12 BRCA2 mutations, 5 mutations (42%) were outside the OCCR, as shown in Table 2.
Table 2. Characteristics of Women who had Ovarian Carcinoma with BRCA1 and BRCA2 Mutations (n = 32)
|BRCA1|| || || || || || |
|187delAG||Ashkenazi Jewish||56||E, S||IIIC||—||Yes|
|187delAG||Mixed European||76||S||IIIC||48, 58||Yes|
|2800delAA||Mixed European||43||S, C||IIIC||—||Yes|
|K679X||Mixed European||45||E, C||IIIC||—||Yes|
|BRCA2|| || || || || || |
|E49X||Mixed European||49||S||IV||26, 41||Yes|
Of the nine women in the study (3.9%) who reported Ashkenazi Jewish ancestry, 4 women tested positive for a BRCA mutation, including 1 woman who had the BRCA1 187delAG mutation, 2 women who had the 6174 delT BRCA2 mutation, and 1 woman who had a nonfounder mutation, Q2009X.
The Prevalence of BRCA1 and BRCA2 Variants of Uncertain Significance
Variants of uncertain significance (VUS) were detected in 21 of the women enrolled in the study (9.1%), including 2 BRCA1 variants (0.86% of women) and 19 BRCA2 variants (8.2% of women), as shown in Table 3. Nineteen of 21 variants were detected in women who had invasive ovarian carcinoma, and 2 of 21 variants were detected in women who had borderline ovarian tumors. Table 3 shows that all of these variants were missense mutations. Each of the 19 detected BRCA2 variants was located outside of the OCCR. Information regarding the classification of these variants was provided by Myriad Genetic Laboratories (see Table 3). Seven of 19 BRCA2 variants (36.8%) had previously been observed with ≥ 1 deleterious mutations.
Table 3. Variants of Uncertain Significance Detected in 23 Women from the Tampa Bay Ovarian Cancer Study
|BRCA1|| || || |
|BRCA2|| || || |
| D596Hc||26-50||No||Western European|
| G1529R||26-50||Yes (multiple mutations)||Western European|
| S384F||> 50||Yes (multiple mutations)||Western European|
| N3124I||2-10||No||Western European|
| E2856A||> 50||No||Western European|
| I2285V||26-50||No||Ashkenazi Jewish|
| D2110A||One additional family||No||—|
| S2148Y||2-10||No||Western European|
| S2835P||2-10||No||Western European, Latin American, and African|
| S384F||50||Yes (multiple mutations)||Western European|
| R2973Hc||One additional family||No||Western European|
| E2586A||> 50||No||Western European|
| P655R||11-25||Yes (1 mutation)||Ashkenazi Jewish|
| S326R||> 50||Yes (multiple mutations)||Western European|
| A75P||11-25||Yes (1 mutation)||Western European|
| I2944F||> 50||Yes (multiple mutations)||African|
Personal and Family History Characteristics: Invasive Tumors (n = 209)
The average age at diagnosis was 53 years among BRCA1 carriers (n = 20 women), 58 years among BRCA2 carriers (n = 12 women), and 57 years among women with sporadic tumors (n = 177 women) (Table 4). Similarly, the age distribution for ovarian carcinoma diagnosis varied between BRCA1 carriers, BRCA2 carriers, and women with sporadic tumors, with 90%, 50%, and 56% of ovarian carcinomas, respectively, diagnosed in women younger than age 60 years. Women who were diagnosed at ages 51–60 years had the highest frequency of mutations (40.6%), and women who had ovarian carcinoma diagnosed when they were younger than age 40 years had a lower prevalence of mutations (3.1%).
