Breast Cancer Incidence After Risk-Reducing Salpingo-Oophorectomy in BRCA1 and BRCA2 Mutation Carriers

Women with a proven BRCA1 or BRCA2 mutation have a risk for developing breast cancer of 40% to 80% by the age of 70 (1, 2). Ovarian cancer risk by the age of 70 is 40% for BRCA1 and 18% for BRCA2 mutation carriers (1, 2) A study by van der Kolk and colleagues (3) conducted at our centre found cumulative risks of breast cancer of 71% [95% confidence interval (CI), 67%–82%) in BRCA1 and 88% (95% CI, 82%–93%) in BRCA2 mutation carriers by the age of 70. These risks were at the high end of the spectrum compared with other studies, especially for BRCA2 mutation carriers. The authors hypothesized that this could be due to skewing towards testing in women affected with cancer, competing risks of mortality from ovarian cancer in BRCA1 mutation carriers, or missing of BRCA2 mutation families with lower cancer penetrance that fail to meet the age-related criteria for mutation testing (3).

BRCA1 and BRCA2 mutation carriers are counseled on different risk-reducing strategies, either screening or prophylactic surgery. Breast cancer screening with clinical breast examination (CBE), mammography, and MRI is effective in detecting early-stage breast cancer in mutation carriers (4–8). For ovarian cancer, current screening protocols are ineffective in detecting early-stage ovarian cancer (9, 10). Risk-reducing salpingo-oophorectomy (RRSO) is effective in reducing the risk of ovarian cancer (HR, 0.21; 95% CI, 0.12–0.39) (11). After counseling, RRSO is chosen by almost all BRCA1 and BRCA2 mutation carriers at our centre.

Besides reducing ovarian cancer risk, RRSO has been shown to reduce breast cancer risk in mutation carriers. A large meta-analysis found risk reductions of 50% associated with RRSO in both BRCA1 and BRCA2 mutation carriers (BRCA1: HR, 0.47; 95% CI, 0.35–0.64; BRCA2: HR, 0.47; 95% CI, 0.26–0.84; ref. 11). The effect of RRSO on breast cancer risk was suggested to be stronger in women at premenopausal age, as greater risk reductions were found in women who had surgery before the age of 40 to 50 years (12, 13).

However, it has been suggested that the effect of RRSO differs between BRCA1 and BRCA2 mutation carriers (14, 15). In a study of Shah and colleagues (15), in 51 BRCA1 and 41 BRCA2 mutation carriers, 11 new breast cancers were found, all within BRCA1 mutation carriers. The percentage of women with RRSO was equal in women who did and did not develop breast cancer (87% vs. 82%, P = 0.754; ref. 15).

We aimed to study the incidence of breast cancer after RRSO at premenopausal age in BRCA1 and BRCA2 mutation carriers in the Northern Netherlands in a prospective cohort.

All women with 1 or 2 breasts in situ, who were 51 years old or younger (mean menopausal age in the Netherlands; ref. 16) at the time of RRSO, were consecutively selected from a prospective cohort of BRCA1/2 mutation carriers who were enrolled in the surveillance program for hereditary breast and ovarian cancer at the University Medical Center Groningen (UMCG; Groningen, The Netherlands) from September 1995 until January 2011. Women with previous breast cancer were included, when breast cancer screening continued on the remaining breast tissue. Women in whom ovarian cancer was found on the RRSO specimen were not included. Before this study, we analyzed a previous version of this database (data up until 30 September, 2009) to determine the effectiveness of breast cancer screening after RRSO (17).

RRSO is advised from the age of 35 to 40 years in BRCA1 mutation carriers and 40 to 45 years in BRCA2 mutation carriers (18). However, patients are counseled individually, and actual timing of RRSO depends on personal circumstances such as previous breast cancer, family history of cancer, previous or planned pregnancies, and mental acceptance of this definitive procedure. The operative procedure is conducted by laparoscopy (19).

Breast cancer screening in BRCA1/2 mutation carriers was done according to the Dutch guidelines, with annual complete breast examination (CBE), and mammography and MRI alternating by 6 months since 2008 (18). Before 2008, CBE was conducted biannually and MRI was conducted in women participating in the MRISC trial (MRI screening for breast cancer in women with familiar or genetic predisposition, ref. 20).

