Article Text

Download PDFPDF

Screening for familial ovarian cancer: poor survival of BRCA1/2 related cancers
  1. D G Evans1,
  2. K N Gaarenstroom2,
  3. D Stirling3,
  4. A Shenton1,
  5. L Maehle4,
  6. A Dørum5,
  7. M Steel6,
  8. F Lalloo1,
  9. J Apold7,
  10. M E Porteous3,
  11. H F A Vasen8,
  12. C J van Asperen9,
  13. P Moller4
  1. 1
    Academic Unit of Medical Genetics and Regional Genetics Service, St Mary’s Hospital, Manchester, UK
  2. 2
    Department of Gynaecology, Leiden University Medical Center, Leiden, The Netherlands
  3. 3
    South East of Scotland Genetics Service, Western General Hospital, Edinburgh, UK
  4. 4
    Section for Inherited Cancer, Department of Medical Genetics, Rikshospitalet Radiumhospitalet Clinical Center, Oslo, Norway
  5. 5
    Department of Gynecologic Oncology, The Norwegian Radium Hospital, Rikshospitalet HF, University of Oslo, Oslo, Norway
  6. 6
    University of St Andrews, Bute Medical Buildings, St Andrews, UK
  7. 7
    Centre of Medical Genetics and Molecular Medicine, Haukeland University Hospital, and Institute of Clinical Medicine, University of Bergen, Bergen, Norway
  8. 8
    The Netherlands Foundation for the Detection of Hereditary Tumours and the Department of Gastroenterology, Leiden University Medical Center, The Netherlands
  9. 9
    Center for Human and Clinical Genetics, Department of Clinical Genetics, Leiden University Medical Center, Leiden, The Netherlands
  1. Correspondence to Professor D G Evans, Academic Unit of Medical Genetics and Regional Genetics Service, St Mary’s Hospital, Manchester M13 OJH, UK; gareth.evans{at}


Aim: To assess the effectiveness of annual ovarian cancer screening (transvaginal ultrasound and serum CA125 estimation) in reducing mortality from ovarian cancer in women at increased genetic risk.

Patients and methods: A cohort of 3532 women at increased risk of ovarian cancer was screened at five European centres between January 1991 and March 2007. Survival from diagnosis of ovarian cancer was calculated using Kaplan–Meier analysis and compared for proven BRCA1/2 carriers with non-carriers and whether the cancer was detected at prevalence or post-prevalent scan. Screening was performed by annual transvaginal ultrasound and serum CA125 measurement.

Results: 64 epithelial ovarian malignancies (59 invasive and 5 borderline), developed in the cohort. 26 tumours were detected at prevalent round; there were 27 incident detected cancers and 11 interval. 65% of cancers were stage 3 or 4, however, stage and survival were little different for prevalent versus post-prevalent cancers. Five year and 10 year survival in 49 BRCA1/2 mutation carriers was 58.6% (95% CI 50.9% to 66.3%) and 36% (95% CI 27% to 45%), which was significantly worse than for 15 non-BRCA carriers (91.8%, 95% CI 84% to 99.6%, both 5 and 10 year survival p = 0.015). However, when borderline tumours were excluded, the difference in survival between carriers and non-carriers was no longer significant.

Conclusion: Annual surveillance, by transvaginal ultrasound scanning and serum CA125 measurement, in women at increased familial risk of ovarian cancer is ineffective in detecting tumours at a sufficiently early stage to influence substantially survival in BRCA1/2 carriers.

Statistics from

Advances in the understanding of the genetics of ovarian cancer, especially relating to the BRCA1 and BRCA2 genes, have led to an increased public and professional awareness of risks associated with a family history of the disease. Demand for ovarian cancer screening has risen, yet there is no clear evidence of its effectiveness in groups at increased risk.1 2 3 4 5 6 7 8 While many studies have reported on the stage at diagnosis, these studies are small in terms of the numbers of ovarian cancers and none have reported on survival. Despite the paucity of evidence of efficacy of ovarian screening, many reviews remain upbeat about the potential outcomes of surveillance.9

Epithelial tumours account for over 90% of all ovarian cancers and the lifetime risk is estimated to be between 1–1.5%. This risk increases with the number of affected relatives and the closeness of relationship. The lifetime risk estimate for women with one first degree relative is 2–5% and when there is more than one close relative affected, the risk increases to 10–23%.1 2 10 Highly penetrant cancer predisposing genes account for 8–13% of all epithelial ovarian cancers in non-founder populations,2 11 12 but this increases to 23% in Norway and even higher in the Jewish population.13 Those identified to date include BRCA1 and BRCA2, associated with breast and ovarian cancer, and the mismatch repair genes (MMR) associated with hereditary non-polyposis colon cancer (HNPCC). Germline mutations in these two classes of genes can confer an increased risk of ovarian cancer of up to 60% and 12%, respectively11 14 15 In most instances, increased genetic risk is recognised first from family history alone and mutation screening may follow. For women at increased risk, management options include ovarian screening and prophylactic surgery.2 This study has examined the efficacy of annual ovarian screening from five European centres (representing most of the published high risk screening previously published) in reducing ovarian cancer mortality.

