You are here

Article Text

Article menu
Cancer genetics
Original article
Ovarian cancer familial relative risks by tumour subtypes and by known ovarian cancer genetic susceptibility variants
  1. Sarah Jervis1,
  2. Honglin Song2,
  3. Andrew Lee1,
  4. Ed Dicks2,
  5. Jonathan Tyrer2,
  6. Patricia Harrington2,
  7. Douglas F Easton1,2,
  8. Ian J Jacobs3,
  9. Paul P D Pharoah1,2,
  10. Antonis C Antoniou1
  1. 1Department of Public and Primary Care, Centre for Cancer Genetic Epidemiology, University of Cambridge, Cambridge, UK
  2. 2Department of Oncology, Centre for Cancer Genetic Epidemiology, University of Cambridge, Cambridge, UK
  3. 3Faculty of Medical & Human Sciences, Institute of Human Development, The University of Manchester and Manchester Academic Health Science Centre, Manchester, UK
  1. Correspondence to
    Dr Antonis C Antoniou, Strangeways Research Laboratory, Department of Public Health and Primary Care, Worts Causeway, Cambridge CB1 8RN, UK; antonis{at}srl.cam.ac.uk

Abstract

Background Family history is one of the most important risk factors for epithelial ovarian cancer (EOC). Little is known, however, on how EOC familial relative risks (FRRs) vary by factors such as tumour subtype or the combined effects of common EOC susceptibility alleles. In addition, no data currently exist on the FRRs associated with EOC after exclusion of BRCA1 or BRCA2 mutation carriers.

Methods EOC FRRs were computed from observed EOCs in relatives of 1548 patients with EOC recruited between 1999 and 2010 from a population-based cohort study with known BRCA1 and BRCA2 mutation status and tumour subtype, compared with the number expected in the general population.

Results The EOC FRR to all first-degree relatives was estimated to be 2.96 (95% CI 2.35 to 3.72) but there was no evidence of difference in the FRRs for mothers, sisters and daughters. There was significant evidence that the FRR for relatives of patients with EOC diagnosed under age 50 years is higher than that for older patients (4.72 (95% CI 3.21 to 6.95) and 2.53 (95% CI 1.91 to 3.35), p-diff=0.0052) and a suggestion that the FRR in relatives of patients with serous disease is higher than that for non-serous tumours (3.64 (95% CI 2.72 to 4.87) and 2.25 (95% CI 1.56 to 3.26), p-diff=0.0023). The FRR to relatives of cases without a deleterious mutation in BRCA1 or BRCA2 was estimated to be over twice that of the general population (2.24 (95% CI 1.71 to 2.94)). BRCA1 and BRCA2 mutations were estimated to account for about 24% of the EOC FRR to first-degree relatives. FRRs were found to increase with increasing polygenic risk score of the index patient, although the trend was not significant.

Conclusions These estimates could be useful in the counselling of relatives of patients with ovarian cancer.

  • Genetic screening/counselling
  • Genetic epidemiology
  • Obstetrics and Gynaecology
  • Oncology
  • Other oncology

https://doi.org/10.1136/jmedgenet-2013-102015

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Epithelial ovarian cancer (EOC) is still a disease with poor survival rates compared with other female cancers, as tumours are often undetected until an advanced stage and difficult to treat. Like many other female cancers, there is evidence that family history of ovarian cancer is one of the strongest risk factors for the disease. The risk of ovarian cancer in first-degree relatives of patients with ovarian cancer has been estimated to be approximately three times greater than the risk of women in the general population.1 Several ovarian cancer susceptibility genetic variants have been identified to date that contribute to the observed ovarian cancer familial relative risks (FRRs). These include deleterious mutations in BRCA1 and BRCA2 which confer high risks of ovarian cancer,2 mutations in the DNA mismatch repair genes MLH1, MSH2 and MSH63–5 and rare variants in RAD51C, RAD51D and BRIP1 which confer moderate risks of ovarian cancer.6–9 A total of 11 common ovarian cancer susceptibility alleles have also been identified through genome-wide association studies.10 These confer a small increase in the risk of ovarian cancer and are estimated to account for approximately 4–5% of the genetic component of ovarian cancer risk.

    Several studies have previously estimated ovarian cancer FRRs,1 ,11 but few have examined and compared the risks with different categories of first-degree relatives simultaneously or by tumour subtypes. In addition, there are no direct estimates of the FRRs to relatives of women diagnosed with ovarian cancer and found not to carry BRCA1 or BRCA2 mutations. Under the assumption that the residual familial clustering of ovarian cancer due to as yet unidentified genetic factors is polygenic, it is expected that the ovarian cancer risk in relatives of patients with ovarian cancer would increase with the increasing number of common ovarian cancer susceptibility alleles carried by the patient. However, there are no published estimates of the FRRs by combined polygenic risk profile based on the known common susceptibility alleles. Having estimates of these risks would be useful in the counselling process of relatives of patients with ovarian cancer. However, validated risk-prediction models for familial ovarian cancer such as the Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA) or BRCAPRO12 ,13 do not include information on genetic effects other than BRCA1 and BRCA2.

