Public health aspects of breast cancer gene testing in Canada, part 1: risks and interventions

Headnote

Abstract

The risk (penetrance) of breast and ovarian cancer in carriers of the BRCA1 or BRCA2 genes are high, but it is likely that estimates based on selected large multicase families are inflated by selection bias. Estimates based on a population survey of Ashhenazi Jews are lower, but other population-based estimates are still not available. The proportion of breast or ovarian cancers related to the genes is similarly lower in population-based samples than in referred selected families, and, even for subjects with cancer onset at young ages or with a family history, it is quite small. Other genes with lower prevalence are also important, and there is evidence of some gene-environmental interactions. The management of female BRCA gene carriers includes intensive surveillance, prophylactic surgery and the use of tamoxifen. Apart from screening justified by randomized trials in the general community, such as mammography, recommendations for surveillance and prophylactic surgery are based only on expert opinion, and there has been little consideration of risk-benefit or cost-benefit comparisons. Tamoxifen reduced breast cancer incidence in one trial of high-risk women, but not in two other smaller trials, and the effect on mortality has not been determined. The limitations of genetic testing, and particularly of intervention strategies, deserve close scrutiny.

Key words: BRCA1; BRCA2; breast neoplasms; Canada; genes; genetic screening; mass screening; ovarian neoplasms; predictive value; risk

Introduction and Methods

This is the first of three papers that will address some key issues in regard to genetic testing for cancer susceptibility in Canada, from an epidemiologic and public health perspective.1,2 They will concentrate on the genes for breast and ovarian cancers, BRCA1 and BRCA2, as these illustrate most of the issues that will apply to similar new developments. Part 1 reviews background information on genetic risks and on interventions for test-positive subjects.

This work is based on an initial literature review, primarily using MEDLINE, discussions with many authorities in Canada in August-October 1997 and an update of the literature review to November 1998. Key words searched were breast neoplasms, ovarian neoplasms, BRCA1, BRCA2, genetic screening, mass screening, genetic counselling, predictive value, attitude to health, decision making, risk, genes and Canada.

Genes Conferring High Risks of Breast Cancer

Cancer develops as a result of accumulated damage to one or more genes, producing either somatic mutations within the cells that will form the cancer or germ line mutations that confer susceptibility by raising the probability of transformation to cancer cells and of progression. Two genetic loci conferring an increased risk of breast cancer have been identified and cloned, and several other genes relate to breast cancer risk. The work has been facilitated by the International Breast Cancer Linkage Consortium (IBCLC), which has collated data on over 200 high-risk families worldwide.

In 1990, the breast susceptibility gene BRCA1 was located on chromosome 1 7q in 23 families, with an average of six breast cancer cases per family; linkage was confined to those families with a mean age of onset of breast cancer of less than 46 years.3 This gene was found in almost all families showing both breast cancer and ovarian cancer,4,5 but only in about half of the families showing breast cancer alone.

The BRCA1 locus was cloned in 1994.6 Several hundred different mutations have been described so far. Most are frameshift, nonsense, splice or regulatory mutations that give a truncated protein product, so tests for these common types of mutations may detect around 80% of all gene mutations.7 The tests are most reliable where a specific mutation has been identified in a cancer case in the family, so that other family members are tested for that mutation. The most common mutations so far described are the 185delAG and 5382insC mutations. BRCA1 is likely to be a tumour suppressor gene;8,9 thus gene therapy, introducing the active wild type gene, could in principle restore tumour suppressor function.10

The second breast cancer susceptibility gene, BRCA2 on chromosome 13q, was identified in families with apparent dominant inheritance showing no linkage to BRCA1, and it was isolated in 1995. n So far, all identified mutations involve premature termination of protein synthesis.9 BRCA2 appears to be linked to an increase in male breast cancer, but not to a large excess of ovarian cancer. Studies of high-risk families suggest that BRCA2 is involved in 12% of families with four or five cases of breast cancer and in 61% with six or more cases.12 A specific mutation of BRCA2, 6174delT, is common in Ashkenazi Jews, occurring in about 1% of that population. Its penetrance may be less than that of BRCA1 185delAG, but together these two mutations may account for 25% of early onset breast cancer in Ashkenazi women.9

BRCA1 and BRCA2: Risk of Cancers in Carriers

Until recently, the most widely quoted estimates of cancer risks were based on the selected high-risk families involved in the IBCLC studies (Table 1). Ford et al.13 analyzed data from 33 families, each with at least four cases of breast or ovarian cancer diagnosed before age 60, and estimated risks from the occurrence of second cancers in subjects with breast cancer. The cumulative risk of breast cancer (the penetrance) in carriers by age 50 was estimated at 73%, and by age 70 at 87%, and the corresponding risks of ovarian cancer were 29% and 44%. There were also increases in risk of colon and prostate cancers in carriers. Substantially different estimates were derived from the same data by an independent method, by maximizing the LOD score over a range of possible penetrance functions;14 this produced lower estimated risks at young ages. A better fit to the data was given by assuming two different alleles (Table 1). The variability in these estimates, their quite wide confidence limits and, most importantly, the fact that these families are highly selected, all limit the general application of these estimates.

Other results from these studies are that the risk of a second, contralateral breast cancer in carriers who have one breast cancer is high (48% by age 50, 64% by age 74) and their risk of ovarian cancer is also greatly increased (29% by age 50, 44% by age 70). 13 The risk may vary with the precise mutation: for ovarian cancer, mutations in the 3' portion of the BRCA1 gene confer a lower risk than do other mutations.7 The BRCA1 gene does not appear to be associated with male breast cancer. As ovarian cancer is not easily diagnosed early and has a very poor survival rate, BRCA1 carriers may have a greater loss of life expectancy from ovarian cancer rather than from breast cancer.

