Variation in Mutation Spectrum Partly Explains Regional Differences in the Breast Cancer Risk of Female BRCA Mutation Carriers in the Netherlands Free

Breast cancer is the most prevalent type of cancer in women. In the Netherlands, women in the general population have a lifetime risk of approximately 11% of developing breast cancer and 1% of ovarian cancer (1), whereas carriers of a pathogenic BRCA1 or BRCA2 mutation have a much higher risk of developing breast and/or ovarian cancer. Two international meta-analyses have estimated a mean breast cancer risk by age 70 of 57% and 65% for BRCA1, and 45% and 49% for BRCA2 mutation carriers (BRCA1/2 carriers; refs. 2, 3). Because referred hereditary cases may have an even higher lifetime risk (4), counseling of BRCA1/2 carriers in the Netherlands is still based on an estimated lifetime breast cancer risk of 60% to 80% (5).

In a previous study in the Northern Netherlands, relatively high risks were observed in BRCA carriers, especially in BRCA2 carriers older than 60 (6). The cumulative incidence for breast cancer by age 70 with inclusion of the index cases was 71% [95% confidence interval (CI), 67%–76%] in BRCA1 carriers and 88% (95% CI, 82%–93%) in BRCA2 carriers. After excluding the index cases, these risks were 60% (95% CI, 55%–66%) and 78% (95% CI, 69%–88%), respectively. These figures, both with and without index cases, are at the high end of published risk estimates in several studies (2, 3, 6–10). Especially, the risk in BRCA2 carriers observed in the Northern Netherlands is high, i.e., higher than the Dutch risks used in counseling (5).

Potential explanations for regional or national differences in breast cancer incidence rates are differences in genetic and nongenetic population factors, such as founder mutations, genetic modifiers, demographic composition, and lifestyle factors, as well as differences in ascertainment and methodology (10–16). The contribution of (founder) mutations in the BRCA genes in the ovarian cancer cluster region (OCCR) could, for instance, affect regional risk differences, because mutations within the OCCR are associated with a lower breast cancer risk (17, 18). Regional differences in the population breast cancer incidence, i.e., difference in the baseline risk, could also affect risk differences among BRCA1/2 carriers. We aimed to determine whether there are regional differences in the breast cancer risk in BRCA1/2 carriers in the Netherlands, and if so, whether they could be explained by mutation spectrum or population background risk.

The Family Cancer Clinic of the University Medical Center Groningen (UMCG; Groningen, the Netherlands) provides multidisciplinary counseling to residents of the Northern Netherlands (19). This is a stable population (20), and, therefore, feasible to assess the role of mutation spectrum in the existence of regional risk differences (21). The population of the study cohort included all (N = 1,050) women with a pathogenic BRCA mutation (either by testing or obligate carriers, i.e., a women with a child as well as a parent or sibling carrying a mutation) known at the UMCG by September 2011. Ascertainment and the DNA testing procedure were described previously (6). Women were included if they were born after 1910, and from a minimum age of 18 years. The index cases were women with: (i) the first detected BRCA1/2 mutation in their family, or (ii) the youngest age at diagnosis of breast or ovarian cancer when the first BRCA1/2 mutation in the family was detected in multiple females at the same time. No index case was defined in the family, when the first mutation was detected in a male or an unaffected woman.

In the previous study, data were collected on the occurrence of breast and ovarian cancer in BRCA1/2 carriers in the Northern Netherlands and their first-degree relatives up to March 2008 (6). For this study, the dataset was updated by retrieving the patient files, and extended to September 2011. The update included data of: known carriers (N = 233), new carriers (N = 122) in known families and carriers (N = 442) in newly ascertained families.

Information on women seen at the clinic was collected from their clinical genetic records, which also contained extensive and regularly updated pedigree information. Information on obligate carriers was obtained, with consent, from medical records. For all the women included, we recorded: date of birth, date of last contact/death, and data on the familial gene mutation (according to HGVS nomenclature) and individual carrier status, date of breast cancer diagnosis, date of ovarian cancer diagnosis, date of risk-reducing mastectomy (RRM) and, date of risk-reducing salpingo-oophorectomy (RRSO). These data were retrieved from the medical records of patients and entered into a separate, anonymous, password-protected database. According to the Dutch law, this means no further approval from the Institutional Review Board was needed.