    We have used data from a large, population-based EOC case series, the Studies of Epidemiology and Risk factors in Cancer Heredity (SEARCH), to calculate FRRs, first for mothers, sisters and daughters of probands, then for first-degree relatives of probands diagnosed before and after age 50 years. We also computed relative risks for first-degree relatives of probands with different tumour subtypes and for first-degree relatives of probands who were BRCA1 and BRCA2 carriers and non-carriers. Finally, we constructed polygenic risk scores (PRS) on the basis of the known common ovarian cancer susceptibility alleles and estimated FRRs to first-degree relatives of probands with different levels of PRS.

    Material and methods

    Study population

    In brief, the data used consists of 1548 ovarian cancer cases (probands) recruited between 1999 and 2010, along with information on their first-degree and second-degree relatives ascertained through an epidemiological questionnaire which also included questions on reproductive and medical history, oral contraceptive use and hormone replacement therapy. The probands were all drawn from SEARCH as described for patients with breast cancer.14 SEARCH is a large population based study with cases ascertained through the Eastern Cancer Registration and Information Centre. Half-sibling status and relative type to the proband, age at cancer diagnosis, cancer site, vital status, status age (the age at death if deceased, the current age if alive) and year of birth were recorded for all probands and relatives. Tumour subtype information for all probands was extracted from routine pathology reports. Tumour subtypes were defined as clear-cell, endometrioid, high-grade and low-grade serous carcinoma (HGSC and LGSC), mixed, mucinous, non-epithelial, unclassified or other.

    BRCA1 and BRCA2 mutation screening and common SNP genotyping

    SEARCH ovarian cancer probands were screened for BRCA1 and BRCA2 mutations as part of a separate project to evaluate the contribution of rare high-risk and moderate-risk variants to overall ovarian cancer risk in the general population (unpublished). Briefly, this involved targeted sequence library preparation using multiplexed 48.48 Fluidigm access arrays and sequenced the library on Illumina HiScan according to the manufacturer's protocol. BRCA1 and BRCA2 mutation status information was available on all 1548 probands in our dataset. The following alterations were considered pathogenic: protein-truncating insertion/deletion variants, consensus splice-site variants and missense variants with reported damaging effect on protein function (Breast Cancer Information Core). For the purpose of our analysis, BRCA1 and BRCA2 mutation status was recorded simply as mutation-positive or negative, with no distinction between different mutation types by location or functional effect.

    A total of 1376 probands, including 37 BRCA1 carriers and 52 BRCA2 carriers, had also been genotyped for the 11 SNPs known to be associated with ovarian cancer risk (see online supplementary table S1). Samples were genotyped using the Illumina iCOGS custom array as part of an experiment performed by the Collaborative Oncological Gene Environment Study.10 Where genotypes for any of these SNPs were missing, they were replaced with inferred genotypes via imputation. BRCA1 and BRCA2 mutation status and SNP data were not available for any of the relatives.

    Statistical methods

    FRRs to first-degree relatives

    Ovarian cancer FRRs were estimated as the ratio of the observed number of cases to the expected number of cases under population incidence rates from a cohort analysis of the relatives of SEARCH index case probands. All at-risk women entered the cohort on 1 January 1960 or their birth date, whichever was later, and all individuals were censored at the first occurrence of death, ovarian cancer diagnosis or age 85 years. Relatives with missing year of birth were given birth dates depending on their relationship to the proband. For example, parents were assigned a birth year 30 years prior to the proband and daughters a birth year 25 years after the proband. Missing ages were then estimated as the lag between birth year and recruitment date and edited using decennial life expectancies when vital status is ‘dead’ or missing. For example, mothers of probands born from 1940 to 1950 had their ages censored at the average total life expectancy for 30-year-old women in that decade. All ovarian cancer diagnosis ages were censored at age 85 years.

    The expected numbers of cancers, among first-degree relatives, labelled 1–K, were computed as Embedded Image where rk is the expected cumulative ovarian cancer incidence in individual k. For a relative with birth date Bk and censored age Nk, Embedded Image where Embedded Image is the population incidence of ovarian cancer among people in the age group Embedded Image in year Embedded Image. The population incidences used were obtained from the age, sex and time-period-specific incidence rates for England and Wales published in Cancer in Five Continents (http://ci5.iarc.fr/CI5i-ix/ci5i-ix.htm). FRRs were first calculated for all first-degree female relatives and separately for mothers, sisters and daughters; then for first-degree relatives of probands diagnosed before and after age 50 years. Ovarian cancer relative risks to relatives of probands by tumour subtype and by mutation status (BRCA1, BRCA2 or neither) were also computed. The proportion of risk accounted for by BRCA1 and BRCA2 mutations was estimated as 1−(R0/RT) where R0 and RT are the logarithms of the FRRs to non-carriers’ and all probands’ relatives, respectively, under the assumption that the contributions to the FRR are additive on the log scale.