For BRCA2, Easton et al.15 assessed risks in two large families, using a maximum LOD score method, estimating breast cancer risks of 60% by age 50 and 80% by age 70, and a risk of breast cancer in male carriers of 6% by age 70 (Table 1). They reported an increased risk of ovarian cancer based on three cases, and also excesses of laryngeal and prostate cancers.

Lower risks have been seen in population-based series. Struewing et al.16 assessed the prevalence of BRCA1 and BRCA2 mutations in a population-based sample of 5318 Ashkenazi Jewish subjects in Washington, DC, recruited through the media; 120 (2.3%) had a mutation. The risks of cancer, estimated by comparing the family history data for carriers and non-carriers, were similar in BRCA1 and BRCA2 carriers and were substantially less than the earlier estimates based on IBCLC families, reaching only 56% by age 70 (Table 1). No excess of colon cancer was seen. Thirty-one carriers (25%) had no cancer history in firstor second-degree relatives, which was not due simply to small family size. This sample, while community-based, was probably biased by the method of recruitment toward inclusion of subjects with strong family histories, and the reports of cancer in relatives were not verified. Even so, it suggests that the risks estimated from large multicase families referred to major research centres may be substantially higher than the risks in mutation carriers in general.

BRCA1 and BRCA2: Proportion of Cancers due to the Genes

Based on the IBCLC analyses, Ford et al.5 calculated a gene frequency of 0.0006 for BRCA1 . They estimated that around 5.3% of population cases of breast cancer occurring below the age of 40 are related to the BRCA1 gene, compared to 2.2% between ages 40 and 49, and 1.1% from age 50 to 70; the estimates for ovarian cancer were 5.7%, 4.6% and 2.1% respectively. Overall, about 1.7% of all breast cancers and 2.8% of ovarian cancers at all ages up to 70 occur in BRCA1 carriers. The IBCLC studies suggest that in 50% of families having at least four affected members with breast cancer under age 60 and in over 80% of families with both breast cancer and ovarian cancer, the cancers are due to BRCA1.17 These estimates are lower than earlier estimates from fewer families.18,19

More direct estimates come from testing series of patients for gene mutations (Table 2). Several series are based on subjects referred to genetic clinics. Couch et al.20 reported on women referred to assessment centres in the eastern US who were not members of known large multicase families. BRCA1 mutations were found in 16% of 169 women who attended because of a family history and in 13% of 94 women who had breast cancer diagnosed before age 40. The authors evaluated factors related to BRCA1 carrier status, providing a table of estimated probabilities. The prevalence of BRCA1 mutations increased with a family history of both breast and ovarian cancers, lower age at diagnosis and Ashkenazi Jewish ancestry. It did not vary with the occurrence of bilateral breast cancer or with the number of breast cancers in the family. The results suggest that gene testing in women who have even several relatives with breast cancer would have a low yield (<10%) unless at least one relative has had breast cancer before age 35.

In a recent international study of 798 referred women from suspected high-risk families, Shattuck-Eidens et al. reported that deleterious mutations were found in 12.8% of these women, using complete sequence analysis, and 50 new genetic alterations were found.21 Factors predicting BRCA1 mutations included the number of relatives with breast or ovarian cancer.

Results from population-based series are of particular value and tend to give lower frequencies of gene disorders. In a study of a hospital series of 73 women with breast cancer diagnosed before age 32 in the Boston area, Krainer et al.22 found that only 2 (2.7%) had BRCA2 mutations, whereas 9 (12%) had BRCA1 mutations. If these two mutations have similar population prevalences (as they do in Ashkenazi Jews), these results show a lower penetrance for BRCA2 than for BRCA1. This is consistent with other literature showing BRCA2 mutations in only 13% of 75 breast cancer kindreds where BRCA1 linkage had been excluded.12·23

Längsten et al.24 studied 80 women identified through a population-based cancer registry in Washington (state), who had breast cancer before age 35. Using DNA sequencing, they found that six women (8%) had BRCA1 mutations, of whom two had no family history of breast or ovarian cancer. Another four women had rare sequence variants of unknown importance. A further study based on the same registries by Malone et al. showed 6% BRCA1 mutations in 193 breast cancer cases under age 35, and 7% in 208 cases with breast cancer before age 45 with a first-degree relative also affected. In the study of Ashkenazi Jews noted earlier, Struewing et al.16 found a prevalence of 14% BRCA1 or BRCA2 mutations in subjects with breast or ovarian cancer prior to age 50 (Table 2).

Frequency of Carriers in the Population

Ford et al.5 estimated the frequency of the BRCA1 gene by assuming that the excess risks of ovarian cancer in first-degree relatives of breast cancer patients, and of breast cancer in first-degree relatives of ovarian cancer patients, were all due to BRCA1 : this gave a gene frequency (f) of 0.0006, with 95% confidence limits of 0.0002 and 0.001. The prevalence of BRCA1 mutations in the general population is then 2f-f2, or very nearly 2f = 0.0012, that is, 1.2 per 1000 population. The prevalence seems similar in the British and US studies.2 However, in the study of Ashkenazi Jews noted earlier, the prevalence of BRCA1 or BRCA2 mutations was 2.3%.16 The prevalence of one particular mutation, 185deLAG, was 0.9 % in an Ashkenazi Jewish population in the US.16,27 Population-based screening programs have been developed in this community, despite some reservations.27

This high frequency may be due to a founder effect;9 that is, most of the community may be descended from a few ancestors who had a high frequency of the gene. Extensive studies of BRCA1 and BRCA2 mutations in different populations have now been published. They show considerable variations in the proportion of high-risk families with BRCA1 mutations and in the frequency of particular mutations in European groups, due to founder effects;28 although some of the variations seen may be due to selection factors related to referral and to sampling variation.