Patients and methods

Five cancer genetics centres participated: Edinburgh and Manchester (UK), Leiden (The Netherlands), and Oslo and Bergen (Norway). These centres were among 10 centres involved in a European Biomed project funded in 1995 (there were insufficient data and cancers in the other centres to increase power). All women presenting to one of these, between January 1991 and March 2007, had their family history documented, assessed and confirmed, as far as practicable, by Clinical Genetic Services. Women assessed as being at increased risk of ovarian cancer (usually at least a 10% lifetime risk, requiring more than just a single close relative with ovarian cancer) were offered an annual screening programme comprising CA125 serum monitoring and transvaginal ultrasound, starting at either 30 or 35 years of age (screening in Norway was performed in several local hospitals). All screening was in accordance with local ethical review board approval. Early results from these centres and more details on ascertainment and screening protocols are available elsewhere.4 5 6 7 8

Genetic testing of BRCA1 and BRCA2 and MMR genes was only available to families meeting the high risk criteria (where there was a living affected relative). Women diagnosed with ovarian cancer during the study were offered genetic testing after diagnosis.

Information regarding stage and histology were documented for each cancer occurring at initial screening assessment (prevalence) and during follow-up. Interval cancers were those occurring symptomatically within 12–14 months of a previously normal screening round, and incidence cancers were those detected either by CA125, ultrasound or both after the initial prevalent round. Kaplan–Meier survival analysis was used to assess survival. Hospital notes and cancer and other registries were used to establish date at death and current vital status as of 1 March 2007. This was a prospective observational study without a control group, and with the observed group stratified according to results of mutation testing.


A total of 3532 women have received annual ovarian screening for up to 16 years. There were 981 known BRCA1/2 carriers in this cohort; 64 epithelial ovarian malignancies (59 invasive and five borderline) developed in the cohort. This represents an additional 17 cancers (nine in Manchester, eight in Norway) from the previous combined reports of 47 cancers.5 6 7 8 As previously reported, the great majority of cancers were serous, although there were seven endometrioid tumours and two clear cell carcinomas. All tumours were of epithelial origin. All but four cases had been fully tested for the appropriate genes and 49 BRCA1/2 carriers and two MMR gene carriers were identified. Twenty-six tumours were detected at prevalent round; there were 27 incident detected cancers and 11 interval. Sixty-five per cent of cancers were stage 3 or 4 at diagnosis. There was little evidence of down staging from prevalent round with 65/63% of both prevalent and follow-up cancers being at stage 3 or 4 (table 1).

Table 1

Stage of cancers detected and proportion who have since died

This increased to 77% and 65% if borderline tumours were excluded. There was no difference in survival curves between prevalent versus post-prevalent cancers (detected during screening) (figure 1); this was also true for BRCA1/2 cancers when treated separately. Twenty-four of 64 (38%) women had died by March 2007 and the proportion increased to 24 of 59 (41%) of invasive cancers. Survival by stage showed a highly significantly worse survival for stage 3 and 4 combined compared with stage 1 and 2, as expected (figure 2; 5 year survival: stage 1–2, 83%; stage 3–4, 55%; log rank p = 0.0032). Fewer than 20% of women with advanced stage ovarian cancer were alive at 10 years. The two women with MSH2 mutations were detected at stage 1c at incident screens aged 32 and 34 years 2653 C>T (endometriod – 2653 C>T Q885X); clear cell – 942+3 A>T). BRCA1/2 interval cancers faired particularly poorly (table 2)—only 2/9 are still alive and they are both within 2 years of diagnosis. The two non-BRCA ovarian interval cancers were both low grade stage 3c and are still alive 3.5 and 6 years post-diagnosis.

Figure 1

(A) Prevalence versus post-prevalent survival analysis (including borderline). (B) Prevalence versus post-prevalent survival analysis (excluding borderline).

Figure 2

(A) Stage 1–2 versus stage 3–4 survival. (B) Stage 1–2 versus stage 3–4 survival comparison with borderlines excluded.

Table 2

Stage and mutation status of interval cancers

Taking the BRCA1/2 set as a separate entity, there was some evidence of down staging in cancers to 61% post prevalence (table 1), but overall survival was extremely poor and significantly worse than for the assumed non-BRCA cases, although the significance disappeared when borderline tumours were excluded (figure 3). Five year and 10 year survival in 49 BRCA1/2 mutation carriers was 58.6% (95% confidence interval (CI) 50.9% to 66.3%) and 36% (95% CI 27% to 45%), which was significantly worse than for 15 non-carriers (91.8%, 95% CI 84% to 99.6%, both 5 and 10 year survival p = 0.015) including the borderline tumours. However, the number of invasive cancers in individuals not known to harbour a BRCA1/2 or MMR mutation was only eight out of over 2500 (0.3%) women compared to 49/981 (5%) BRCA1/2 carriers. Nonetheless, many of the individuals not known to harbour a mutation have not been mutation screened and this may over-represent the difference in detection rates.

