toag.5.2.95

Familial screening for gynaecological and breast cancer Diana M Eccles

This article covers basic principles of cancer genetics including carcinogenesis and genetic susceptibility and taking and interpreting a family history of cancer. The main emphasis is on common situations where a family history of ovarian, breast or endometrial cancer has practical implications for patients and several illustrative examples of family pedigrees are used to help clarify key points. Genes that are known to confer a high risk of certain types of cancer are discussed.

Introduction to the basic principles of genetic susceptibility to cancer

Although all cancers arise following an accumulation of cellular DNA damage most cancer is not hereditary. Environmental risk factors for cancers differ depending on the primary site and they are well described. In the general population most cancers increase in fiequency with increasing age. Simplistically, ths is due to increasing time to acquire cellular genetic damage and a reduction in the efficiency of immune surveillance.

Population-based epidemiological studies have long recognised the association between a positive family history of a cancer and susceptibility to the same cancer.Around 10-25% of common cancers arise because of some sort of genetic predisposition, but less than 5% will arise as a result of a single highly penetrant cancer predisposition gene. Epidemiological studies often find that a family history of a given type of cancer raises an individual’s risk by two to three times.The problem with these statistics is they are not at all specific to the individual and reflect a much more complex situation.

Familial clustering may be due to:

chance shared environment

The human genome contains an estimated 30 000-100 000 genes, many of which have now been identified, although their function is often

low penetrance, high fiequency genes high penetrance low fiequency genes.

unknown. Many cancer genes have been identified by examining the molecular characteristics of cancer specimens with the multiplicity of techniques available to molecular biologists. In broad principle, genes found to be important in cancer development are usually involved to some extent in regulating cellular proliferation. Disruption of the normal functioning of a number of key genes will lead to the emergence of a malignant clone and if the clone manages to avoid the many check points that are in place to identify and destroy genetically abnormal cells, the cancer will become clinically detectable.

Summary

Cancer arises as a result of a stepwise process of accumulated DNA damage over time. Genes involved in the normal process of DNA replication and repair and cellular proliferation are those most often found to be damaged in cancers. A minority of common adult cancers arise on a background of genetic susceptibility. Environmental and genetic influences interact to alter susceptibility to many common diseases includmg cancer.

Hereditary or familial cancers

Less than 5% of common cancers are likely to be the result of high-risk genes (termed hereditary cancer predisposition). In this context high-risk genes are gene mutations, whch, if inherited, give rise to a high chance of certain cancers developing (hgh penetrance). However, where such high-risk genes are present within a family the effect can be devastating. It is this group of

0 2003 Royal College of Obstetricians and Gynaecologists

SCREENING REVIEW

The Obstetrician 81 Gynaecologist

2003;5:95-102

Keywords cancer, genetic testing, heredity, risk, treatment.

Author details

Diana M Eccles MD FRCP,

Consultant and Senior Lecturer in Cancer Genetics, Wessex Clinical Genetics Service, The Princess Anne Hospital, Coxford Road, Southampton SO16 SYA UK. email: delOroton.ac.uk

95

SCREENING REVIEW

The Obstetrician ' Gynaeco'ogist

2003;5:95-102

high-risk genes that will be the main subject of this article. In order to find which families are most likely to benefit fiom genetic investigations, molecular tests and preventive measures, it is important to understand the process of genetic assessment. It is a l~o important to recognise that different types of clinical intervention may be appropriate at Merent levels of risk.