Table 4. Personal and Family History of Cancer among Patients with Invasive Ovarian Carcinoma (n = 209)
|Proband|| || || || || || || || |
|Age at diagnosis|| || || || || || || || |
|≤ 40 yrs||0 (0.0)||1 (8.3)||1 (3.1)||9 (5.8)||1 (4.8)||10 (5.6)||—||—|
|41-50 yrs||8 (40.0)||2 (16.7)||10 (31.3)||35 (22.4)||4 (19.0)||39 (22.0)||—||—|
|51-60 yrs||10 (50.0)||3 (25.0)||13 (40.6)||40 (25.6)||9 (42.9)||49 (27.7)||—||—|
|> 60 yrs||2 (10.0)||6 (5f0.0)||8 (25.0)||72 (46.2)||7 (33.3)||79 (44.6)||—||—|
|Mean age (yrs)||53||58||55||58||56||57||—||—|
|Prior breast carcinoma||6 (30.0)||3 (25.0)||9 (28.0)||8 (5.1)||1 (4.8)||9 (5.1)||7.2||< 0.0001|
|Family history|| || || || || || || || |
|FDR or SDR with breast or ovarian carcinoma||13 (65.0)||9 (75.0)||22 (69.0)||65 (41.7)||12 (57.1)||77 (43.5)||2.9||0.0085|
|≥ 1 FDR with breast carcinoma||4 (20.0)||5 (42.0)||9 (28.0)||30 (19.2)||3 (14.3)||33 (18.7)||1.2||0.2181|
|≥ 1 FDR with ovarian carcinoma||6 (30.0)||2 (17.0)||8 (25.0)||8 (5.1)||1 (4.8)||9 (5.1)||6.2||0.0001|
|≥ 1 SDR with breast carcinomac||8 (40.0)||4 (44.0)||12 (38.0)||38 (24.4)||8 (38.1)||46 (26.0)||1.7||0.1808|
|≥ 1 SDR with ovarian carcinomac||3 (15.0)||1 (8.0)||4 (13.0)||7 (4.5)||1 (4.8)||8 (4.5)||3.0||0.0741|
|No FDR or SDR with breast or ovarian carcinoma||7 (35.0)||3 (25.0)||10 (31.0)||91 (58.3%)||9 (42.9)||100 (56.5)||2.9||0.0085|
Almost 30% (n = 9 patients) of the women with BRCA1 and BRCA2 mutation-positive ovarian carcinoma had a personal history of breast carcinoma, with an average age at breast carcinoma diagnosis of approximately 43 years. This compared with 5% (n = 8 patients) of the 156 women with sporadic tumors who had a personal history of breast carcinoma, with an average age at diagnosis of 52.8 years.
The proportion of women with invasive carcinoma who had a positive family history of breast and/or ovarian carcinoma in a first-degree or second-degree relative was 65.0% among BRCA1 carriers (n = 20 patients), 75.0% among BRCA2 carriers (n = 12 patients), and 43.5% among women with sporadic tumors (n = 177 patients) (Table 4). Thirty-one percent of mutation carriers had no family history of breast or ovarian carcinoma in any first-degree or second-degree relative.
Reproductive and Other Factors and Ovarian Cancer Risk
Comparisons of reproductive and other factors (such as age at menarche, age at first live birth, parity, history of infertility, duration of hormone replacement therapy, and body mass index) among carriers and noncarriers (with invasive or borderline tumors) showed no statistically significant differences between the groups.
The majority (63%) of the BRCA1 and BRCA2-associated tumors were of serous histology, and none were mucinous or borderline tumors (Table 5). Three histologic subtypes seen among the BRCA2 carriers were not observed among BRCA1 carriers: primary peritoneal, Brenner, and clear cell. The distribution by stage of BRCA carriers overall was similar to that seen in the general population, in which approximately 30% present with Stage I or II disease, and 70% present with Stage III or IV disease (Table 5). When stratifying BRCA1 and BRCA2 carriers by stage, 20% (4 of 20 women) of BRCA1 carriers and 42% (5 of 12 women) of BRCA2 carriers presented with early-stage disease (i.e., Stage I or II). The majority of BRCA carriers had poorly differentiated tumors.