    To estimate the FRRs by combined SNP profile, we constructed a PRS assuming that all loci interact multiplicatively on the relative risk scale. For this purpose, the genotype gk for each SNP was coded as the number of risk alleles carried (0, 1 or 2) if known or the expected genotype if imputed. The PRS was computed as Embedded Image where βk is the logarithm of the OR for the risk associated with each allele of the kth SNP. The log-ORs used are shown in online supplementary table S1. FRRs were estimated for all first-degree relatives of probands stratified by quartile of the PRS. Separate estimates were obtained for the relatives of mutation-negative or serous-tumour probands.

    To address the correlation between individuals from the same family, for all analyses robust SEs were computed and used to calculate 95% CIs and perform significance tests of the difference in risk to relatives of probands with different characteristics. Under the assumption that the FRRs were all log-normally distributed, the statistic used to test for difference between FRRs r1 and r2 wasEmbedded Image where σi is the robust SE for log (ri.) Statistical analyses were carried out using the Stata V.11 software.

    Results

    The numbers of SEARCH ovarian cancer probands and their first-degree relatives, the number of ovarian cancers diagnosed in each group and other sample characteristics are summarised in table 1 and online supplementary table S2.

    View this table:
    Table 1

    Number of probands, families and relatives used in the analysis, and number of ovarian cancer cases in each group

    Familial relative risks by relative type and proband age

    Data from 1548 ovarian cancer cases recruited into the SEARCH study were used for our analyses. These probands’ first-degree relatives included 1340 mothers, 1404 sisters and 1144 daughters, of whom 80 were also diagnosed with ovarian cancer.

    Table 2 summarises the estimated FRRs by relationship and proband diagnosis age. The estimated ovarian cancer risks to any type of first-degree relative, to the mother and to a sister of a proband were all very close to three times the population risk 2.96 (95% CI 2.35 to 3.72), 2.88 (95% CI 2.19 to 3.79) and 2.98 (95% CI 2.03 to 4.37), respectively.

    View this table:
    Table 2

    Estimated ovarian cancer familial relative risks (FRRs) by first-degree relative and proband's age at diagnosis

    The relative risk to daughters was estimated to be somewhat larger at 4.78 (95% CI 1.54 to 14.83), although this was not significantly different from the other FRRs (p=0.2). The FRR estimates for all relatives, mothers and sisters of probands diagnosed under age 50 years, were 4.72, 4.67 and 5.11, respectively, and were all noticeably larger than those for relatives of probands diagnosed over age 50 years, at 2.53, 2.31 and 2.71, respectively. These differences in risk were statistically significant for all relatives grouped together and individually for mothers (p=0.0052 and 0.0073, respectively) but not for sisters (p=0.09).

    Familial relative risks by tumour subtype

    The FRR estimates to relatives of probands with different tumour subtypes are displayed in table 3. Comparing the ovarian cancer FRRs for probands with serous and non-serous tumours, there was a suggestion that first-degree relatives of patients with serous tumours are at higher risk of developing ovarian cancer than relatives of patients with a non-serous tumour (FRR for all relatives=3.64 vs 2.25, p=0.023). These differences were also seen when analyses were stratified by the age at diagnosis of the proband and by type of relative (table 3). The FRRs to daughters of probands diagnosed after age 50 years were estimated to be larger for non-serous than for serous tumours, but the difference in estimates were not significantly different.

    View this table:
    Table 3

    Estimated ovarian cancer familial relative risks (FRRs) by serous and non-serous ovarian cancer subtype

    Table 4 shows the FRR estimates for relatives of probands further stratified by the eight different tumour subtypes recorded in SEARCH. Relatives of probands with endometrioid, high-grade serous and ‘other’ tumours were at significantly increased risk of developing ovarian cancer (FRR=3.81 (95% CI 2.27 to 6.39), 3.85 (95% CI 2.84 to 5.21) and 3.4 (95% CI 1.52 to 7.64), respectively). Although the FRR estimates for relatives of probands with other tumour subtypes were >1, the number of probands falling into these categories and number of ovarian cancers in relatives were relatively small. Hence these estimates were associated with wide CIs. FRR estimates for mothers and sisters and for relatives of probands stratified by age were also all significantly different from 1 for relatives of probands diagnosed with endometrioid carcinoma or HGSC (see online supplementary table S3). However, the numbers for relatives of probands with other tumour subtypes were too small to draw definitive conclusions.