Current understanding of cancer genes suggests that in addition to the smal l proportion of high- risk hereditary cancers there is a much larger proportion, perhaps one-quarter to one-third of all common adult cancers, that have some genetic component (termed familial cancer pre- disposition). Much genetic predisposition is likely to be due to common genetic variants within the population. Such variation may, for example, affect enzymatic metabolism of environmental carcinogens. On an individual basis the increased cancer risk is marginal, meaning that the associated excess risk is s m a l l

I+ 11:l

reeking advice I:1 I:2 90

ll:2 11:3 11:4 69 63

ovarian cancer age 65 years

I IV: 1 IV:2 IV:3

Figure 1. Population risk for ovarian cancer A single affected individual may be difficult to assess but if there are clearly several unaffected relatives on the same side of the family and the affected individual is not unusually young then there are no grounds to suspect any significant genetic component to the disease. Even for a single unusually young affected relative where the family history is clear on both sides it is extremely uncommon for this to be attributed to a high-risk susceptibility gene

Recommendation Referral to the NHS breast screening programme; if over 50 years of age and interested in ovarian screening trials enquire about UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS).aU

'

but in a large population study the effect may be measurable. Such genetic variants may be common in the population at large, but they are also inherited and wdl therefore be more likely to cluster in families. Predictive genetic testing for this type of susceptibility is unlikely to be clinically useful for the individual. Current interest has been directed in part towards elucidating the effect such genetic variation is having on the susceptibility of high-risk gene carriers to disease. In other words, these variants are interesting potential candidates for genetic modifiers of penetrance. In reality, full characterisation and cataloguing of all genetic susceptibility to cancers at various sites is a long way off.'

The clinical genetic assessment process

The clinical service

Most regional genetics centres provide a service for families with a history of cancer that seems to be in excess of what might be reasonably ascribed to chance.

Who should be referred?

Most centres provide guidelines for referral and will give advice by letter or telephone if the relevant information about family structure and dlagnoses can be provided. Cancer Research UK has produced information leaflets on familial breast and ovarian cancer.2 Although there are published guidelines with an explanation about how risk categories are derived, most centres prefer to use their own (broadly similar) guidelines that have usually been agreed with local service providers.' Because these are guidelines rather than rules interpretation will need to take account of the resources available. Therefore, local guidelines may be more or less stringent than published guidelines.'

Risk assessment

Many centres use a simple questionnaire to document family history in some detail before advising by letter (for lower-risk patients) or arranging a clinic review (for higher-risk patients). In the genetics clinic family trees are constructed as far as third-degree relatives (cousins, great aunts and uncles) and any key cancer diagnoses are verified as far as possible using death certificates or medical records. This can be time consuming and requires permission &om living relatives, so may not always be possible.

96

Information about treatment and outcome for a given family member can provide clues to the veracity of a proffered diagnosis. For example, benign ovarian tumours are usually not relevant to genetic predisposition, but may be perceived as malignant in the understanding of a relative. Family histories are not always what they seem at first telling and are dynamic and may evolve over time; this may cause problems in assessing the likely diagnosis and, therefore, the risk.

Family histories are assessed with all the available information, giving due weight to confirmed diagnoses and taking into account any clear doubts about diagnoses.An estimation is made of the likelihood of the reported, and ideally confirmed, diagnoses being due to chance (Figure l), familial clustering (Figure 2) or due to a hereditary cancer predisposition (Figures 3-6).

In many centres the cancer geneticist interacts with the multidisciplinary, site-specific cancer teams in discussions regarding clinical management in patients with a high genetic risk who have been diagnosed with cancer. For example, a breast cancer diagnosis in a premenopausal woman with a mother with breast cancer and a maternal aunt with ovarian cancer may carry a BRCA1 or a BRCA2 mutation. Scrutiny of the hstopathology of the tumours and the general structure of the pedigree can indicate the likely gene and give an idea of associated risks long before genetic testing can be completed. Such a patient might then be offered an oophorectomy as part of her breast cancer treatment rather than have a radiation induced menopause.

Summary

Take a three-generation family history when assessing a f a d y history of cancer.

Note ages at affection and type of cancer and age for unaffected relatives (especially intervening female relatives). Ask questions about treatment and outcome that may indicate the veracity of a reported cancer in a relative.

*

Key features suggesting a high-risk genetic suscep tibihty :

Multiple cases of the same or related cancers in more than one generation. Younger than average age at onset than in the general population.