Table 5. Tumor Histology: Subtype, Stage, and Grade among BRCA Mutation Carriers and Noncarriers
|Tumor type|| || || || || || || |
|Serous||135 (58.2)||121 (57.9)||14 (60.9)||14 (70)||6 (50)||20 (63)||115 (57.5)|
|Endometrioid||30 (12.9)||29 (13.9)||1 (4.3)||1 (5.0)||2 (16.8)||3 (9.0)||27 (13.5)|
|Transitional||3 (1.3)||3 (14.4)||0 (0.0)||1 (5.0)||0 (0.0)||1 (3.0)||2 (0.1)|
|Mucinous||19 (8.2)||13 (6.2)||6 (26.1)||0 (0.0)||0 (0.0)||0 (0.0)||19 (9.5)|
|Mixed||20 (8.6)||18 (8.6)||2 (8.7)||4 (20.0)||1 (8.3)||5 (16)||15 (7.5)|
|Peritoneal||6 (2.6)||6 (2.9)||0 (0.0)||0 (0.0)||1 (8.3)||1 (3.0)||5 (2.5)|
|Brenner||3 (1.3)||3 (1.4)||0 (0.0)||0 (0.0)||1 (8.3)||1 (3.0)||1 (1.0)|
|Clear cell||9 (3.9)||9 (4.3)||0 (0.0)||0 (0.0)||1 (0.0)||0 (0.0)||9 (4.5)|
|Unknown||7 (3.0)||7 (3.3)||0 (0.0)||0 (0.0)||1 (8.3)||1 (3.0)||6 (3.0)|
|Stage|| || || || || || || |
|I||47 (20.3)||28 (13.4)||19 (82.6)||1 (5)||4 (33.4)||5 (15.6)||42 (21.0)|
|II||20 (8.6)||20 (9.6)||0 (0.0)||3 (15)||1 (8.3)||4 (12.5)||16 (8.0)|
|III||136 (58.6)||132 (63.15)||4 (17.4)||13 (65)||6 (50)||19 (59.4)||117 (58.5)|
|IV||28 (12.1)||28 (13.4)||0 (0.0)||3 (15)||1 (8.3)||4 (12.5)||24 (12.0)|
|Unknown||1 (0.4)||1 (0.45)||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||1 (0.5)|
|Grade|| || || || || || || |
|1||46 (20.4)||27 (12.9)||19 (82.6)||3 (15)||3 (25)||6 (18.75)||40 (20.0)|
|2||43 (19.0)||42 (20.1)||2 (8.7)||3 (15)||1 (8.3)||4 (12.5)||39 (19.5)|
|3||137 (60.6)||136 (65.1)||1 (4.35)||14 (70)||8 (66.7)||22 (68.75)||115 (57.5)|
|Unknown||6 (2.6)||4 (1.9)||1 (4.35)||0 (0.0)||0 (0.0)||0 (0.0)||6 (3.0)|
Although previous estimates have suggested that at least 10% of invasive ovarian carcinomas are due to mutations in BRCA1 and BRCA2, results from the current study indicate that the proportion actually may be as high as 15% when a comprehensive mutation-detection strategy is employed. The largest population-based study to date was performed in Ontario, Canada,6 and was based on 649 unselected women with ovarian carcinoma, including 515 women with invasive ovarian carcinoma. The methodology for mutation detection in that study included screening of exon 11 of BRCA1, screening of exons 10 and 11 of BRCA2 through protein truncation testing, and screening for 11 common mutations by rapid multiplex methods (including 3 Jewish founder mutations and 6 French-Canadian mutations). Using this limited mutation-detection strategy, among the 515 women with invasive ovarian carcinoma, 60 mutations (11.7%; 95% confidence interval [95% CI], 9.2–14.8%) were identified, including 39 mutations that were identified in BRCA1 and 21 mutations that were identified in BRCA2. More recently, in a Swedish population-based study, 161 women with invasive epithelial ovarian carcinoma were investigated, and 13 patients (8%) were identified with deleterious mutations.13 Although the strategy for mutation detection in that study included several molecular techniques, full gene sequencing was not performed; thus, mutations may have been missed. In a Polish study, 364 unselected women with ovarian carcinoma were investigated for 3 common founder BRCA1 mutations, and 49 deleterious mutations were detected (13.5%).14
In the current study, among the 209 women with invasive ovarian carcinoma, 32 mutations were identified (15.3%; 95% CI, 11–21%), including 20 mutations in BRCA1 and 12 mutations in BRCA2. The higher rate of mutation detection compared with the Ontario study6 was due in part to the more comprehensive analysis performed in the current study, including regions outside the OCCR. Specifically, only 42% of BRCA2 mutations were in the OCCR region in our study, compared with 62% in the Ontario study.