All BRCA1/2 mutation carriers who visit our Family Cancer Clinic are consecutively included in a prospective cohort. If these women met our inclusion criteria, we retrospectively retrieved relevant data from the patient files in the hospital. Physician's letters, pathology reports, and imaging reports were used for data collection.

Information on the type of mutation, date of birth, date of RRSO, and ever-use of hormonal replacement therapy after RRSO was collected. For previous breast cancers, age of diagnosis and conducted surgical procedure (breast-conserving therapy or mastectomy) were recorded. For breast cancers diagnosed after RRSO, age at diagnosis and pathologic features were recorded: tumor type according to the WHO classification, tumor size in millimeters, tumor grade (Elston–Ellis modification of Bloom–Richardson grading system), receptor status [estrogen receptor (ER), progesterone receptor (PR), and HER2 receptor], presence of ductal carcinoma in situ (DCIS), and presence of lymph node metastases. Patient data were entered into an SPSS database. Protection of the patient's identity was guaranteed by a patient-specific number. Those numbers were only retraceable to individual women when entered in the password-protected hospital database, which is only accessible for hospital employees who have a personal account. Complying with Dutch law, no further Institutional Review Board approval was needed.

To present the study population and characteristics of breast cancers discovered after RRSO, contingency tables were used. Breast cancers detected after RRSO could be prevalent, possibly prevalent or incident. Prevalent breast cancers were defined as breast cancers detected within 6 months after RRSO. Possibly prevalent breast cancers were defined as breast cancers detected more than 6 months after RRSO that we considered prevalent because the time after RRSO was shorter than the estimated growing time of the breast cancer. For all the invasive breast cancers diagnosed more than 6 months after RRSO, we estimated the duration of time a breast cancer had been growing and compared this with the time interval between RRSO and detection of breast cancer. To estimate this time, we used the formula for estimating tumor-doubling time in mutation carriers based on age at tumor diagnosis as presented by Tilanus-Linthorst and colleagues (21): log₂ [doubling time (years)] = −7.75 + 0.12 age. On the basis of the tumor-doubling time, we estimated the time a tumor was growing from the theoretical size at baseline of 5 mm (which is the assumed detection threshold for invasive breast cancer on MRI; refs. 22, 23) to the size at detection. If this estimated time period was longer than the time period between RRSO and tumor detection, we considered a breast cancer possibly prevalent. Incident breast cancers were defined as breast cancers detected more than 6 months after RRSO and not possibly prevalent.

Duration of follow-up was calculated for each woman from RRSO to diagnosis of incident breast cancer, to prophylactic mastectomy, to one year after the last screening visit, or to January 1, 2011. The analyses were conducted using the SPAW software package, version 18.0 for Windows (SPSS).

To estimate the number of women who develop incident breast cancer after RRSO based on womens' ages at RRSO and their ages during follow-up, we used the penetrance curves for breast cancer in mutation carriers in our centre as published by Van der Kolk and colleagues (3) and a 50% risk reduction for breast cancer expected after RRSO [BRCA1/2: HR, 0.49; (95% CI, 0.37–0.65); BRCA1: HR, 0.47; (95% CI, 0.35–0.64) and BRCA2: HR, 0.47 (95% CI 0.26–0.84); ref. 11].

We analyzed the data of 162 women, 104 BRCA1 mutation carriers and 58 BRCA2 mutation carriers (Table 1). At the time of RRSO, median age was 41 years (range 30–51) for BRCA1 mutation carriers and 42 years (range 33–51) for BRCA2 mutation carriers (P = 0.013). At the time of RRSO, 25% of the BRCA1 mutation carriers and 12% of the BRCA2 mutation carriers had a history of breast cancer (P = 0.066). After RRSO, hormonal replacement therapy was prescribed to 47% (68/146) of the women, all without previous breast cancer. Total follow-up in this study was 6,389 months (532 women-years); 4,309 months (359 women-years), and 2,080 months (173 women-years) for BRCA1 and BRCA2 mutation carriers, respectively.