Genetic testing - principles

At the present time molecular investigation for an underlying genetic predisposition is offered to f a d e s who fall into the hereditary category. Such investigation is only possible where an affected living relative is able and d n g to give a blood sample for analysis and where a specific hereditary syndrome, for which a gene has been identified, is suspected.

SCREENING REVIEW

The Obstetrician & Gynaecologist

2003;5:95-102

Genetic testing for cancer predisposition is not straightforward. A uselid analogy is to consider each gene as being simdar to a book in a large library. If on assessing the pedigree we suspect that there may be a ‘spehng mistake’ in one of those books (there are one or two for each ‘syndrome’) which might be implicated and of which we know the location, we can take these books off the shelf and go through them using various techniques. Because none of techniques is 100% sensitive, if the outcome of this search is negative we are left not knowing whether we

I 0 Ma’e I 0 Female

seeking advice

ll:2 11:3 11:4 61 65 75

ovarian cancer at 61 years

Ilk1 Ilk2 Ilk3 Ilk4 r\ 5a 55 50 45

Breast cancer aged 51 years

Figure 2. Moderate risk for ovarian/breast cancer Both the cancers have occurred at ages similar to what is expected in the general population. There are a number of older female relatives who could have inherited a genetic predisposition and if this were due to a high penetrance gene it would have been more likely to have led to a cancer than not. Genetically determined risk factors such as those determining natural oestrogen levels, age at menarche, etc. may cluster in families increasing susceptibility to hormonally influenced cancers by a small amount. These factors would be good candidates for moderate risk genes and a plausible explanation for this pattern of family history.

Recommendation Referral to the NHS breast screening programme starting at 50 years of age. Some centres may recommend earlier breast screening but many would not. The breast cancer would need to have been diagnosed before 50 years of age to fulfil the criteria for the UK Familial Ovarian Cancer Screening Study (UKFOCSS).’)

97

SCREENING REVIEW

11:l

The Obstetrician & Gynaecologist

ll:2 11:3 11:5 11:4

breast cancer age 42 ovarian cancer age 55

breast cancer age 48

2003; 5:95-102 1; I

reeking advice T breast cancer age 37

died age 41 age 33

Figure 3. High-risk breast and ovarian cancer This family history is typical of a BRCAl gene carrier family; note the early average age of breast cancers (42 years) and the dual primary breast and ovarian cancer with early onset of the first primary tumour in the mother.

Recommendation Refer to a clinical genetics centre for consideration of genetic testing (the aunt would need to be willing to be involved and to give a blood sample).

Refer for early mammography (from 35 years of age), recommend breast awareness and prompt investigation of symptoms.

Eligible for UKFOCSS from the age of 35 years and may wish to discuss prophylactic oophorectomy once child bearing is complete.

missed the relevant fault or whether the mutation is in a gene (book) that is somewhere else in the library, as yet unidentified.The weaker the family history criteria used for selection of samples for testing the more complicated the possibilities become, as lower penetrance familial genes may be a possible explanation. A further possibility is that the individual whose DNA was tested had a sporadic cancer and was never a gene carrier in the first place, even though a high-risk gene may be present among the other cancer-affected individuals in the family.

Sutnniary

Presymptomatic (predictive) genetic testing can be offered to family members once a specific gene fault has been identified that the geneticist is confident is causing the disease predisposition in the family.This involves three stages: first an assessment of the likely genetic diagnosis to direct the laboratory to the relevant gene(s), next a mutation search in a DNA sample taken &om a cancer-affected individual likely to be a gene carrier in the f d y and finally if a causative (pathogenic) mutation is found a simple single test for the presence of that mutation in a DNA

sample h m an unaffected individual wishing to clarify his or her future cancer risks.

Practical points to be considered before embarking on mutation searching:

the likely genetic diagnosis (risk assessment, clinical assessment) the scale of the proposed mutation search the cost, sensitivity and specificity of the proposed search interpretation of a negative or equivocal result.