BRCA VUS were detected in almost 10% of TBOCS participants, which is higher than previously reported.13, 20 This likely is attributable to the more comprehensive mutation-detection strategy we employed. The high proportion of women with VUS in this study highlights the importance of presenting this possible testing outcome during the pretest counseling session. The majority of VUS (90.5%) detected in the current study were in BRCA2. Data from Myriad Genetic Laboratories suggest that almost 40% of the BRCA2 missense variants detected in this study are likely to be of limited clinical significance, because they have been observed previously with ≥ 1 deleterious mutations, which is consistent with other reports.34 Results from a survey of cancer genetic counselors suggest that the clinical interpretation and recommendations for individuals with a VUS in BRCA1 and BRCA2 continues to be challenging.35 Improved understanding of the functional significance of missense and other types of VUS would allow more informed recommendations regarding medical management for women with these variants and their families.
Hereditary carcinoma syndromes often are characterized by a younger age of onset, multiple cancer-affected family members, and a characteristic histopathologic profile. For ovarian carcinoma, previous studies have shown that, although BRCA1-associated ovarian carcinomas occur an average of 5–10 years prior to sporadic ovarian carcinomas, BRCA2-associated ovarian carcinomas occur, on average, at the same age as sporadic ovarian cacinomars.6 Findings in the current study were similar to those reported previously and showed that the mean age of onset of ovarian carcinoma in BRCA1 carriers was approximately 5 years younger than the mean age of onset in BRCA2 carriers and in women with sporadic ovarian carcinoma.
In clinical practice, family history of cancer is used as a key factor in estimating the likelihood of a BRCA1 or BRCA2 mutation and in guiding recommendations for patient testing. However, there have been conflicting reports regarding the sensitivity of family history as a reliable predictor of mutation carrier status. Our results suggest that relying on a family history of breast or ovarian carcinoma to indicate the need for genetic counseling and testing would result in missing > 30% of BRCA-associated ovarian carcinomas. In comparison, a recent Polish study reported a positive family history of breast and/or ovarian carcinoma in only 41% of mutation carriers and in only 9% of noncarriers.14 In contrast, a Swedish population-based study of ovarian carcinomas reported the presence of a family history of breast and/or ovarian carcinoma in > 90% of mutation carriers, compared with only 24% of noncarriers.13 Thus, similar to recommendations made by Risch et al.,6 these results suggest that it may be appropriate to offer genetic counseling and testing of BRCA1 and BRCA2 to all patients with invasive, nonmucinous epithelial ovarian carcinomas, regardless of the presence or absence of a family history of breast or ovarian carcinoma.