    View this table:
    Table 4

    Estimated ovarian cancer familial relative risks (FRRs) by ovarian cancer subtype in the proband

    Familial relative risks by BRCA1 and BRCA2 mutation status

    Table 5 displays the FRRs to first-degree relatives of probands stratified by BRCA1 and BRCA2 mutation status. As expected, relatives of mutation carriers were at a substantially increased risk of developing ovarian cancer. The FRR in relatives of BRCA1 mutation carriers was estimated to be 20.97 (95% CI 11.94 to 36.81) and 9.57 in relatives of BRCA2 mutation carriers (95% CI 5.25 to 17.45) (p=0.031 for difference between BRCA1 and BRCA2). After excluding probands with identified BRCA1 or BRCA2 mutations, there were 57 ovarian cancer cases in first-degree relatives of probands. These were spread evenly among 54 families, 3 of which had 2 ovarian cancer cases in relatives of the probands. The rest of the families contained only one first relative who had been diagnosed with ovarian cancer. Relatives of probands in whom no BRCA1 or BRCA2 mutation was identified were also at significantly increased risk of developing ovarian cancer (FRR=2.24, 95% CI 1.71 to 2.94). However, this FRR was significantly lower than the FRRs for relatives of BRCA1 and BRCA2 mutation carriers (p=1.2 × 10−12 and p=8 × 10−6, respectively). When further stratified by the age at diagnosis of proband, relatives of non-BRCA1 or non-BRCA2 carriers diagnosed at a younger age were at a significantly increased risk of ovarian cancer compared with relatives of non-carriers diagnosed at older ages (FRR=3.83 vs FRR=1.88, p =0.011). Relatives of BRCA1 carriers diagnosed under age 50 years appeared to be at a lower risk of developing ovarian cancer than relatives of BRCA1 carriers diagnosed over age 50 years, though not significantly so (p=0.07). The risk to BRCA2 carriers’ relatives, in contrast, did not vary by the diagnosis age of the proband. When analyses were stratified by the proband's tumour subtype (serous vs non-serous), we observed elevated ovarian cancer risks in relatives of non-carrier patients with serous and non-serous tumours. However, the FRR for relatives of non-carriers diagnosed with serous ovarian cancer were higher, though not significantly so (FRR=2.56 vs 1.94, p=0.15). There were no significant differences in the FRR estimates for relatives of BRCA1 and BRCA2 mutation carriers when stratified by serous and non-serous tumour subtypes.

    View this table:
    Table 5

    Estimated ovarian cancer familial relative risks (FRRs) by the proband BRCA1 and BRCA2 mutation status

    Our results suggest that 25.4% of the observed FRR of ovarian cancer in first-degree relatives is accounted for by BRCA1 and BRCA2 mutations.

    FRRs by PRS quartile

    Table 6 displays the FRRs to first-degree relatives of probands stratified by quartile of the observed PRS distribution for all 1376 probands with SNP genotype data, and restricted to probands with serous tumours and probands with no BRCA1 or BRCA2 mutation. Overall, the patterns were suggestive of increasing FRR with increasing PRS in the proband. However the test for trend was not significant. The estimated FRRs increased from the first PRS quartile to the third PRS quartile but relatives of probands in the fourth PRS quartile were estimated to be at a somewhat lower ovarian cancer risk compared with relatives of probands in the 25–75th centiles of the PRS distribution. Similar patterns were observed when analyses were restricted to probands diagnosed with serous ovarian cancer, although FRRs were generally higher. Relatives of non-BRCA1 and non-BRCA2 carrier probands with low PRS were still at a significantly increased risk of developing ovarian cancer (FRR>2).

    View this table:
    Table 6

    Estimated ovarian cancer familial relative risks (FRRs) by the distribution of the polygenic risk score (PRS) in the proband

    Discussion

    We have used a population-based series of EOC cases to estimate the relative risk of ovarian cancer for first-degree relatives of patients with ovarian cancer with different tumour subtypes, different BRCA1 and BRCA2 mutation status and different levels of polygenic background based on the combined effects of common low-penetrance susceptibility alleles. We found strong evidence that the ovarian cancer risks to any first-degree relative (including mothers and sisters) of patients with ovarian cancer are all about threefold greater compared with the general population. Our analyses also showed some evidence of age dependence, with the risk to relatives of patients diagnosed younger than 50 years being as much as twice that to relatives of older patients with ovarian cancer. We found that the most common (classified) types, endometrioid and high-grade serous, which make up approximately 16% and 41%, respectively, of all ovarian tumours, are associated with a significant increase in familial risk while our sample size for each of the rarer LGSC, clear-cell, mixed, mucinous tumours and non-epithelial tumours was too small to draw definitive conclusions for individual subtype categories.