Data on the many ethnically diverse communities within Canada are therefore important. Some data on French-Canadian families have been presented.12,29 In recent abstracts, Wong et al.30 report that 10.3% of 117 Ashkenazi Jewish women with breast cancer before age 65 seen in Montreal and Toronto had BRCA1 mutations. Tonin et al.31 assessed 94 French-Canadian families with breast cancer in women under age 65, ovarian cancer or male breast cancer; mutations have been identified so far in 34 families (36%), 19 with BRCA1 and 15 with BRCA2 mutations. Seventy-nine percent of the families with mutations had four or more cases of breast or ovarian type cancer.

Looking at other ethnic groups, BRCA1 mutations may be rarer in Japanese families.32 In Iceland, most families with combined breast and ovarian cancers are linked to BRCA2 and to a specific mutation, 999del5;33'34 these families may show an excess of pancreatic cancer.

Other Genes Conferring Increased Risk of Breast Cancer

Several studies have found that about 20% of multicase high-risk families show no linkage to either BRCA1 or BRCA2 after extensive testing/·35 This could be due to as yet unidentified mutations of these genes, but it is likely that other genes (BRCA3, etc.) will be identified in some of these families. Another dominant gene condition with an increased risk of breast cancer is the Li-Fraumeni syndrome, related to the suppressor gene p53, which may account for up to 1% of breast cancer cases occurring under the age of 35.9'36 Other rarer conditions include Cowden disease, an androgen receptor mutation that has been linked to male breast cancer, and genes associated with HNPCC (hereditary non-polyposis colorectal cancer).9

Several other genes are reasonably common, but have low penetrance, that is, the risk of breast or other cancers is only moderately increased. Ataxia telangiectasia heterozygotes have a moderately increased risk of breast cancer (relative risk of 3.9 in one overview37), conferring a cumulative risk of some 20-30% by age 70. Since between 0.5% and 1% of the population may have this genetic risk factor, it may account for roughly 2-8% of breast cancer in the population. There is some evidence that individuals with p53 and AT mutations may be more sensitive to ionizing radiation, and the breast cancer risks in mothers of AT children (who are heterozygous for the gene) are increased by radiation exposure.3 '3

The HRAS1 gene carries an increased risk of breast cancer of about twice the normal risk; since the carrier frequency is between 5% and 20%, about 3-8% of all breast cancer could be attributed to HRAS1. This association may be greater for black women and for women with estrogen negative tumours.40 Carcinogen-metabolism loci encode enzymes involved in the metabolism of tobacco, alcohol, occupational solvents and dietary constituents; elevated risks of breast cancer have been reported in association with specific genotypes.40

Interactive Effects of Genetic and Environmental Factors

In a case-control study of breast cancer among subjects known to be BRCA1 or BRCA2 carriers, smoking was negatively associated with breast cancer risk, with an odds ratio of 0.46 (95% confidence interval = 0.27-0.80) for subjects who had more than four pack-years of smoking. x The authors of this study, co-ordinated from Toronto, suggest that the anti-estrogenic effect of tobacco may be responsible. Previously, smoking in the general population appeared to be a risk factor for postmenopausal breast cancer among slow acetylators but was protective among rapid acetylators, the difference relating to the JV-acetyltransferase 2 polymorphism.42

In a case-control study based on the Nurses' Health Study in the US, smoking was related to an increase in breast cancer risk among subjects with genotypic variants of cytochrome P450 1A1, which affect aryl hydrocarbon hydroxylase activity, suggesting a causal effect of smoking in a genetically susceptible population.43

Validity of Laboratory Tests for Genetic Susceptibility

As with any other test, its ability to correctly identify subjects with a genetic susceptibility (the sensitivity) and the risks of the test giving a positive result when a true genetic susceptibility does not exist (the false positive rate, or 1 -specificity) are important parameters. They cannot be measured directly, as all subjects with a true genetic susceptibility cannot be identified, but may be inferred from linkage studies. Simpler tests suitable for widespread use can be compared with more extensive and detailed tests (such as DNA sequencing) that may only be applicable in research situations.

For BRCA1, many specific mutations have been shown;44 only detailed and expensive testing would detect all of these, and in a newly investigated family, there is a high likelihood of finding a previously unrecognized mutation. Families have been found with very strong evidence of linkage to BRCA1 but without an identified mutation, suggesting that further mutations are yet to be found.9 So the sensitivity of BRCA1 testing is clearly less than perfect. A negative test result is clinically useful in a family in which a specific BRCA1 mutation has been identified, so that failure to find that mutation in an individual can be taken as a valid negative result. Population-based screening is only feasible for specific mutations, for example, the 185delAG mutation in BRCA1 common in Ashkenazi Jews.27

A negative test for a gene such as BRCA1 in a subject from a high-risk family could be due to several possible factors: a mutation was missed because the test used did not detect that specific mutation; another known gene was responsible; another, as yet unidentified, gene was responsible; the familial risk was due to multiple, low penetrance genes; the "familial" aggregation was due to shared environmental risk factors, or to chance; or this was a sporadic case, despite being from a high-risk family. A false positive genetic test can also occur, since sophisticated testing may demonstrate mutations (i.e. differences from the normal gene sequence) that may not be related to increased risk and may not be informative.

A discussion of the different types of genetic tests that can be used is outside the scope of this review. Much of the literature is based on high-quality studies in research centres, using optimum methods such as confirmation sensitive gel electrophoresis and DNA sequencing. Simpler and cheaper tests, such as the protein truncation test, will detect the great majority of mutations, but the risk of a false negative result must be considered when the tests are interpreted.