Practical points to be considered before embarking on predictive genetic testing:

The patient must have full information about the implications of a positive test outcome including:

associated cancer risks

options for prevention The test may affect insurance. The test outcome will have implications for other family members’ risk and sharing of information with the wider family is likely to be important.

available options for early detection

98

0 Female

reeking advice

70 I ovarian cancer age 65 years colon cancer age 69 years

I r

111:l “r‘ 111:2

ovarian cancer age 46 years

IV:2 I V 3 IV4 I V 5 IV6 IV7

ovarian cancer age 26 years

Figure 4. Site-specific ovarian cancer This example is a family where confirmation of diagnosis was key to the final interpretation. The initial medical details were of three ovarian cancers in three generations at increasingly young ages. Over a period of three years and with considerable difficulty, the medical records for the three individuals with ovarian cancer were retrieved. Individual 11:4 and Ilk2 had epithelial ovarian carcinomas. Reviewins the available records for individual IF3 revealed that she initialiy had an ovarian tumour which was ckssified as a stage 1 endometrioid ovarian carcinoma but also was found to have hyperplasia of the endometrium with severe atypia on uterine curettage. She was treated with unilateral oophorectomy and progestogens to preserve fertility but presented a year later with a large fixed pelvic carcinoma from which she died. In the meantime both Ilk3 and Ilk4 had presented with rectal bleeding and been diagnosed with colorectal cancers aged 56 and 54 years respectively. Initial searches for BRCAt and BRCA2 mutations had been negative and in the light of these developments a diagnosis of hereditary non polyposis colorectal cancer (HNPCC) was suspected. This was confirmed when a truncating mutation was detected in the hMLHl gene.

Recommendation In this family genetic testing could eventually be offered to clarify the risk. Recommendations for gene carriers in HNPCC are given with Figure 6. Because ovarian cancer had occurred this family were eligible for the UKFOCSS trial and were offered annual measurement of serum CA125 levels and transvaginal ultrasound scanning as part of this study. Prophylactic oophorectomy and hysterectomy were also offered.

Specific cancer sites and high-risk qenes

Ovarian cancer

The lifetime risk of developing ovarian cancer is around one in 100 women and varies across geographic locations, even within the UK. As for other common adult tumours, most cases of ovarian cancer are not heredtary but arise as a result of multiple acquired genetic mutations withm an ovarian cell that leads to uncontrolled clonal proliferation (Figure 1 and Figure 2). Risk factors include nulliparity, previous diagnosis of breast cancer and a farmly history of ovarian cancer. Population-based studies, using segregation analysis, have identified an

inheritance pattern for ovarian cancer. The most likely scenario to account for this pattern is that rare genes are inherited as autosomal dominant traits. Carriers of these genes have a high risk of developing ovarian ~ancer .~

The genes BRCAl and BRCAZ are currently known to cause the highest genetic risks for ovarian cancer (Figure 3). The BRCA1 gene confers the hghest ovarian cancer risk, estimated at 40-60%, and the risk of ovarian cancer in BRCAZ gene carriers is around 15-20%.6 There is some evidence that the level of risk might vary accordmg to the position of the causative mutation in the gene.’ Although the majority of families with a dominant pattern of ovarian cancer and breast plus ovarian cancer are due to

SCREENING REVIEW

The 0 bstet r ician & Gynaecologist

2003;5:95-102

99

SCREENING REVIEW ~~~ ~~ ~

The Obstetrician & Gynaecologist

2003;5:95-102

Fl-? lung cancer

age 49 years

ll:1 11:2 k3 11:4 11:5

breast cancer brain tumour age 27 years age 5 years

3

age 3 years adrenocortical carcinoma

Figure 5. The Li-Fraumeni syndrome This typically involves childhood soft tissue sarcoma and early onset breast cancer. In this family the presenting individual was the child but the family history was already suggestive. Typical tumoun indude soft tissue sarcomas and adrenocortical carcinoma in the first decade, osteosarcoma in the second decade, breast, lung and brain tumours in the third and fourth decades. t he penetrance is very high in classical Li-Fraumeni syndrome families; in females it is virtually 100% by the age of 50 years, it is a little lower in males.