During the post-RRSO screening period, 18 breast cancers in 18 women were detected (34/1,000 women-years). Of these, 16 were found in BRCA1 mutation carriers (45/1,000 women-years) and 2 in BRCA2 mutation carriers (12/1,000 women-years).

Of the 18 breast cancers diagnosed after RRSO, 3 were found within 6 months after RRSO and were considered prevalent breast cancers. On the basis of tumor-doubling time calculations, we considered 2 breast cancers to be possibly prevalent (Table 2). Leaving out 3 prevalent breast cancers and 2 possibly prevalent breast cancers, there were at least 13 incident breast cancers (24/1,000 women-years) during the study period, 12 in BRCA1 mutation carriers (33/1,000 women-years) and 1 in a BRCA2 mutation carrier (6/1,000 women-years). Table 3 illustrates the number of women expected to develop breast cancer compared with the number observed. Table 4 illustrates the characteristics of women who did and did not develop incident breast cancer after RRSO.

Of the 13 incident breast cancers detected after RRSO, 11 (86%) were invasive and 2 (15%) were DCIS (Table 5). In 4 of 11 (36%) invasive cancers, axillary lymph nodes were positive. Histologic grade was higher in breast cancers from BRCA1 mutation carriers than from BRCA2 mutation carriers (70% grade 3 vs. none, respectively). Of the 11 invasive breast cancers, 9 (82%) were ER, PR, and HER2 negative (triple negative), all in BRCA1 mutation carriers.

In a consecutive group of 104 BRCA1 and 58 BRCA2 mutation carriers with RRSO at premenopausal age, breast cancer screening at our Family Cancer Clinic revealed 18 breast cancers in 18 women during 532 women-years (34/1,000 women-years), of which, 13 were incident breast cancers (24/1,000 women-years). On the basis of breast cancer incidence curves and a 50% risk reduction associated with RRSO, we expected to find 8 (range 6–10) incident breast cancers (15/1,000 women-years; refs. 3, 11). Several factors may have contributed to this difference in expected and observed breast cancer incidence after RRSO.

An important finding is that 12 incident breast cancers developed in BRCA1 mutation carriers compared with 5 (range 3–6) expected. For BRCA2 mutation carriers, 1 incident breast cancer developed compared with 3 (range 2–5) breast cancers expected. The risk reduction associated with RRSO might be smaller in BRCA1 than in BRCA2 mutation carriers. The same hypothesis was suggested by Shah and colleagues (15); they found 9 new tumors after RRSO in 45 BRCA1 mutation carriers and none in 35 BRCA2 mutation carriers. Kauff and colleagues (24) also showed a significant breast cancer risk reduction for BRCA2 mutation carriers but not for BRCA1 mutation carriers [BRCA1: HR, 0.61 (95% CI, 0.30–122), BRCA2: HR 0.28 (95% CI, 0.08–0.92)] in a large prospective study with 190 BRCA1 and 113 BRCA2 mutation carriers at risk for breast cancer after RRSO.

One explanation for this difference is that the effect of RRSO might be less marked in BRCA1 mutation carriers, because their breast cancers are often ER and PR negative (14, 24). Of the invasive breast cancers found in BRCA1 mutation carriers in this study, 90% were ER and PR negative. Nevertheless, a protective effect of RRSO in BRCA1 mutation carriers has been shown in several studies (12, 13).

Others suggested that RRSO might inhibit breast cancer growth in BRCA1 mutation carriers at tumorigenesis: growth of ER and PR negative cells might be indirectly induced by paracrine signals from ER- and PR-positive cells that are influenced by estrogen and progesterone (15, 25). Thus, breast cancer risk reduction after RRSO may take longer to establish than the duration of follow-up in this study, which may be a second explanation why we saw more breast cancers than expected.