Histologic findings in the current study are consistent with previous reports, which have suggested that BRCA-associated ovarian carcinomas are predominantly of serous histology and that mutations are seen rarely in mucinous or borderline tumors.6, 36–42 In the Ontario study,6 pathologic analysis showed that 56 of 60 mutation carriers (93%) had invasive serous tumors, and the remaining 4 women had endometrioid tumors. The Swedish population-based study showed similar results, with 62% serous carcinomas, 23% endometrioid carcinomas, 8% clear cell carcinomas, 8% serous surface papillary carcinomas, and no mucinous carcinomas.13 In the current study, the results showed that 20 of 32 mutation carriers (63%) had invasive serous carcinomas, 5 women (16%) had mixed carcinomas, and 3 women (9%) had endometrioid-type carcinomas. No mutation carriers in our study had borderline or mucinous tumors. In addition, most mutation carriers (almost 70%) had poorly differentiated tumors, consistent with previous reports.43
To our knowledge, the current study is the first to provide detailed information regarding disease stage at presentation in a population-based sample of BRCA1-associated and BRCA2-associated ovarian carcinomas.22 Our sample was representative of the stage at presentation of women with ovarian carcinoma in the general population, with 71% presenting with Stages III and IV disease. The stage at presentation of the BRCA carriers was similar to that seen in the general population.
Limitations of this study included a lower than desired participation rate (64%). Given the low survival rates for women with ovarian carcinoma, we decided to bypass the state cancer registry for patient ascertainment. Instead, we developed a rapid case-ascertainment system through gynecologic oncologists. However, we contend that the rapid ascertainment among clinics that diagnose > 85% of the cases in the catchment area generated a more representative sample than is obtainable through typical population-based studies of rapidly fatal carcinomas.6 Another limitation was the limited access to minority participants. However, our population was representative of the demographics of our catchment area (see Table 1).
Our mutation-detection strategy involved full gene sequencing and testing for common rearrangements, which is supposed to detect approximately 90% of mutations. It is possible, however, that some mutations may have been missed with this strategy, which is the most comprehensive testing currently available on a clinical basis. Thus, the hereditary fraction may be higher than what we reported.
Another possible weakness in this study is that family history information was obtained by personal interview and was not confirmed by either pathology report or other medical records. Studies suggest that the diagnosis of cancer reported in relatives is likely to be accurate, at least for first-degree relatives.24–26 However, we acknowledge that reports of certain types of malignancies, such as ovarian carcinoma, may not be as accurate.27 Self-reported family history should not affect our conclusions about the frequency of BRCA mutations in this population, because the patients were not selected on the basis of family history. Rather, because the study was population based, the reported family histories are likely representative of those of all women with ovarian carcinoma, and not of families that were selected for their high incidence of carcinoma.
Epithelial ovarian carcinoma results in the death of more American women than all other gynecologic malignancies combined. To our knowledge, the current report is the first U.S. population-based ovarian carcinoma study that comprehensively evaluates the role of the BRCA1 and BRCA2 genes in the etiology and pathology of this deadly disease. Findings from the current study suggest that 1) the incidence of hereditary ovarian carcinoma attributed to BRCA1 and BRCA2 mutations may be greater than reported previously; 2) BRCA2 mutations account for a higher percentage of hereditary ovarian carcinomas than reported previously, and previous studies may have underestimated the contribution of BRCA2 to ovarian carcinoma, especially mutations outside the OCCR; 3) approximately 30% of patients with hereditary ovarian carcinoma have no obvious family history to suggest hereditary carcinoma susceptibility, hence, it appears reasonable to offer genetic counseling to all women who have invasive, nonmucinous epithelial ovarian carcinoma.
The authors thank Kathy Eyring, C.C.R.P., Margaret (Peggy) R. Maggiacomo, B.A., Daniel Shippy, B.S., and Tricia Holtje, M.S.P.H., for data management assistance; Jennifer Pickard, M.S. for genetic counseling assistance; Elizabeth Allen, A.R.N.P., Dominique Garvin, A.R.N.P., Laura Adamson, R.N., Mary Reilly, R.N., Regina White, R.N., Peggy Rice, R.N., and Shirley Antonek for assistance with patient recruitment; and Jael Rodriguez and Justice Pitts for chart abstraction and data entry. They also thank Dr. Thomas Sellers for helpful discussions and careful review of this article. The authors especially thank all of the women who participated in this study for their time, efforts, and insight.