Interventions for Subjects at High Risk

The assumption (often implicit) behind screening for genetic cancer susceptibility is that, in those individuals identified as mutation carriers, the cancers anticipated can be prevented, or their morbidity and mortality risk can be reduced by earlier diagnosis from surveillance. As in all issues of screening for chronic disease, there is a need for high-quality objective evidence to show the net benefit of the screening modality; this is best founded on population-based randomized trials, since observational studies are open to various biases. There are no randomized trials assessing the value of screening and intervention specifically for people tested for genetic markers. Whether the results of trials in other groups, such as the general population, can be applied without modification to mutation carriers is debatable. Trials, observational studies, and routine data collection and monitoring of the results of interventions in mutation carriers are of major priority.

Clinical Management of Carriers ofBRCAl and BRCA2: Breast Cancer Surveillance

The options available for clinical management include screening, prophylactic surgery and chemoprevention. Recently (1997) the Cancer Genetics Studies Consortium (CGSC), a US-based multidisciplinary group, published consensus statements on the management of BRCA mutation carriers.45 The CGSC recommends frequent use of screening procedures, starting at early ages but based on only grade 3 evidence (expert opinion only) [Table 3]. Since there are no available randomized trials of gene carriers or other high-risk groups, these recommendations are based on extrapolation from general population results, carrying the implicit assumption that a lower level of evidence of benefit is acceptable for subjects at higher risk.

For example, breast self-examination (BSE) is recommended by the CGSC despite the fact that the evidence for benefit is based on studies open to selection biases; the results from a non-randomized trial in Britain and from randomized trials in Russia and China all suggest no benefit.46-48 The apparent benefit seen in observational studies is likely due to other characteristics of those who practice BSE, rather than to the examination itself, and any benefit is likely to be lowest in women involved in other intense monitoring.49'50 Nevertheless, in a case-control study within the randomized Canadian National Breast Screening Study, there was a strong association between the risk of fatal or metastatic breast cancer and the non-performance of particular components of recommended self-examination procedures.51

The specific contribution of clinical breast examination is also difficult to assess because the quality of clinical examination is likely to vary greatly. Data from four randomized trials that compared clinical examination plus mammography with mammography alone showed improved cancer detection rates and sensitivity with the use of both methods, although mammography performed better than clinical examination alone.52

Mammographie screening for breast cancer is the best justified intervention proposed for carriers of BRCA genes, as several randomized trials have reported reduced mortality in women in the general population over age 50.53 However, there is considerable controversy about the value of mammography in women under age 50.53'54 At younger ages, the sensitivity is lower, with mortality benefits appearing later and being smaller. Although some of the most recent analyses do show a useful mortality reduction even at younger ages,55'56 much of this may be due to screening after age 50.5 '57 There is no evidence on the effects of mammography in women under age 40. Women with a genetic predisposition may be more sensitive to radiation from mammography, although any effect is unlikely to be substantial.38

Outcome of Mammographie Screening in High-risk Women

The effects of family history of breast cancer on the results of mammography screening were assessed for 31,814 asymptomatic women aged 30 or older who had a first screening examination between 1985 and 1992 in San Francisco.59 In women under 50, the false positive rate was increased by about 1 0% in those with a positive family history, while the cancer detection rate was increased more, giving screening a higher predictive value for those with a family history (8.3%) than for those without (3.9%). For women over 50, the predictive value of screening was 19.6% for those with a family history and 13.5% for those without. The sensitivity of mammographie screening, estimated by documenting interval cancers in a 13 -month period after a negative screen, was significantly lower for women under 50 with a family history (69%) than for those with no family history (88%). This effect was not related to breast density. The authors concluded that it was due to a faster tumour growth in women with a family history of breast cancer, and they recommended annual screening for such women. Family history had no effect for women over age 50, where the sensitivity was 95% for those with a family history and 93% for those without.

Surveillance for Other Cancers in BRCA Carriers

In the case of ovarian cancer, there is no good evidence for the benefits of any screening technique, although trials are in progress in the general population.61-63 The CGSC recommends the use of ultrasound with colour Doppler and CA 125 annually or semi-annually beginning at age 25 or 35, again based only on expert opinion.

BRCA carriers appear to be at moderately increased risks of prostate and colon cancers (see above). For prostate cancer, the CGSC recommends only providing information regarding options, given the lack of evidence for any benefit from prostate screening.64-66 The recommendation for colon cancer is the same as for the general population:67 the use of annual fecal occult blood testing and flexible sigmoidoscopy every 3-5 years beginning at age 50, which is supported by evidence from randomized trials68-70 and case-control studies.71'72

Prophylactic Surgery

For subjects with very high cancer risks, prophylactic surgery will remove the target organ to prevent the cancer from arising. Such surgery is as extensive (or more so) than that used to treat a cancer if it does occur, and it will be performed many years earlier. Because many mutation carriers will escape disease or die from other causes, a proportion of any prophylactic surgery done will be unnecessary. Moreover, since all the relevant tissue cannot be removed and since mutation carriers may have increased risks of several cancers, the protection given will be only partial.

The CGSC makes no recommendation for or against prophylactic mastectomy or ovarian removal. Breast cancer has been documented in women after prophylactic mastectomy, as residual breast tissue remains and breast tissue also occurs at other sites.73-75

A National Institutes of Health (NIH) consensus conference63 recommended that women with two or more first-degree relatives with ovarian cancer be offered ovarian removal after completion of child-bearing or at age 35. Ovarian removal will protect against ovarian cancer, and it also may decrease breast cancer risk. However, these benefits need to be compared with the problems of early surgical menopause, the likely increased risks of osteoporosis and cardiovascular disease and the effects of any replacement hormones used.