Another explanation for finding more breast cancers than expected may be the intensification of the breast cancer screening regime since 2008, when MRI screening for all BRCA1/2 mutation carriers was introduced at our centre. Van der Kolks study contains information on breast cancer incidence up to March 2008. Before 2008, MRI screening was used in a small selection of women participating in the MRISC study. It is known that after introducing a new effective screening regimen, more breast tumors are found. Warner and colleagues found a higher incidence of breast cancer in the first 3 years after the introduction of MRI screening in BRCA1/2 mutation carriers and an overall higher incidence of DCIS (26). This effect has also been shown in the general population after introduction of the population-based breast cancer screening with mammography (27, 28). This explanation is weakened by the fact that this effect was not seen in BRCA2 mutation carriers.

A fourth explanation might be that RRSO is chosen by women with especially high breast cancer risks. Although eventually RRSO is chosen by almost all mutation carriers at our centre, timing might be affected by family history and previous breast cancers. As can be seen in Table 4, of the women who developed breast cancer after RRSO, 54% were younger than 40 years and 31% had previous breast cancer, compared with 41% and 23% in the women who did not develop breast cancer.

A last explanation might be the use of hormonal replacement therapy after RRSO. In our study population, 47% of the women used hormonal replacement therapy after RRSO and this percentage was not higher in women with incident breast cancer. Although it has been shown that short-term use of hormonal replacement therapy does not negate the effect of RRSO (29) and one study found an inversed relation in hormonal replacement therapy use and breast cancer incidence in BRCA1 mutation carriers (30), hormonal replacement therapy does increase the risk of new and recurrent breast cancer in women with previous breast cancer (31, 32). On the contrary, a higher mortality was seen in women with bilateral oophorectomy before the age of 45 who did not use hormonal replacement therapy, compared with those women who did (33). Theoretically, it is plausible that hormonal replacement therapy use after RRSO partially negates the risk-reducing effect of RRSO, but bias might be introduced if hormonal replacement therapy is more often prescribed to women with less breast cancer in their family, or if survival is shorter in women who do not use hormonal replacement therapy. Our study sample is too small and follow-up to short to draw conclusions on this issue. The effects of hormonal replacement therapy after RRSO on the long-term breast cancer risk and survival should be monitored carefully.

We aimed to study the incidence of breast cancer after RRSO at premenopausal age. The strength of this study is that we used an algorithm that incorporates an estimate of growing time to exclude the effect of prevalent breast cancer on the incidence of new breast cancers after RRSO. Small sample size and short follow-up limit the possibilities to identify risk factors for breast cancer after RRSO. Because menopausal status at RRSO was not known for all women, we chose to include all women with the age at RRSO of 51 years or younger, which can be a limitation. Although this is the mean age of menopause in the Netherlands, some of these women might have been postmenopausal at time of RRSO, either due to natural or chemotherapy-induced menopause. Furthermore, we estimated the amount of breast cancer after RRSO with breast cancer penetrance curves from our own centre. An ideal design would be a randomized controlled trial with a RRSO and a surveillance group. As surveillance is not effective and RRSO is chosen by almost all mutation carriers at our centre, this design would be unethical.

To conclude, the breast cancer incidence after premenopausal RRSO is still high, especially in BRCA1 mutation carriers. We could not confirm the expected risk reduction as described by other authors. As a consequence, after RRSO, continued surveillance with mammography and MRI for breast cancer in BRCA1 and BRCA2 mutation carriers is warranted.

No potential conflicts of interest were disclosed.

Conception and design: I.E. Fakkert, M.J.E. Mourits, J.C. Oosterwijk, G.H. de Bock

Development of methodology: I.E. Fakkert, M.J.E. Mourits, G.H. de Bock

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): I.E. Fakkert, M.J.E. Mourits, L. Jansen, K. Meijer, J.C. Oosterwijk, B. van der Vegt

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): I.E. Fakkert, L. Jansen, M.J.W. Greuter, G.H. de Bock

Writing, review, and/or revision of the manuscript: I.E. Fakkert, M.J.E. Mourits, L. Jansen, D.M. van der Kolk, K. Meijer, J.C. Oosterwijk, B. van der Vegt, M.J.W. Greuter

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): I.E. Fakkert, M.J.E. Mourits, G.H. de Bock

Study supervision: M.J.E. Mourits, M.J.W. Greuter, G.H. de Bock

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.