The carrier reference population consisted of all BRCA1/2 carriers from the rest of the Netherlands as registered in the Hereditary Breast and Ovarian Cancer Research Group Netherlands (HEBON) study (N = 2,013), a national cohort of BRCA1/2 families up to 2006 (10, 22). The general reference populations were the general population from the Northern Netherlands and the general population in the rest of the Netherlands. Data on the population background risk, i.e., age-related breast cancer incidence, in both regions were collected from the Dutch Comprehensive Cancer Centre (CCN/IKNL) for the period 2005 to 2010 (1).

In the analyses, obligate carriers and proven carriers were considered as one group. The mutation spectrum was assessed by stratifying the mutations for both genes in three previously defined groups: 5′ to the OCCR, in the OCCR, and 3′ to the OCCR (10, 17, 18). The BRCA1 OCCR was defined as nucleotides 2,282 to 4,071, and the BRCA2 OCCR as nucleotides 2,831 to 6,401 (10, 17, 18). The regions downstream and upstream of the OCCR were defined as the first nucleotide at the prime side up to the nucleotide before the start of the OCCR. Large rearrangements were classified in the according category based on their location.

Descriptive statistics were used to describe the study population and the carrier reference population. Only the first primary breast cancer was considered as an event and censoring was applied at the age of RRM (N = 177), RRSO (N = 342), ovarian cancer diagnosis (N = 298), last moment of contact (N = 914), or death (N = 33). To estimate the cumulative incidence of breast cancer in BRCA1 and BRCA2 carriers in both regions, crude (unadjusted) age-related breast cancer rates were calculated using Kaplan–Meier survival analyses, and the corresponding confidence intervals were calculated with a log–log approach based on Greenwood SEs (23). To adjust for a possible birth cohort effect and mutation spectrum, normalized weights (one divided by the number of combination of the categories of the confounding variables) were calculated and applied in the Kaplan–Meier analyses. Through this weighing, the percentage of women in the different birth cohort categories (≤1920, 1921–1930, 1931–1940, 1941–1950, 1951–1960, 1961–1970 and ≥1971) and the mutation spectrum (5′ to OCCR, OCCR, and 3′ to OCCR) were identical in each category and also in the study and the carrier reference population, without affecting the total sample size.

Regional differences in the age-related breast cancer incidences in BRCA1/2 carriers in the Northern Netherlands compared with the rest of the Netherlands were analyzed using Cox regression analyses. All HRs were adjusted for birth year and mutation spectrum. To correct for ascertainment bias, these analyses were performed with and without the index cases. The assumption of proportionality was checked by assessing the log-minus-log plots and Schoenfeld residuals. The amount of confounding by the mutation spectrum was calculated as the change in the HR (24):

Differences in the background breast cancer risk in the general population in the UMCG catchment area compared with the rest of the Netherlands were tested using ORs. The cumulative incidence in the general population was calculated using the cumulative rates, which were calculated as the sum of the age-specific breast cancer incidence rates in the general population, multiplied by the width of the age group (i.e., 5 years; ref. 25).

All analyses were performed using SPSS version 20 (SPSS Inc.) and R version 3.0.1 (26), and statistical significance was defined as P < 0.05.

The study population consisted of 1,050 BRCA carriers: 656 BRCA1 and 394 BRCA2 (Table 1, data excluding index cases; Supplementary Table S1, data including index cases). These included 29% BRCA1 and 32% BRCA2 index cases. The carrier reference population from the rest of the Netherlands consisted of 1,580 BRCA1 carriers, including 31% index cases, and 433 BRCA2 carriers, including 30% index cases. The study population consisted of relatively more BRCA2 carriers, but the family size of BRCA families was not different between the two populations. The average birth year in the study population was earlier in BRCA1 carriers (1953 vs. 1955, P = 0.025) and BRCA2 carriers (1952 vs. 1957, P < 0.001) compared with the carrier reference population.

In the study cohort, including index cases, 46% BRCA1 and 44% BRCA2 carriers developed breast cancer. In the carrier reference population 45% BRCA1 and 42% BRCA2 carriers developed breast cancer (Supplementary Table S1). BRCA1 and BRCA2 carriers in the study population were diagnosed with breast cancer at an older mean age than carriers in the reference population, for BRCA1 42.4 (SD 10.1) versus 40.2 (SD 8.5) years (P = 0.001) and for BRCA2 47.1 (SD 10.6) versus 43.4 (SD 8.6) years (P < 0.001).