Recommendations Refer for expert advice from clinical cancer genetics service. Avoid screening tests that use X-irradiation whenever possible, magnetic resonance imaging may be better and is often more sensitive. There is no screening regimen that has been proven to be effective in this situation, many clinicians operate an open door policy for those at risk and rely on annual review plus prompt symptom assessment and targeted investigation.

BRCA 1 or BRCAZ, there are a small subset that may not be due to either gene and other high risk ovarian cancer genes are being sought (Figure 4). Borderline tumours and isolated early-onset ovarian cancers, which are more likely to be borderline and mucinous, are not generally associated with BRCAZ or BRCAZ gene carriers."

Having identified an increased genetic risk of ovarian cancer the options available to patients are relatively 1imited.They may involve entering screening trials using serum tumour markers and transvaginal ultrasound scanning. Current trials in the UK aim to determine whether the available methods for screening are effective or not.'O Prophylactic oophorectomy is an option often chosen by high-risk gene carriers and the reduction in risk is substantial, but not 100% as there is a small residual risk of primary peritoneal carcinoma. There is also an increased risk of cancer of the fdopian tube, so these must be removed with the ovaries." There are many

anecdotal cases of metastatic carcinoma after pmphylactic oophorectomy, but at least some of these are related to microscopic adenocarcinoma identified retrospectively in resected ovaries. Thorough histopathological assessment of ovaries after removal is imperative in high-risk patients.Lz Chemoprevention using the oral contraceptive pill may be effective in high-risk gene carriers, but there also appears to be an increased breast cancer risk associated with oral contraceptive pill use in BRCA1 gene carriers, so it is difiicult to actively recommend this as a therapeutic strategy.

Breast cancer

Breast cancer is more common in the general population than ovarian cancer, with a lifetime incidence of around one in 1 1 in the UK. Because of this high incidence, clustering of breast cancer cases in families may occur by chance, because of a low penetrance genetic susceptibility (Figure 2) or because of a high-risk gene mutation (Figure

100

I:1 I I:2

hysterectomy age 36 years rectal cancer age 52 years

ll:2 I 11:3 11:4 ll:1

endometrial cancer age 41 years colon cancer age 36 years adenoma of colon age 49 years I

0 Male

0 Female

patient A reeking advice

k 111:1 Ilk2 Ilk3

caecal cancer age 32 years

Figure 6. Typical family with hereditary non polyposis colorectal cancer (HNPCC) Note the key features: there is more than one generation affected, affected individuals are all directly related to each other, the average age for colon cancer is 40 years, endometrial cancer has occurred in some female family members.

Recommendation Refer to clinical genetics service for further investigation and genetic testing. Screening involves colonoscopy every two years from the age of 25 years increasing to annual screening a t 40 years of age for those carrying the gene, affected with cancer or at a 50% risk. Close investigation of any abnormal uterine bleeding with hysteroscopy and biopsy is usually recommended in the absence of any evidence of benefit from screening for endometrial cancer. A similar policy of watchfulness is practiced regarding other extracolonic malignancies in HNPCC, which include urothelial and hepatobiliary tumours.

3). As before the general principles apply, the greater the number of cases and the younger the average age at onset the more likely it is that the cancer is related to an underlying genetic predisposition. Again, BRCA 1 and BRCA2 are the two genes currently known to confer a high Metime risk of breast cancer. According to a currently used estimate, breast cancer occurs in 80% of BRCA1 and BRCA2 gene carriers. However, these risks may be exaggerated because of the method used to investigate the families h m which these risks are derived.')