    Our study is large compared with the majority of previous studies, with 1548 index patients used in the analysis compared with only 824 cases in the largest case-control study15 and 599 in11 one of the few previous cohort studies. Nevertheless, many of our results are consistent with previous publications. A meta-analysis of 15 cohort and case-control studies estimated a relative risk of 3.1 for all ages and relatives.1 However, published estimates of relative risk by age of the affected proband are inconsistent. Auranen et al estimated relative risks to sisters of probands diagnosed younger and older than 50 years to be 5.1 and 3.4, respectively.11 In contrast, Easton et al estimated the risk to relatives where the index case was diagnosed with ovarian cancer before the age of 40 years, and found it to be lower than if she was diagnosed with ovarian cancer at an older age (1.7 vs 3.8).16

    We found the highest relative risks were in relatives of probands with high-grade serous or endometrioid cancer. These findings are similar to the few published studies that have reported ovarian cancer FRRs by tumour pathology.17 ,18 As HGSC is the subtype primarily associated with BRCA1 and BRCA2 mutations, one would expect the FRR to be higher for this subtype.

    When analyses were performed by BRCA1 and BRCA2 mutation status of the proband, we found clear evidence of elevated ovarian cancer risks in relatives of non-mutation carriers. Our FRR estimate of 2.24 for relatives of non-carriers may be a slight overestimate due to the fact that the sensitivity of BRCA1and BRCA2 mutation screening is not 100%. We have estimated that the sensitivity of the BRCA1 and BRCA2 mutation screening performed in the current study is approximately 83% for BRCA1 and 76% for BRCA2 (Song et al, Personal Communication). Adjusting for this reduced sensitivity the FRR in relatives of non-BRCA1 and non-BRCA2 mutation carriers would be slightly lower at 2.19. Moreover, if we were to exclude all families in which the proband had a first-degree relative diagnosed with breast cancer, a likely indication of the presence of a BRCA1 or BRCA2 mutation, the FRR estimate in relatives of non-BRCA1 and non-BRCA2 mutation carriers is still elevated at 2.15 (95% CI 1.62 to 2.87). The higher FRRs in relatives of BRCA1 mutation carriers compared with relatives of BRCA2 mutation carriers are consistent with previous penetrance estimates for these mutations where ovarian cancer risks conferred by BRCA1 mutations are much higher than those conferred by BRCA2 mutations.2 However, published data on the ovarian cancer risk in relatives of probands that are BRCA1 and BRCA2 mutation negative are more limited. A recent study investigated risks of ovarian cancer in families without identifiable BRCA1 and BRCA2 mutations but these families were ascertained on the basis of primarily having multiple members diagnosed with breast cancer.19 One study of 382 women from 56 multicase ovarian cancer families reported a relative risk of 12.20 Another study of 339 probands estimated the relative risks to first-degree relatives of BRCA1 mutation-negative, BRCA1 mutation carriers and all patients with ovarian cancer to be 1.9 (95% CI 1.0 to 4.0), 11 (95% CI 3.6 to 36) and 2.5 (95% CI 1.6 to 4.0), respectively.21

    Our study is the first of its kind to report FRRs by the combined effects of SNPs associated with ovarian cancer. The upwards trend in estimated FRRs seen from the probands in the first to third quartile of the PRS distribution (and seen also in analyses restricted to serous tumours) is consistent with the hypothesis of polygenic susceptibility to ovarian cancer. This trend was not observed for relatives of probands in the highest PRS quartile, but the associated 95% CI is wide. Furthermore, the range of PRSs in the probands is quite small based on only 11 SNPs (range: −0.5 to 1.3) and the effect on the FRR would also be expected to be limited. It is also possible that the effect in the top quartile of the PRS reflects a deviation from the polygenic model at the extremes of the distribution. Much larger series of patients with ovarian cancer and additional SNPs would be required to clarify this. One of our key findings is that relatives of patients with low PRS have significantly higher risks of developing ovarian cancer, even after excluding BRCA1 and BRCA2 mutations. Rare mutations in other known ovarian cancer susceptibility genes, such as RAD51C, RAD51D, BRIP1 and mutations in the DNA mismatch repair genes cannot explain this elevated risk, because they account for only a small proportion of the observed familial aggregation of ovarian cancer (<5%).22 These suggest that additional genetic variants remain to be identified, either common or rare. Future segregation analyses incorporating the explicit effects of known genetic effects will aim to address this question.

    One shortcoming of a retrospective cohort study design is the possibility of errors in the reporting of family cancer history. For first-degree relatives, previous studies have however found reported ovarian cancer history to be reasonably accurate (83.3% probability of agreement between reported ovarian cancer status in first-degree relatives and status according to reference standard, 95% CI 72.8 to 93.8).23 ,24 A further drawback of our study is the lack of information on tumour subtypes in relatives. Thus we were able to estimate FRRs of overall ovarian cancer risk in relatives of patients with different tumour subtypes, but we were unable to obtain subtype-specific FRRs. Future studies of this kind are warranted that would clarify further the genetic architecture of different ovarian cancer subtypes.

    In conclusion, our results will be informative in the counselling of relatives of women diagnosed with ovarian cancer. Our results demonstrate clearly that female relatives of women diagnosed with ovarian cancer and without BRCA1 or BRCA2 mutations are at elevated risk of developing ovarian cancer. Our findings provide insights into the architecture of familial ovarian cancer and such findings would be of high value in the construction of risk prediction models for familial ovarian cancer.