The mutation spectrum was known for all BRCA1 and BRCA2 carriers in the Northern Netherlands, and for 88% of BRCA1 and 83% of BRCA2 carriers in the rest of the Netherlands. In the Northern Netherlands, 39% BRCA1 carriers had a mutation within the OCCR versus 23% carriers in the reference population (P < 0.001; Fig. 1). For the BRCA2 mutation in the Northern Netherlands, 18% mutations were within the OCCR versus 65% in the rest of the Netherlands (P < 0.001). This large difference in the mutation spectrum between BRCA2 carriers in both populations was due mainly to the presence of three apparent northern founder mutations (c.9672dup, c.7419_7420delTG, and c.1310_1313del: all three outside the OCCR).

For BRCA1 carriers, the crude cumulative incidence of breast cancer by age 70 was 58% (95% CI, 50%–66%) in the Northern Netherlands, and 68% (95% CI, 62%–73%) in de rest of the Netherlands (Table 2). At 70 years of age, 24 (5%) and 28 (3%) women were still at risk of developing cancer in the BRCA1 study and reference population, respectively (Supplementary Table S2). When the index cases were included in both populations, these risks were 72% (95% CI, 66%–78%) and 76% (95% CI, 71%–79%), respectively. Weighted breast cancer risks that excluded the index cases and incorporated the differences in birth cohorts and mutation spectrum between both populations, resulted in lower risk estimates: 53% (95% CI, 45%–61%) in the Northern Netherlands and 55% (95% CI, 47%–63%) in the rest of the Netherlands (Table 2, Fig. 2; Supplementary Table S3).

In BRCA2 carriers, the breast cancer risk at age 70 was 64% (95% CI, 50%–75%) in the Northern Netherlands and 61% (95% CI, 50%–69%) in the rest of the Netherlands. At 70 years of age, 11 (4%) and 14 (5%) women were still at risk of developing cancer, respectively. When the index cases were included, these risks were 78% (95% CI, 69%–85%) and 72% (95% CI, 64%–78%), respectively. For BRCA2 carriers, the weighted breast cancer risk, excluding the index cases, resulted in higher estimates than the crude risks: by age 70 these were 66% (95% CI, 54%–78%) in the Northern Netherlands and 64% (95% CI, 51%–77%) in the rest of the Netherlands (Table 2).

Weighted cumulative incidence curves stratified by mutation spectrum and the related relative risks are presented in Fig. 3 and in Supplementary Table S4, respectively. The different categories of the mutation spectrum were related to different breast cancer risks, with lower breast cancer risk estimates for mutations within the OCCR. These differences were more outspoken in BRCA2 carriers, and opposite effects (increased vs. decreased risks) were shown in the study population and reference population. No significant differences were revealed in regional comparisons of each specific mutation, but numbers were small (data not shown).

BRCA1 carriers (excluding index cases) in the Northern Netherlands had a lower risk of developing breast cancer than those in the rest of the Netherlands, HR, 0.66 (95% CI, 0.54–0.81; Table 3, data excluding index cases; Supplementary Table S5, data including index cases). When only data from women with a known mutation spectrum were included, this HR was 0.73 (95% CI, 0.59–0.89). The amount of confounding by the mutation spectrum was 3%. Correction for the mutation spectrum increased the HR with this amount, and resulted in an HR of 0.75 (95% CI, 0.61–0.93).

For the incidence curves of the BRCA2 populations, the assumption of proportionality in the Cox model was not met; the lines crossed at approximately age 60 (Fig. 2). Below age 60, the breast cancer risk of BRCA2 carriers in the Northern Netherlands was lower than that in the rest of the Netherlands, both excluding index cases (HR, 0.64; 95% CI, 0.51–0.81) and including index cases (HR, 0.62; 95% CI, 0.49–0.78). Strikingly, correcting for the mutation spectrum reduced the HR by 32%, which resulted in a 50% lower breast cancer risk in the Northern Netherlands (HR, 0.50; 95% CI, 0.33–0.77). From age 60 onward, however, the breast cancer risk of BRCA2 carriers in the Northern Netherlands was at least twice as high: the HR was 3.99 (95% CI, 1.11–14.35), excluding the index cases, and 2.56 (95% CI, 1.91–5.48), including the index cases. When we looked at only carriers with known mutation spectrum data, the HRs decreased and became nonsignificant. Correcting for the mutation spectrum reduced the HR by 11% to a HR of 2.61 (95% CI, 0.68–10.0).