Families with an inherited mutation in the TP53 gene are rare, but mutations in TP53 confer a high risk of breast cancer at a young age (typically in the third or fourth decade of life) and also up to a 20% risk of malignancy in childhood or adolescence, a pattern known as the Li-Fraumeni syndrome (Figure 5)." Only a small proportion of all the attributable genetic risk in a population is explained by detectable mutations in BRCA1, BRCA2, TP53 or other r m breast cancer genes.There are undoubtedly

further high-risk breast cancer genes to be disc~vered.'*'~

Although breast screening mammography is offered in many centres to women at increased genetic risk fiom about 35 years of age this may not be the ideal technique for imaging younger women with a high likelihood of developing high-grade tumours.'6 Screening of the breasts by magnetic resonance imaging is currently being tested in many centres in the UK and elsewhere, alongside conventional screening." Some women take up chemoprevention options, although this approach may be too little too late for those at very high risk.'* Some women chose to reduce their risk as much as possible by having a risk-reducing mastectomy and this seems to be a more popular choice in proven high-risk gene carriers with young ~hildren.'~

Endometrial cancer

Endometrial cancer occurs with similar fkequency and age distribution to ovarian cancer.

SCREENING REVIEW

The Obstetrician & Gynaecologist

2003;5:95-102

101

SCREENING REVIEW

The Obstetrician ’ Gynaeco’ogist

2003;5:95-1 O2

Risk factors include obesity and unopposed oestrogenic stimulation, such as with oestrogen- only hormone replacement therapy and with tamoxifen. Endometrial cancer is not a feature of BRCA1 or BRCAZ gene carriers, but is a prominent feature in hereditary non polyposis colorectal cancer (HNPCC). HNPCC arises b m inherited mutations in one of the mismatch repair gene family, commonly hMLHl or LMSHZ (Figure 4 and Figure 6).These genes are involved in maintaining the integrity of DNA copying fidelity. This deficiency leads to a rapid accumulation of defective, h s h mutations in dividing cells. The most common cancers resulting from this are cancer of the endometrium and colorectal cancers, 70% of which are proximal to the splenic flexure, unlike in sporadic colorectal cancer. A significant proportion, around 20%, of cases of endometrial cancer arise before 50 years of age and in some families this may be more common than colorectal cancer.”*’ There is no proven method of screening for endometrial cancer and where a high risk is identified some women opt for prophylactic hysterectomy. The ovarian cancer risk in HNPCC is increased and some studies have estimated a lifetime risk of 7% for ovarian

cancer.m Opinions are divided as to whether there is an increased risk of breast cancer, but it is udkely to be greatly increased and hormone replacement therapy is not contraindicated for a young HNPCC gene carrier electing to have prophylactic surgery.

Conclusion

The assessment of genetic risk requires a fidl family history, taking information about all unaffected individuals’ ages and all affected individuals’ ages at onset of cancer, treatment and outcomes. It is then possible to deduce whether a given family history may have arisen by chance or if it is more likely to reflect a significant genetic predisposition. Most regional cancer genetics services offer guidelines as to what type of family history they would be interested in assessing further. For women identified as being at a moderately increased risk genetic testing is unlikely be helpful. For those who appear high risk, and where there is a suitable blood sample available, genetic testing may help to clarify the risks for family members. In both cases options for surveillance

rn and prevention can be discussed.

References

1.

2.

3.

4.

5.

6.

7.

8.

Nathanson KL,Weber BL. “Other” breast cancer susceptibility genes: searching for more holy grail. Hum Mol Genet 2001;10715-20. Cancer Research UK. [www.cancemsearchuk.org/] Eccles DM, Evans DG, Mackay J. Guidelines for a genetic risk based approach to advising women with a family history of breast cancer. UK Cancer Family Study Group (UKCFSG). J Med Genet 2000,37203-9. Hurwin B. Legal and political considerations of clinical practice guidelines. BMJ 1999;318:661-4. Eccles DM, Forabosco F‘, Williams A, Dunn B, Williams C, Bishop DT, et a / . Segregation analysis of ovarian cancer using diathesis to include other cancers. Ann Hum Gene1 1997;61:243-52. Ford D, Easton DF, Stratton M, Narod S, Goldgar D, Devilee P, et a/. Genetic heterogeneity and penetrance analysis of the BRCA 1 and BRCAZ genes in breast cancer families. Am] Hum Genet 1998;62:676-89. Gayther SA, Mangion J, Russell PA, Seal S, Barfoot R, Ponder BAJ, et al. Variation of risks of breast and ovarian cancer associated with different germline mutations of the BRCAZ gene. Nar Genet 1997;15:103-5. Gotlieb WH, Friedman E, Bar-Sade RB, Kruglikova A, Hirsh-Yechezkel G, Modan B, et al. Rates of Jewish ancestral mutations in BRCA 1 and BRCAZ in borderline ovarian turnouts.] Natl Cancer Inrt 1998;90995-1OOO.