Schrag et al.76 have reported a Markov chain-based decision analysis of prophylactic mastectomy and oophorectomy for mutation carriers. They used three levels of cancer risk (penetrance): the highest were the risks based on the multicase family linkage analyses,5'14 and the others were the estimated risk and the lower 95% confidence limit reported by Struewing et al.16 The Schräg analysis assumed an 85% reduction in breast cancer risk after prophylactic mastectomy and a 50% reduction in ovarian cancer risk after oophorectomy. No effect on non-cancer outcomes such as heart disease was assumed, nor any effect of hormonal replacement therapy.

The main results of the Schräg analysis76 are that the gain in life expectancy from prophylactic mastectomy in mutation carriers is substantial if performed before age 40 and that oophorectomy can be delayed until this age without loss of benefit. The estimated gains for a 35 -year-old mutation carrier are similar to the benefits from successful reduction of high cholesterol levels and are greater than smoking cessation or the benefits from using adjuvant chemotherapy after breast cancer diagnosis. However, this decision analysis did not assess quality of life. In a decision analysis incorporating utility measures for quality-adjusted life years, using time trade-off methods, prophylactic surgery was shown to be cost-effective in comparison with surveillance for years of life saved, but not for quality-adjusted years.

Chemoprevention: Tamoxifen

Another option for high-risk women is to prevent breast cancer by hormonal or other interventions. The anti-estrogen tamoxifen appeared to reduce breast cancer risk by 49% in a randomized trial with 69 months' follow-up of women at high risk of breast cancer, based on age over 60, age 35-59 with high risk on the Gail et al. predictive model78 or a history of lobular carcinoma in situ. All the reduction was in estrogen receptor-positive tumours. These women showed no change in ischemic heart disease incidence, but had an increased incidence of endometrial cancer. 9 This study, the US National Surgical Adjuvant Breast and Bowel Project (NSABBP), was therefore closed early in April 1998, so its ability to assess tamoxifen's affect on mortality has been lost.

However, neither a British nor an Italian trial of tamoxifen in high-risk women showed any reduction in breast cancer incidence.80,81 Since these European studies were smaller and the Italian study had limited compliance, longer follow-up for mortality is needed.82 Tamoxifen has been approved by the US Food and Drug Administration (FDA) for reduction of short-term incidence of breast cancer in women at increased risk, but it has not been supported for a wider marketing for breast cancer prevention.83 Although a range of other chemopreventive agents have been considered for breast cancer and other cancers, none has reached a similar stage of having randomized trial results available.84

Cancer Prognosis in Mutation Carriers

A Seattle study reported better survival in breast cancer patients aged 21^5 with a first-degree family history compared with women without a family history, after adjustment for disease stage, mammogram history and other major confounders.85 In a small study in Scotland, 35 breast cancer patients with BRCA1 experienced better-than-expected survival.86 However, other studies have recorded similar or worse prognosis in mutation carriers.34'87 A Canadian study of 1 17 Ashkenazi Jewish women with breast cancer under age 65 (of whom 10% had BRCA1 mutations)30 and a French series of breast cancer patients under age 36 (of whom 15 were BRCA1 positive)88 both showed a worse prognosis in the BRCA1 carriers. The pathology of breast cancer, assessed in a large series by the IBCLC, differs among BRCA1 and BRCA2 carriers in several respects,89 although it is not clear what overall prognostic effect the differences would have. In a small series of Ashkenazi Jewish women, the BRCA1 carriers were less often estrogen receptor-positive and had a higher nuclear grade, implying a worse prognosis.90

Discussion

The risks of cancer in gene mutation carriers are substantially lower in more recent, population-based studies than in the older studies based on selected multicase families. The frequency of mutations in series of cancer patients is also lower in more recent, population-based studies. There is a need for further empirical information on risks in carriers that will avoid the biases in the research based on selected multicase families. More information is also needed on the results of testing as the amount of testing increases and the referral criteria for testing change. The technical aspects of genetic testing develop rapidly. Therefore, the sensitivity of tests suitable for routine use, compared with optimal procedures, must be assessed in routine practice. In addition, the impact of tests for newly recognized mutations in known genes and for newly discovered genetic markers requires continual review.

The guidelines for the clinical management of mutation carriers are based largely on older estimates of risk, which are likely to be too high. None of the interventions recommended for mutation carriers is supported by randomized trial evidence, which is often regarded as essential, or at least highly desirable, to support interventions for the general population or for individual therapy.91 Indeed, for several of the CGSC recommendations, the best evidence suggests that a net benefit for general population groups is unlikely. The CGSC recommendations are made on a "best case" argument; it is implied that, although there is only weak evidence for benefit, the use of all these modalities is justified for high-risk women. If a screening modality is effective, its net benefit-to-harm ratio will be greater for high-risk subjects. However, if it is not effective, the risks of both false positive and false negative results may be greater for high-risk subjects than for the general population.

There is no discussion in the CGSC report of the high probability of false positives from using all these screening methods simultaneously or of the potential sequelae in both physical and psychological terms. It may be that the anxiety created by knowledge of the high-risk state is so great that excess examinations and false positives will not increase it, but further assessment of those issues is warranted. Objective consideration of any recommendations is essential, especially for adequate informed consent. The uncertainty of the benefits of surveillance appears to receive relatively little attention in the literature on genetic screening. Current guidelines for management pay little attention to cost and cost-benefit assessments or to quality-of-life issues.

Acknowledgements

This paper is based on a report prepared initially for the Cancer Bureau, Laboratory Centre for Disease Control, Health Canada, under contract 502-8082 and project 502-0205.