    Acknowledgments

    The authors thank all the study participants who contributed to this study and all the researchers, clinicians and technical and administrative staff who have made possible this work. In particular, the authors thank Craig Luccarini, Caroline Baynes, the SEARCH team and the Eastern Cancer Registration and Information Centre. ACA is a Cancer Research UK Senior Cancer Research Fellow. IJJ is a National Institute for Health Research Senior Investigator.

    References

    1. ↵
      1. Stratton JF,
      2. Pharoah P,
      3. Smith SK,
      4. Easton D,
      5. Ponder BA
      . A systematic review and meta-analysis of family history and risk of ovarian cancer. Br J Obstet Gynaecol 1998;105:493–9.
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Antoniou A,
      2. Pharoah PD,
      3. Narod S,
      4. Risch HA,
      5. Eyfjord JE,
      6. Hopper JL,
      7. Loman N,
      8. Olsson H,
      9. Johannsson O,
      10. Borg A,
      11. Pasini B,
      12. Radice P,
      13. Manoukian S,
      14. Eccles DM,
      15. Tang N,
      16. Olah E,
      17. Anton-Culver H,
      18. Warner E,
      19. Lubinski J,
      20. Gronwald J,
      21. Gorski B,
      22. Tulinius H,
      23. Thorlacius S,
      24. Eerola H,
      25. Nevanlinna H,
      26. Syrjakoski K,
      27. Kallioniemi OP,
      28. Thompson D,
      29. Evans C,
      30. Peto J,
      31. Lalloo F,
      32. Evans DG,
      33. Easton DF
      . Average risks of breast and ovarian cancer associated with BRCA1 or BRCA2 mutations detected in case Series unselected for family history: a combined analysis of 22 studies. Am J Hum Genet 2003;72:1117–30.
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Stratton JF,
      2. Thompson D,
      3. Bobrow L,
      4. Dalal N,
      5. Gore M,
      6. Bishop DT,
      7. Scott I,
      8. Evans G,
      9. Daly P,
      10. Easton DF,
      11. Ponder BAJ
      . The genetic epidemiology of early-onset epithelial ovarian cancer: a population-based study. Am J Hum Genet 1999;65:1725–32.
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Lynch HT,
      2. Casey MJ,
      3. Snyder CL,
      4. Bewtra C,
      5. Lynch JF,
      6. Butts M,
      7. Godwin AK
      . Hereditary ovarian cancer: molecular genetics, pathology, management, and heterogeneity. Mol Oncol 2009;3:97–137.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Zhai Qœ,
      2. Rosen DG,
      3. Lu K,
      4. Liu J
      . Loss of DNA mismatch repair protein hMSH6 in ovarian cancer is histotype-specific. Int J Clin Exp Pathol 2008;1:502–9.
      OpenUrlPubMed
    6. ↵
      1. Meindl A,
      2. Ditsch N,
      3. Kast K,
      4. Rhiem K,
      5. Schmutzler RK
      . Hereditary breast and ovarian cancer: new genes, new treatments, new concepts. Dtsch Arztebl Int 2011;108:323–30.
      OpenUrlPubMed
    7. ↵
      1. Clague J,
      2. Wilhoite G,
      3. Adamson A,
      4. Bailis A,
      5. Weitzel JN,
      6. Neuhausen SL
      . RAD51C Germline mutations in breast and ovarian cancer cases from high-risk families. PLoS One 2011;6:e25632.
      OpenUrlCrossRefPubMed
    8. ↵
      1. Song H,
      2. Ramus SJ,
      3. Kjaer SK,
      4. Hogdall E,
      5. DiCioccio RA,
      6. Whittemore AS,
      7. McGuire V,
      8. Hogdall C,
      9. Jacobs IJ,
      10. Easton DF,
      11. Ponder BA,
      12. Dunning AM,
      13. Gayther SA,
      14. Pharoah DP
      . Tagging single nucleotide polymorphisms in the BRIP1 gene and susceptibility to breast and ovarian cancer. PLoS One 2007;2:e268.
      OpenUrlCrossRefPubMed
    9. ↵
      1. Kuusisto KM,
      2. Bebel A,
      3. Vihinen M,
      4. Schleutker J,
      5. Sallinen SL
      . Screening for BRCA1, BRCA2, CHEK2, PALB2, BRIP1, RAD50, and CDH1 mutations in high-risk Finnish BRCA1/2-founder mutation-negative breast and/or ovarian cancer individuals. Breast Cancer Res 2011;13:R20.
      OpenUrlCrossRefPubMed
    10. ↵
      1. Pharoah PD,
      2. Tsai YY,
      3. Ramus SJ,
      4. Phelan CM,
      5. Goode EL,
      6. Lawrenson K,
      7. Buckley M,
      8. Fridley BL,
      9. Tyrer JP,
      10. Shen H,
      11. Weber R,
      12. Karevan R,