There seemed to be no significant differences in the breast cancer incidence in the general population between the Northern Netherlands and the rest of the Netherlands: the OR over the ages 20 to 75 was 1.03 (95% CI, 0.97–1.09), younger than 60, 1.02 (95% CI, 0.94–1.11); and older than 60, 1.03 (95% CI, 0.97–1.09). Neither were there birth cohort–related differences in the regional OR, and the cumulative incidence curves were parallel over time (Fig. 2).

We aimed to determine whether there were regional differences in the breast cancer risk in BRCA1/2 carriers ascertained and analyzed in one identical way, and if so, whether they could be explained by the mutation spectrum or population background risk. We found that all BRCA1 and the BRCA2 carriers younger than 60 had a significantly lower breast cancer risk in the Northern Netherlands, HRs excluding index cases were 0.66 (95% CI, 0.54–0.81) and 0.64 (95% CI, 0.51–0.81), respectively. From age 60, the breast cancer risk in BRCA2 carriers in the Northern Netherlands was higher than in the rest of the Netherlands (HR excluding index cases, 3.99; 95% CI, 1.11–14.35). Adjustment for the mutation spectrum changed these HRs for BRCA1, BRCA2 < 60, and BRCA2 ≥ 60 by −3%, 32%, and 11%, respectively. The background breast cancer incidence did not explain any risk difference: OR younger than 60 was 1.02 (95% CI, 0.94–1.11) and OR older than 60 was 1.03 (95% CI, 0.97–1.09).

This study confirmed regional breast cancer risk differences, especially for BRCA2 carriers. We estimated a cumulative incidence of breast cancer by age 70, excluding indexes, of 58% (95% CI, 50%–66%) for BRCA1 carriers and 64% (95% CI, 50%–75%) for BRCA2 carriers in the Northern Netherlands. For BRCA1 carriers this was similar to the previous estimate, but the BRCA2 risk decreased from 78% to 64% (6). The lower risk in this BRCA2 cohort can be explained by the fact that now more carriers (+9%) have a mutation within the OCCR and more carriers (+8%) are born in the ≥1971 cohort as compared with the previous study. Besides, to enable comparison with the national HEBON cohort data, we applied censoring for all RRSOs, whereas in the previous study, censoring was only applied when RRSO was performed before 50 years of age.

Published estimations of the breast cancer life time risk of BRCA1/2 carriers show considerable variation: 40% to 87% for BRCA1 and 28% to 88% for BRCA2 carriers (2, 3, 7–9, 27). This variation can be explained by differences in population, in ascertainment and in methodology. Our current estimates of lifetime risks for the rest of the Netherlands are higher than those recently published in the same cohort (10): 68% (95% CI, 62%–73%) versus 45% (95% CI, 36%–52%) for BRCA1, and 61% (95% CI, 50%–69%) versus 27% (95% CI, 14%–38%) for BRCA2 carriers. These low estimates of Brohet and colleagues (10) can be explained by inclusion of older birth cohorts (with lower breast cancer penetrance), and by the applied method (modified segregation analysis has a stronger correction for bias). When they included only cohorts from 1940 onward, their estimates substantially increased: from 45% to 66% for BRCA1 and 27% to 32% for BRCA2 by age 65. Though based on mainly retrospective data, our breast cancer risk estimates of the carriers in the Netherlands are very similar to those published in a recent prospective study: 60% (95% CI, 44%–75%) for BRCA1 and 55% (95% CI, 41%–70%) for BRCA2 carriers by age 70 (27). However, in this prospective study the peak incidence among BRCA2 carriers was in the age group 40 to 49 years, which is earlier than in the Netherlands.

We found that differences in the mutation spectrum between the Northern Netherlands and the rest of the Netherlands partly explained the regional breast cancer risk differences, which was seen in the adjusted HRs and illustrated with weighted risk estimates. The magnitude of confounding by the mutation spectrum was largest (32%) among BRCA2 carriers below 60, as compared with 3% and 11% in BRCA1 and BRCA2 carriers older than 60, respectively. Our analyses showed that BRCA1 carriers and BRCA2 carriers younger than 60 that live in the Northern Netherlands have a lower breast cancer risk than those in the rest of the Netherlands: the HRs adjusted for birth year and OCCR were 0.75 and 0.50, respectively. Controversially, above age 60, BRCA2 carriers in the Northern Netherlands had a higher, though statistically nonsignificant, breast cancer risk (adjusted HR, 2.61; 95% CI, 0.68–10.00). When the crude and adjusted BRCA2 ≥60 model only included carriers with a known mutation spectrum, the risk difference was insignificant, which was probably due to a lack of power. The differences in the breast cancer risks between each gene region, and the different distribution of mutations between the Northern Netherlands and the rest of the Netherlands, underline the importance of the confounding effect of mutation spectrum. Strikingly, in BRCA2 the HRs of the 3′ versus the 5′ region were significantly different in the opposite direction in both the study and the reference population. This might indicate that smaller defined regions on the BRCA2 gene could explain breast cancer risk differences better, making the (previously found) regional differences in the frequency of specific BRCA1 or BRCA2 mutations in the Netherlands even more important (12).