Thanks are expressed to Don Wigle, Christina Mills and Janet Beauvais of the Laboratory Centre for Disease Control; Ian McDowell, Fay Draper and Mariella Pica of the University of Ottawa; Richard Gallagher, Ivo Olivotto, Karen Panabaker and Mary McCullum of the British Columbia Cancer Agency; Jean-François Boi vin and Renaldo Battista of McGill University; Nancy Quattrocchi and Hussein Noorani of the Canadian Coordinating Office for Health Technology Assessment; and Steven Narod of the University of Toronto, for useful discussions and for practical assistance in preparing the report, and also to several reviewers.

References

References

1 . Elwood JM. Public health aspects of breast cancer gene testing in Canada. Part 2: Selection for and effects of testing. Chronic Dis Can 1999;20(1): 14-20.

2. Elwood JM. Public health aspects of breast cancer gene testing in Canada. Part 3: A model of potential need and demand. Chronic Dis Can 1999;20(l):21-5.

3 . Hall JM, Lee MK, Morrow J, et al. Linkage analysis of early onset breast cancer to chromosome 17q21. Science 1990;250:1684-9.

4. Narod SA, Feunteun J, Lynch HT, et al. Familial breast-ovarian cancer locus on chromosome 17ql2-q23. Lancet 1991;338:82-3.

5. Ford D, Easton DF, Peto J. Estimates of the gene frequency of BRCA1 and its contribution to breast and ovarian cancer incidence. Am J Hum Genet 1995;57:1457-62.

6. Miki Y, Swensen J, Shattuck-Eidens D, et al. A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science 1994;266:66-71.

7. Cannon- Albright LA, Skolnick MH. The genetics of familial breast cancer. Semin Oncol 1996;23:1-5.

8 . Holt JT, Thompson ME, Szabo C, et al. Growth retardation and tumour inhibition by BRCA1. Nat Genet 1996;12:298-302.

9. Greene MH. Genetics of breast cancer. Mayo Clin Proc 1997;72:54-65.

10. Holt JT. Breast cancer genes: therapeutic strategies. Ann ? YAcadSci\991;m2>-3A-4\.

11. Wooster R, Bignell G, Lancaster J, et al. Identification of the breast cancer susceptibility gene BRCA2. Nature 1995;378:789-92.

12. Phelan CM, Lancaster JM, Tonin P, et al. Mutation analysis of the BRCA2 gene in 49 site-specific breast cancer families. Nat Genet 1996;13:120-2.

13. Ford D, Easton DF, Bishop DT, Narod SA, Golgar DE, Breast Cancer Linkage Consortium. Risks of cancer in BRCA1 -mutation carriers. Lancet 1994;343:692-5.

14. Easton DF, Ford D, Bishop DT, Breast Cancer Linkage Consortium. Breast and ovarian cancer incidence in BRCA-1 mutation carriers. Am J Hum Genet 1995;56:265-71.

15. Easton DF, Steele L, Fields P, et al. Cancer risks in two large breast cancer families linked to BRCA2 on chromosome I3ql2-13. Am J Hum Genet 1997;61:120-8.

16. Struewing JP, Hartge P, Wacholder S, et al. The risk of cancer associated with specific mutations of BRCA1 and BRCA2 among Ashkenazi Jews. ? EnglJ Med 1997;336:1401-8.

17. Easton DF, Bishop DT, Ford D, Crockford GP, Breast Cancer Linkage Consortium. Genetic linkage analysis in familial breast and ovarian cancer: results from 214 femilies. Am J Hum Genet 1993;52:678-701.

18. Easton D, Ford D, Peto J. Inherited susceptibility to breast cancer. Cancer Surv 1993;18:95-113.

19. Claus EB, Risch N, Thompson WD. Genetic analysis of breast cancer in the Cancer and Steroid Hormone Study. Am J Hum Genet 1991;48:232-41.

20. Couch F J, DeShano ML, Blackwood MA, et al. BRCAl mutations in women attending clinics that evaluate the risk of breast cancer. ? EnglJ Med 1997;336:1409-15.

21. Shattuck-Evans D, Oliphant A, McClure M, et al. BRCAl sequence analysis in women at high risk for susceptibility mutations: risk factor analysis and implications for genetic testing. JAMA 1 997;278: 1242-50.

22. Krainer M, Silva-Arrieta S, FitzGerald MG, et al. Differential contributions of BRCAl andBRCA2 to early-onset breast cancer. ? Engl J Med 1997;336:1416-21.

23. Couch FJ, Farid LM, DeShano ML, et al. BRCA2 germline mutations in male breast cancer cases and breast cancer families. Nat Genet 1996;13:123-5.

24. Langston AA, Malone KE, Thompson JD, Daling JR, Ostrander EA. BRCAl mutations in a population-based sample of young women with breast cancer. ? Engl J Med 1996;334:137-42.

25. Malone KE, Daling JR, Thompson JD, O'Brien CA, Francisco LV, Ostrander EA. BRCAl mutations and breast cancer in the general population. JAMA 1998;279:922-9.

26. Eeles R. Testing for the breast cancer predisposition gene, BRCA1. Documenting the outcome in gene carriers is essential. BMJ 1996;313:572-3 .

27. Bryant H. Genetic screening for breast cancer in Ashkenazi women. Lancet 1996;347: 1638-9.

28. Szabo CI, King M. Population genetics of BRCA1 and BRCA2. Am J Hum Genet 1997;60:1013-20.

29. Simard J, Tonin P, Durocher F, et al. Common origins of BRCA1 mutations in Canadian breast and ovarian cancer families. Nat Genet 1994;8:392-8.