9. Stratton JF,Thompson D, Bobrow L, Dalal N, Gore M, Bishop DT, ef al. The genetic epidemiology of early-onset epithelial ovarian cancer: a population-based study. Am J Hum Genef 1999;65:1725-32.

10. Jacobs IJ. Skates SJ, MacDonald N, Menon U, Rosenthal AN, Davies M, et d. Screening for ovarian cancer: a pilot randomised controlled trial. Lanut 1999;353:1207-10.

Nedelcu R, McLaughlin J. ef a/. A genetic epidemiological study of carcinoma of the fallopian tube. Cynecol Oncol2001;80:341-5.

12. Lu KH. Garber JE, Crarner DW,WelchWR, Niloff J, Schrag D, ef al. Occult ovarian turnouts in women with BRCAl or BRCAZ mutations undergoing prophylactic oophorectomy] Chi Oncol2OOO;1%:2728-32.

13. Antoniou AC, Pharoah PD, McMdan G, Day NE. Stratton MR. Pet0 J. et al. A comprehensive model for familial breast cancer incorporating BRCAl. BRCA2 and other genes. Br] Cancer 2002;8676-83.

14. Varley JM, Evans DGR, Birch JM. Li-Fraumeni syndrome - a molecular and clinical review. Br J Cancer 1997;76: 1-1 4.

15. Easton D. Breast cancer genes - what are the real risks? Naf Genet 1997;16:21&1.

16. Brekelmans CT, Seynaeve C, Bartels CC, Tilanus-Linthorst MM, Meijers-Heijboer EJ, Crepin CM, er a/. Effectiveness of breast cancer surveillance in BRCA 1/2 gene mutation carriers and women with high tVnilial risk.

11. Aziz S, Kuperstein G, Rosen B, Cole D,

~~ _ _ _ _ _ _ _ _ _ ~ ~ ~

] CIin Oncol2001;19:924-30.

Bult P, Hendriks JH, et a/. Magnetic resonance imaging and manunopphy in women with a hereditary risk of breast cancer.] Nafl Cancer

17. Stougesdijk MJ, Boetes C, Jager GJ, Beex L,

rmt 2001;93:i095-102. 18. Wickerham DL,Tan-Chiu E. Breast cancer

chemopmntion: current status and future directions. k i n Oncol2001;28253-9.

19. MeijmHeijboer EJ,Verhoog LC. Brekelmans CTM, Seynaeve C,Tdanus-Linthont MMA, Wagner A, ef al. Presymptomatic DNA testing and prophylactic surgery in families with a BRCA1 or BRCAZ mutation. Lancet 2000,355:2015-20.

JH,Taal BG, Griffioen G, ef 01. Cancer risk in families with hereditary nonpolyposis colorectal cancer diagnosed by mutation analysis. Gashocnterology 1996;1101020-7.

21. Dunlop M, Farrington SM, Camthen AD, Wyllie AH, Sharp L, Burn J, et al. Cancer risk wociated with germline DNA mismatch repair gene mutations. Hum Mol C h e f 1997;6105-10.

Trials Network. [www.nwctognetwork.org.uk/list-of-network -members. htm].

23. National Cancer Research Network Trials Porrfolio. [www.ncrn.org.uk/Pordolio/&tabase.asp].

20. Vasen HFA, Wijnen JT, Menko FH, Kleibeuker

22. North West Obstetrics & Gynaecology Clinical

102