We found that the breast cancer incidence of women in the general population did not differ between the Northern Netherlands and the rest of the Netherlands (OR, 1.03; 95% CI, 0.97–1.09), and this cannot, therefore, explain the regional risk differences among BRCA1/2 carriers. In a recent report from the Dutch Institute for Public Health and the Environment, it was shown that in one of the three northern provinces of the Netherlands, there was a small, but statistically significant, increased breast cancer risk of 1.06 compared with the national average risk, whereas for the other two provinces the 1.02 and 1.03 risk increase was nonsignificant (28).

The strengths of our study are the inclusion of the relevant target population of BRCA1/2 carriers (i.e., carriers who receive the actual counseling), the large sample size, and the partly prospectively collected data. Our study gives a new perspective on the existence and explanation of regional differences in the breast cancer risk of BRCA1/2 carriers, using a two-way approach involving genes (mutation spectrum) and environment (population background risk). Moreover, the study and the carrier reference population are family clinic–based populations, using the same national referral and DNA testing criteria and methodology, with well-documented information on carrier status and cancer development. Differences in the ascertainment between the study and reference carrier population could affect the results. However, comparison of indicators of the ascertainment of both populations showed that 76% of the tests were presymptomatic, and showed no statistically significant differences in proportion of (pre)symptomatic genetic tests, mean age at genetic test, proportion of test above age 60 and proportion of these tests per family. The population in the Northern Netherlands was ascertained up to September 2011 and the Rest of the Netherlands up to 2006, but there was also no indication for testing bias when only carriers ascertained up to 2006 were assessed. This more recent period of ascertainment is a likely explanation for the higher uptake of preventive surgery in the northern region. Up to 2006, women were offered the choice between ovarian screening and RRSO. From 2007 onward, counseling changed and ovarian screening was gradually abandoned because of proven ineffectiveness. Our center was one of the first centers to change this policy. As a result, women more often choose for RRSO and RRM, and also at an earlier age (29). The RRSOs were truly prophylactic because before RRSO all women underwent transvaginal ultrasonography and serum marker CA125 to exclude manifest ovarian cancer (30). Underreporting of preventive surgery in the rest of the Netherlands was very unlikely as medical files of proven carriers were available.

A limitation of our study is that the BRCA1/2 carriers were ascertained through family cancer clinics, which may have resulted in an overestimation of the breast cancer risks. Moreover, the study population included both incident and prevalent breast cancer cases, which also could have led to overestimation. However, the overestimation is negligible, because the follow-up time of known carriers and former untested relatives (N = 273) was extended with 61 months (SD 40), and included 19 new primary breast cancers, which is only 6% of all cases.

Besides, the percentage of ovarian cancer and preventive surgery was higher in the study population as compared with the reference population. This may have led to a higher percentage of censored cases in the study population. Therefore, the breast cancer risk could be underestimated in the study population as compared with the carrier reference population, because after these censoring events some carriers will develop breast cancer (31). Because the medical files were available, underestimation of the risk due to missing data on breast cancer, ovarian cancer, or preventive surgery is unlikely. For the rest of the Netherlands, missing data for the age at cancer diagnosis in the pedigree were imputed on the basis of birth cohort–specific population data (10). However, it is unknown whether this was necessary for the carriers included in this study (born after 1910). Future linking to national pathology and cancer registries will help to identify and quantify missing data.