30. Wong N, Brunet J, Narod SA, et al. Germ-line BRCAl mutation is an adverse prognostic factor in Ashkenazi Jewish women with breast cancer [abstract]. Terry Fox Workshop on Cancer Genetics (Canadian Collaborative Group for Cancer Genetics); 1997 May 31-Jun 1; Toronto (Ont).

31. Tonin PN, Mes-Masson A, Provencher D. Mutation analysis of BRCAl and BRCA2 hereditary breast-ovarian cancer families of French-Canadian descent [abstract]. Terry Fox Workshop on Cancer Genetics (Canadian Collaborative Group for Cancer Genetics); May 31Jun 1; Toronto (Ont).

32. Inoue R, Fukutomi T, Ushijima T, et al. Germline mutation of BRCAl in Japanese breast cancer families. Cancer Res 1995;55:3521-4.

33. Thorlacius S, Olafsdottir G, Tryggvadottir L, et al. A single BRCA2 mutation in male and female breast cancer families from Iceland with varied cancer phenotypes. Nat Genet 1996;13: 117-9.

34. Foulkes WD. BRCAl and BRCA2: penetrating the clinical arena. Lancet 1998;352: 1325-6.

35. Serova OM, Mazoyer S, Puget N, et al. Mutations in BRCAl and BRCA2 in breast cancer families: are there more breast cancer susceptibility genes? Am J Hum Genet 1997;60:486-95.

36. Malkin D, Li FP, Strong LC. Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms. Science 1997;250:1233-8.

37. Easton DF. Cancer risks in A-T heterozygotes. Int J Radiât Biol 1994;55:S 1 77-82.

38. Swift M, Reitnauer PJ, Morrell D, Chase CL. Breast and other cancers in families with ataxia-telangiectasia. ? Engl J Med 1987;316:1289-94.

39. Swift M, Morrell D, Cromarne E, et al. Incidence of cancer in 161 families affected by ataxia-telangiectasia. ? Engl J Med 1991;325:1831-6.

40. Newman B, Millikan RC, King M. Genetic epidemiology of breast and ovarian cancers. Epidemiol Rev 1997;19:69-79.

41. Brunet J, Ghadirian P, Rebbeck TR, et al. Effect of smoking on breast cancer in carriers of mutant BRCAl or BRCA2 genes. J Natl Cancer Inst 1998;90:761-6.

42. Ambrosone CB, Freudenheim L, Graham S, et al. Cigarette smoking, N-acetyltransferase 2 genetic polymorphisms, and breast cancer risk. JAMA 1996;276:1494-501.

43. Ishibe N, Hankinson SE, Colditz GA, et al. Cigarette smoking, cytochrome P450 1A1 polymorphisms, and breast cancer risk in the Nurses' Health Study. Cancer Res 1998;58:667-71.

44. Couch FJ, Weber BL. Mutations and polymorphisms in the familial early-onset breast cancer (BRCAl) gene. Hum Mutat 1996;8:8-18.

45. Burke W, Daly M, Garber J, et al., for the Cancer Genetics Studies Consortium. Recommendations for follow-up care of individuals with an inherited predisposition to cancer. II. BRCAl andBRCA2. JAMA 1997;277:997-1003.

46. Austoker J. Screening and self examination for breast cancer. BMJ 1994;309:168-74.

47. Semiglazov VF, Moiseyenko VM, Bavli JL, et al. The role of breast self-examination in early breast cancer detection (results of the 5-years USSR/WHO randomized study in Leningrad). Eur J Epidemiol 1992;8:498-502.

48. Baines CJ. Reflections on breast self-examination. J Natl Cancer Inst 1997;89:339-40.

49. Hill D, White V, Jolley D, Mapperson ?. Self examination of the breast: is it beneficial? Meta-analysis of studies investigating breast self-examination and extent of disease in patients with breast cancer. BMJ 1988;297:271-5.

50. Cole P, Austin H. Breast self-examination: an adjuvant to early cancer detection. Am J Public Health 1981;71:572-4.

51. Harvey BJ, Miller AB, Baines CJ, Corey PN. Effect of breast self-examination techniques on the risk of death from breast cancer. Can Med Assoc J 1997;157:1205-12.

52. Baines CJ, Miller AB. Mammography versus clinical examination of the breasts. Monogr Natl Cancer Inst 1997;22:125-9.

53. Fletcher SW, Black W, Harris R, Rimer BK, Shapiro S. Report of the International Workshop on Screening for Breast Cancer. J Natl Cancer Inst 1993;85: 1644-56.

54. Ferguson JH. National Institutes of Health Consensus Development Conference statement: breast cancer screening for women ages 40-49, January 21-23, 1997. J Natl Cancer Inst 1997;89:1015-25.

55. Bjurstam N, Björneid L, Duffy SW, et al. The Gothenberg Breast Screening Trial: first results on mortality, incidence, and mode of detection for women ages 39-49 years at randomization. Cancer 1997;80:2091-9.

56. Smart CR, Hendrick RE, Rutledge JH, Smith RA. Benefit of mammography screening in women ages 40 to 49 years. Current evidence from randomized controlled trials. Cancer 1995;75:1619-26.

57. de Koning HJ, Boer R, Warmerdam PG, Beemsterboer PMM, van der Maas PJ. Quantitative interpretation of age-specific mortality reductions from the Swedish breast cancer-screening trials. J Natl Cancer Inst 1995;87:1217-23.

58. Chakraborty R, Little MP, Sankaranarayanan K. Cancer predisposition, radiosensitivity and the risk of radiation-induced cancers. IV. Prediction of risks in relatives of cancer-predisposed individuals. Radiât Res 1998;149:493-507.

59. Kerlikowske K, Grady D, Barclay J, Sickles EA, Eaton A, Ernster V. Positive predictive value of screening mammography by age and family history of breast cancer. JAMA 1993;270:2444-50.