      13. Larson MC,
      14. Song H,
      15. Tessier DC,
      16. Bacot F,
      17. Vincent D,
      18. Cunningham JM,
      19. Dennis J,
      20. Dicks E,
      21. Aben KK,
      22. Anton-Culver H,
      23. Antonenkova N,
      24. Armasu SM,
      25. Baglietto L,
      26. Bandera EV,
      27. Beckmann MW,
      28. Birrer MJ,
      29. Bloom G,
      30. Bogdanova N,
      31. Brenton JD,
      32. Brinton LA,
      33. Brooks-Wilson A,
      34. Brown R,
      35. Butzow R,
      36. Campbell I,
      37. Carney ME,
      38. Carvalho RS,
      39. Chang-Claude J,
      40. Chen YA,
      41. Chen Z,
      42. Chow WH,
      43. Cicek MS,
      44. Coetzee G,
      45. Cook LS,
      46. Cramer DW,
      47. Cybulski C,
      48. Dansonka-Mieszkowska A,
      49. Despierre E,
      50. Doherty JA,
      51. Dork T,
      52. du BA,
      53. Durst M,
      54. Eccles D,
      55. Edwards R,
      56. Ekici AB,
      57. Fasching PA,
      58. Fenstermacher D,
      59. Flanagan J,
      60. Gao YT,
      61. Garcia-Closas M,
      62. Gentry-Maharaj A,
      63. Giles G,
      64. Gjyshi A,
      65. Gore M,
      66. Gronwald J,
      67. Guo Q,
      68. Halle MK,
      69. Harter P,
      70. Hein A,
      71. Heitz F,
      72. Hillemanns P,
      73. Hoatlin M,
      74. Hogdall E,
      75. Hogdall CK,
      76. Hosono S,
      77. Jakubowska A,
      78. Jensen A,
      79. Kalli KR,
      80. Karlan BY,
      81. Kelemen LE,
      82. Kiemeney LA,
      83. Kjaer SK,
      84. Konecny GE,
      85. Krakstad C,
      86. Kupryjanczyk J,
      87. Lambrechts D,
      88. Lambrechts S,
      89. Le ND,
      90. Lee N,
      91. Lee J,
      92. Leminen A,
      93. Lim BK,
      94. Lissowska J,
      95. Lubinski J,
      96. Lundvall L,
      97. Lurie G,
      98. Massuger LF,
      99. Matsuo K,
      100. McGuire V,
      101. McLaughlin JR,
      102. Menon U,
      103. Modugno F,
      104. Moysich KB,
      105. Nakanishi T,
      106. Narod SA,
      107. Ness RB,
      108. Nevanlinna H,
      109. Nickels S,
      110. Noushmehr H,
      111. Odunsi K,
      112. Olson S,
      113. Orlow I,
      114. Paul J,
      115. Pejovic T,
      116. Pelttari LM,
      117. Permuth-Wey J,
      118. Pike MC,
      119. Poole EM,
      120. Qu X,
      121. Risch HA,
      122. Rodriguez-Rodriguez L,
      123. Rossing MA,
      124. Rudolph A,
      125. Runnebaum I,
      126. Rzepecka IK,
      127. Salvesen HB,
      128. Schwaab I,
      129. Severi G,
      130. Shen H,
      131. Shridhar V,
      132. Shu XO,
      133. Sieh W,
      134. Southey MC,
      135. Spellman P,
      136. Tajima K,
      137. Teo SH,
      138. Terry KL,
      139. Thompson PJ,
      140. Timorek A,
      141. Tworoger SS,
      142. van Altena AM,
      143. van den Berg D,
      144. Vergote I,
      145. Vierkant RA,
      146. Vitonis AF,
      147. Wang-Gohrke S,
      148. Wentzensen N,
      149. Whittemore AS,
      150. Wik E,
      151. Winterhoff B,
      152. Woo YL,
      153. Wu AH,
      154. Yang HP,
      155. Zheng W,
      156. Ziogas A,
      157. Zulkifli F,
      158. Goodman MT,
      159. Hall P,
      160. Easton DF,
      161. Pearce CL,
      162. Berchuck A,
      163. Chenevix-Trench G,
      164. Iversen E,
      165. Monteiro AN,
      166. Gayther SA,
      167. Schildkraut JM,
      168. Sellers TA
      . GWAS meta-analysis and replication identifies three new susceptibility loci for ovarian cancer. Nat Genet 2013;45:362.
      OpenUrlCrossRefPubMed
    11. ↵
      1. Auranen A,
      2. Pukkala E,
      3. Makinen J,
      4. Sankila R,
      5. Grenman S,
      6. Salmi T
      . Cancer incidence in the first-degree relatives of ovarian cancer patients. Br J Cancer 1996;74:280–4.
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Antoniou AC,
      2. Pharoah PDP,
      3. McMullan G,
      4. Day NE,
      5. Stratton MR,
      6. Peto J,
      7. Ponder BJ,
      8. Easton DF
      . A comprehensive model for familial breast cancer incorporating BRCA1, BRCA2 and other genes. Br J Cancer 2002;86:76–83.
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Parmigiani G,
      2. Berry D,
      3. Aguilar O