In conclusion, BRCA1/2 carriers in the Northern Netherlands were at lower risk of developing breast cancer than in the rest of the Netherlands, except for BRCA2 carriers older than 60, whose risk was doubled. Differences in the mutation spectrum between the two regions partly explain the risk differences seen in BRCA1/2 carriers, with a somewhat stronger effect in BRCA2 carriers. Both for BRCA1 and BRCA2 carriers, the detected regional risk differences could not be explained by differences in the background cancer risk in their respective general populations. Possible other explanations for the regional risk differences could be differences in (founder) mutations in modifier genes, lifestyle and/or gene–environment interactions. Especially, a large scale international study on the breast cancer risk and mutation position in BRCA carriers can be worthwhile due to the limited sample sizes in this study. The increased risk in BRCA2 carriers older than 60 in the Northern Netherlands should be a good reason to consider extending intensive breast cancer screening to BRCA2 carriers beyond the age of 60 in this region.

No potential conflicts of interest were disclosed.

Conception and design: J.R. Vos, D.M. van der Kolk, M.J.E. Mourits, J.C. Oosterwijk, G.H. de Bock

Development of methodology: J.R. Vos, D.M. van der Kolk, M.J.E. Mourits, F.E. van Leeuwen, J.C. Oosterwijk, G.H. de Bock

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.R. Vos, N. Teixeira, D.M. van der Kolk, M.A. Rookus, M. Collée, C.J. van Asperen, M.G.E.M. Ausems, T.A.M. van Os, E.B. Gómez-Garcia, R.M. Brohet, A.H. van der Hout, J.C. Oosterwijk, H.F. Vasen

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.R. Vos, M.J.E. Mourits, F.E. van Leeuwen, A.H. van der Hout, L. Jansen, J.C. Oosterwijk, G.H. de Bock

Writing, review, and/or revision of the manuscript: J.R. Vos, D.M. van der Kolk, M.J.E. Mourits, M.A. Rookus, F.E. van Leeuwen, M. Collée, C.J. van Asperen, A.R. Mensenkamp, M.G.E.M. Ausems, T.A.M. van Os, H.E.J. Meijers-Heijboer, E.B. Gómez-Garcia, H.F. Vasen, L. Jansen, J.C. Oosterwijk, G.H. de Bock

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.R. Vos, N. Teixeira, C.J. van Asperen, A.R. Mensenkamp, R.M. Brohet, A.H. van der Hout

Study supervision: M.J.E. Mourits, R.M. Brohet, J.C. Oosterwijk, G.H. de Bock

The Hereditary Breast and Ovarian Cancer Research Group Netherlands (HEBON) consists of the following Collaborating Centers: Coordinating center: Netherlands Cancer Institute, Amsterdam, NL: M.A. Rookus, F.B.L. Hogervorst, F.E. van Leeuwen, S. Verhoef, M.K. Schmidt, J.L. de Lange, R. Wijnands; Erasmus Medical Center, Rotterdam, NL: M. Collée, A.M.W. van den Ouweland, M.J. Hooning, C. Seynaeve, C.H.M. van Deurzen, I.M. Obdeijn; Leiden University Medical Center, NL: C.J. van Asperen, J.T. Wijnen, R.A.E.M. Tollenaar, P. Devilee, T.C.T.E.F. van Cronenburg; Radboud University Nijmegen Medical Center, NL: C.M. Kets, A.R. Mensenkamp; University Medical Center Utrecht, NL: M.G.E.M. Ausems, R.B. van der Luijt; Academic Medical Center, NL: C.M. Aalfs, T.A.M. van Os; VU University Medical Center, Amsterdam, NL: J.J.P. Gille, Q. Waisfisz, H.E.J. Meijers-Heijboer; University Hospital Maastricht, NL: E.B. Gómez-Garcia, M.J. Blok; University Medical Center Groningen, NL: J.C. Oosterwijk, A.H. van der Hout, M.J. Mourits, G.H. de Bock; The Netherlands Foundation for the Detection of Hereditary Tumours, Leiden, NL: H.F. Vasen; The Netherlands Cancer Registry: S. Siesling; The Dutch Pathology Registry (PALGA): L.I.H. Overbeek.

HEBON thanks the registration teams of the Comprehensive Cancer Centre Netherlands and Comprehensive Centre South (together the Netherlands Cancer Registry) and PALGA (Dutch Pathology Registry) for part of the data collection.

The authors thank Jackie Senior for editing the final version of the article.

The HEBON study is supported by the Dutch Cancer Society grants NKI1998-1854, NKI2004-3088, and NKI2007-3756, the Netherlands Organization of Scientific Research grant NWO 91109024, the Pink Ribbon grant 110005, and the BBMRI grant NWO 184.021.007/CP46 (to M.A. Rookus and F.E. van Leeuwen).

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.