Figure 3

(A) Comparison of BRCA mutation carriers and non-carriers. (B) Carriers comparison with borderlines excluded.

It was not possible to identify a true prospective control group. Nonetheless, we have identified the stage of 111 ovarian cancers with BRCA1/2 mutations in the Manchester region since 1990 who had not undergone screening. Only 33/111 (30%) were at stage 1/2 but this is little different to the 29% from table 1 of those undergoing screening. However, post-prevalence the proportion at stage 1/2 did improve slightly to 39% but still not statistically better than in the unscreened women.


The current study has combined analysis from most of the previous reported studies on high risk ovarian screening4 5 6 7 8 9 with an additional 17 previously unreported cancers and represents by far the largest report of cancers detected in the context of high risk familial screening. The report includes the majority of published high risk screening for ovarian cancer. It is the first time that such studies have addressed survival rather than simply stage and morphology at diagnosis, and none have so far shown data on long term survival. Our study has shown that in an analysis of ovarian screening in five European centres, annual ovarian screening with transvaginal ultrasound and CA125 is ineffective in improving survival in BRCA1/2 carriers. Survival at 10 years was only 36% which is little different to that expected in an unscreened population. There is very little evidence of down staging, although 5 year survival figures are reasonably good even for BRCA1/2. The comparison of prevalent versus post-prevalent scans shows little advantage. Even though there could be a relative down staging of tumour mass this is not correlated with increased survival in the post-prevalent cases. The lack of actual down staging and almost identical survival curves suggests that the lead time for screening is <12 months. Although there is a suggestion for benefit in the non-BRCA group, 5/15 (33%) of these cancers were borderline where screening is unlikely to affect prognosis, which is extremely favourable. Even if screening is effective in the non-BRCA group, the incidence of ovarian cancer is likely to be extremely small and the economics outside proven HNPCC is likely to be marginal at best. Previous reports including our own3 4 5 6 7 16 17 18 19 20 have shown a high rate of interval cancers and that BRCA carriers in particular are detected at late stage. This may be due to the biology of the tumours spreading along the peritoneal surface rather than creating a mass in the ovary.3 The failure of annual CA125 to detect early stage disease is likely due to the rapid rise from normal of CA125 in 75% of serous tumours.21 This probably reflects a rapid transition through stage 1 to stage 3 in serous ovarian cancers and may explain the lack of lead time and difference in survival between prevalent and post-prevalent cancers. The HNPCC related cancers and other non- BRCA related cancers may be different in this respect. The absence of borderline tumours among the BRCA1/2 cancers and the presence of 5/15 (33%) among the non-BRCA cancers emphasises that these are a different entity and not part of the BRCA1/2 spectrum,22 23 and may not be part of the hereditary spectrum. Detecting borderline tumours almost certainly makes no long term difference to survival, as there is little evidence for progression of these tumours to invasive cancer. It is difficult to draw too many conclusions on the non-BRCA group once borderline tumours are removed as this is only based on 10 patients.

Any hope that an annual screen may at least mean a more curable “low” volume disease are pretty much excluded by our current study in BRCA carriers. If anything BRCA carriers are reported as having improved survival due to better responses to platinum based chemotherapy.24 25 Even if this is the case, current ovarian screening protocols are unlikely to make a difference for any but the occasional patient. While two large multicentre trials with new protocols are examining more frequent CA125 measurement and the assessment of other biomarkers, BRCA1/2 carriers should be counselled that at present there is no evidence of benefit for ovarian screening, and indeed evidence suggests that it is ineffective. Ovarian ultrasound in particular appears to offer little benefit for a labour intensive and expensive test. The failure of annual screening is almost certainly due to the inability to downstage ovarian cancer to stage 1 or 2 as seen from our control group and from fig 2. Women who have not completed their family or who decline bilateral salpingo-oophorectomy (BSPO) should be encouraged to join research trials addressing additional modalities rather than receive ad hoc annual surveillance that has no proven benefit. Thorough information on the limitation of screening must be given, as the mere offer of such screening may be considered by many women as proof of efficacy. It is likely to be difficult to deny women at very high risk some form of surveillance if no research trial is available. Therefore, until there is clear evidence of benefit from new protocols, BRCA1/2 carriers and those at risk of a known BRCA1/2 mutation should be advised of the clear proven efficacy of BSPO, including the reduction in breast cancer risk.26 These women should seriously consider this surgery after completion of their family and preferably before 40 years of age, although this could be delayed a little later in BRCA2 carriers.


We dedicate this paper to Andrew Shenton who died tragically young on 19 February 2008. We also acknowledge the NIHR Biomedical Research Centre at Central Manchester Foundation Trust for support.

Conception—Evans DG, Moller P; Data collection—Evans DG, Shenton A, Vasen HFA, Gaarenstroom KN, Stirling D, Moller P, Apold P, Dørum A, van Asperen C; Data analysis—Evans DG, Shenton A; Manuscript writing—All; Approval of final version—All.



  • Competing interests None declared.

  • Patient consent Obtained.

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

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.