      . Determining carrier probabilities for breast cancer-susceptibility genes BRCA1 and BRCA2. Am J Hum Genet 1998; 62:145–58.
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Pharoah PDP,
      2. Tyrer J,
      3. Dunning AM,
      4. Easton DF,
      5. Ponder BAJ
      . Association between common variation in 120 Candidate genes and breast cancer risk. PLoS Genet 2007;3:401–6.
      OpenUrlWeb of Science
    15. ↵
      1. Purdie D,
      2. Green A,
      3. Bain C,
      4. Siskind V,
      5. Ward B,
      6. Hacker N,
      7. Quinn M,
      8. Wright G,
      9. Russell P,
      10. Susil B
      . Reproductive and other factors and risk of epithelial ovarian cancer: an Australian case-control study. Survey of Women's Health Study Group. Int J Cancer 1995;62:678–84.
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Easton DF,
      2. Matthews FE,
      3. Ford D,
      4. Swerdlow AJ,
      5. Peto J
      . Cancer mortality in relatives of women with ovarian cancer: the OPCS Study. Office of Population Censuses and Surveys. Int J Cancer 1996;65:284–94.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Hemminki K,
      2. Sundquist J,
      3. Brandt A
      . Incidence and mortality in epithelial ovarian cancer by family history of any cancer. Cancer 2011;117:3972–80.
      OpenUrlCrossRefPubMed
    18. ↵
      1. Lorenzo BJ,
      2. Rawal R,
      3. Hemminki K
      . Familial association of specific histologic types of ovarian malignancy with other malignancies. Cancer 2004;100:1507–14.
      OpenUrlCrossRefPubMed
    19. ↵
      1. Ingham SL,
      2. Warwick J,
      3. Buchan I,
      4. Sahin S,
      5. O'Hara C,
      6. Moran A,
      7. Howell A,
      8. Evans DG
      . Ovarian cancer among 8,005 women from a breast cancer family history clinic: no increased risk of invasive ovarian cancer in families testing negative for BRCA1 and BRCA2. J Med Genet 2013;50:368–72.
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Sutcliffe S,
      2. Pharoah PD,
      3. Easton DF,
      4. Ponder BA
      . Ovarian and breast cancer risks to women in families with two or more cases of ovarian cancer. Int J Cancer 2000;87:110–17.
      OpenUrlCrossRefPubMed
    21. ↵
      1. Lee JS,
      2. John EM,
      3. McGuire V,
      4. Felberg A,
      5. Ostrow KL,
      6. DiCioccio RA,
      7. Li FP,
      8. Miron A,
      9. West DW,
      10. Whittemore AS
      . Breast and ovarian cancer in relatives of cancer patients, with and without BRCA mutations. Cancer Epidemiol Biomarkers Prev 2006;15:359–63.
      OpenUrlAbstract/FREE Full Text
    22. ↵
      1. Bahcall OG
      . iCOGS collection provides a collaborative model. Foreword. Nat Genet 2013;45:343.
      OpenUrlCrossRefPubMed
    23. ↵
      1. Ziogas A,
      2. Anton-Culver H
      . Validation of family history data in cancer family registries. Am J Prev Med 2003;24:190–8.
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      1. Mai PL,
      2. Garceau AO,
      3. Graubard BI,
      4. Dunn M,
      5. McNeel TS,
      6. Gonsalves L,
      7. Gail MH,
      8. Greene MH,
      9. Willis GB,
      10. Wideroff L
      . Confirmation of family cancer history reported in a population-based survey. J Natl Cancer Inst 2011;103:788–97.
      OpenUrlAbstract/FREE Full Text

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

      Files in this Data Supplement:

    Footnotes

    • Contributors Conception and design: ACA, PPDP. Analysis: SJ, ACA, HS, AL, ED, JT, PH. Interpretation of data: SJ, ACA, PPDP, DFE, IJJ. Acquisition of data: PPDP, DFE. Drafting the article: SJ, ACA. Critical revision: all authors. Final approval: all authors.

    • Funding This work has been supported by grants from Cancer Research UK (C1005/A12677, C12292/A11174, C490/A10119, C490/A10124) including the PROMISE research programme, the Eve Appeal and the UK National Institute for Health Research Biomedical Research Centre at the University of Cambridge.

    • Competing interests None.

    • Patient consent Obtained.

    • Ethics approval Cambridgeshire 4 research ethics committee.

    • Provenance and peer review Not commissioned; externally peer reviewed.