60. Kerlikowske K, Grady D, Barclay J, Sickles EA, Ernster V. Effect of age, breast density, and family history on the sensitivity of first screening mammography. JAMA 1996;276:33-8.

61. American College of Physicians. Screening for ovarian cancer: recommendations and rationale. Ann Intern Med 1995;121:141-2.

62. Gallion HH, Park R. Developing intervention/prevention strategies for individuals at high risk of developing hereditary ovarian cancer. MonogrNatl Cancer Inst 1995;17:103-6.

63. NIH Consensus Development Panel on Ovarian Cancer. Ovarian cancer: screening, treatment and follow up. JAMA 1995;273:491-7.

64. The prostate question, unanswered still [editorial]. Lancet 1997;349:443.

65. Fintor L, Hall I. Prostate cancer meeting confronts controversies. J Natl Cancer Inst 1995;87: 1743^1.

66. Chodak GW, Thisted RA, Gerber GS, et al. Results of conservative management of clinically localized prostate cancer. NEnglJMed 1994;330:242-8.

67. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology 1997;112:594-642.

68. Hardcastle JD, Chamberlain JO, Robinson MHE, et al. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet 1996;348: 1472-7.

69. Kronborg O, Fenger C, Olsen J, Jorgensen OD, Sondergaard O. Randomised study of screening for colorectal cancer with faecal-occult-blood test. Lancet 1996;348:1467-71.

70. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. NEnglJMed 1993;328:1365-71.

71. Selby JV, Friedman GD, Quesenberry CP Jr, Weiss NS. A case-control study of screening sigmoidoscopy and mortality from colorectal cancer. ? Engl J Med 1992;326:653-7.

72. Selby JV, Friedman GD, Newcomb PA, Weiss NS. Re: Case-control studies of colorectal cancer mortality: is the case made for screening sigmoidoscopy? J Natl Cancer Inst 1993;85:503-40

73. StefanekME. Bilateral prophylactic mastectomy: issues and concerns. J Natl Cancer Inst 1995;17:37-42.

74. Eldar S, Meguid MM, Beatty JD. Cancer of the breast after prophylactic subcutaneous mastectomy. Am J Surg 1984;148:692-3.

75. Goodnight JE Jr, Quagliana JM, Morton DL. Failure of subcutaneous mastectomy to prevent the development of breast cancer. J Surg Oncol 1984 ;26: 198-201.

76. Schräg D, Kuntz KM, Garber JE, Weeks JC. Decision analysis - effects of prophylactic mastectomy and oophorectomy on life expectancy among women with BRCA1 or BRCA2 mutations. NEnglJMed 1997;336:1465-71.

77. Grann VR, Panageas KS, Whang W, Antman KH, Neugut ??. Decision analysis of prophylactic mastectomy and oophorectomy in BRCA1 -positive and BRCA2-positive patients. J Clin Oncol 1998;16:979-85.

78. Gail MH, Brinton LA, Byar DP, et al. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. J Natl Cancer Inst 1989;81:1879-86.

79. Fisher B, Costantino JP, Wickerham DL, et al. Tamoxifen for prevention of breast cancer: report of the National Surgical Adjuvant Breast and Bowel Project P-l Study. J Natl Cancer Inst 1998;90: 137 1-88.

80. Powles T, Eeles R, Ashley S, et al. Interim analysis of the incidence of breast cancer in the Royal Marsden Hospital tamoxifen randomised chemoprevention trial. Lancet 1998;352:98-101.

81. Veronesi U, Maisonneuve P, Costa A, et al. Prevention of breast cancer with tamoxifen: preliminary findings from the Italian randomised trial among hysterectomised women. Lancet 1998;352:93-7.

82. Pritchard KI. Is tamoxifen effective in prevention of breast cancer? [commentary]. Lancet 1998;352:80-1.

83. Ault A. Tamoxifen prevention claim will not be allowed in USA [News item]. Lancet 1998;352:883.

84. Greenwald P, Kelloff G, Burch- Whitman C, Kramer BS. Chemoprevention. CA Cancer J Clin 1995 ;45:3 1^19.

85. Malone KE, Daling JR, Weiss NS, McKnight B, White E, Voigt LF. Family history and survival of young women with invasive breast carcinoma. Cancer 1996;78:1417-25.

86. Porter DE, Cohen BB, Wallace MR, et al. Breast cancer incidence, penetrance and survival in probable carriers of BRCA1 gene mutation in families linked to BRCA1 on chromosome 17ql2-21.5r J Surg 1994;81:1512-5.

87. Johannsson OT, Ranstam J, Borg A, Olsson H. Survival of BRCA1 breast and ovarian cancer patients: a population-based study from southern Sweden. J Clin Oncol 1998;16:397-404.

88. Ansquer Y, Gautier C, Fourquet A, Asselain B, Stoppa-Lyonnet D. Survival in early-onset BRCA1 breast-cancer patients. Lancet 1998;352:541.

89. Breast Cancer Linkage Consortium. Pathology of familial breast cancer: differences between breast cancers in carriers of BRCA1 or BRCA2 mutations and sporadic cases. Lancet 1997;349:1505-10.

90. Karp SE, Tonin PN, Begin LR, et al. Influence of BRCA1 mutations on nuclear grade and estrogen receptor status of breast carcinoma in Ashkenazi Jewish women. Cancer 1997;80:435^11.

91. Kramer BS, Screening Editorial Board of the Physician Data Query. NCI state of the art statements on cancer screening. In: Greenwald P, Kramer BS, Weed DL, editors. Cancer Prevention and Control. New York: Marcel Dekker, 1995:719-22.