Genetic Testing for Breast-Ovarian Cancer Susceptibility: A Regional Trial

  • Published on

  • View

  • Download

Embed Size (px)


<ul><li><p>GENETIC TESTINGVolume 2, Number 4, 1998Mary Ann Liebert, Inc.</p><p>Genetic Testing for Breast-Ovarian Cancer Susceptibility:A Regional Trial</p><p>STARLENE LOADER,1 JEFFREY C. LEVENKRON,2 and PETER T. ROWLEY1</p><p>ABSTRACT</p><p>To evaluate receptivity to testing for a genetic susceptibility to breast-ovarian cancer, information is neededon the response when the offer is open to all qualifying women in a given region. To qualify in this trial, awoman who had not had breast or ovarian cancer had to have at least two first-degree relatives or one first-and one second-degree relative with breast and/or ovarian cancer, whereas a woman who had had breast orovarian cancer had to have at least one first-degree relative with breast or ovarian cancer and a first- or sec-ond-degree relative without cancer willing to be tested. Of 140 women qualifying and interested enough to re-turn questionnaires requesting baseline information, 111 were referred by their physician and 29 were iden-tified from a regional tumor registry. Of these 140, 112 came for pretest education and 98 of these chose tobe tested. Thus, the acceptance rate was 70% for all those returning baseline questionnaires, but 88% forthose interested enough to come for pretest education. The most common reasons for accepting testing wereto take extra precautions if a mutation were found (42.9%) and to determine if offspring were at risk (24.5%).The most common reasons for declining were anxiety and absence of specific interventions. Factors predict-ing who chose testing were years of education (p &lt; 0.005) and family closeness (p &lt; 0.02). Fourteen deleteri-ous BRCA1 or BRCA2 mutations were found in 13 of the 87 families actually tested. If the criteria for test-ing had been three or more affected family members rather than two or more, the number of families testedwould have been reduced by 46%, but the number of families found to have a deleterious mutation wouldhave been reduced by only 9%.</p><p>INTRODUCTION</p><p>The isolation of the breast-ovarian cancer susceptibil-ity genes BRCA1 (Miki et al., 1994) and BRCA2 (Woosteret al, 1995) and the characterization of mutations in these genesin affected families have made it possible to identify women athigh risk for these types of cancers. Widespread public concernabout these types of cancers, especially about breast cancer,means that large numbers of women may seek testing from theirphysicians even though the impact of such testing has yet to befully evaluated.</p><p>For BRCA1, located on chromosome 17q21 (Hall et al.,1990), over 500 different mutations have been reported (BreastCancer Information Core Steering Committee, 1998). Reportsbased on families with large numbers of affected individualssuggest that a woman with a BRCA1 mutation has an 87%chance (confidence interval 72-95%) of developing breast can-</p><p>cer and a 44% chance (confidence interval 28-56%) of devel-oping ovarian cancer by age 70 (Ford et al., 1994). However,penetrance may be lower in families with fewer affected mem-bers or in certain subpopulations (Easton, 1997), e.g., a reported56% risk of breast cancer and 16% risk of ovarian cancer byage 70 in persons of Ashkenazi Jewish descent (Struewing etal, 1997). BRCA2 is located on chromosome 13ql2-13(Wooster et al, 1994). Over 280 different BRCA2 mutationshave been reported (Breast Cancer Information Core SteeringCommittee, 1998). A BRCA2 mutation confers a high risk ofbreast cancer, similar to a BRCA1 mutation, but a lower riskof ovarian cancer than a BRCA1 mutation (Wooster et al,1995). About 2.5% of Ashkenazi Jews have a BRCA1 orBRCA2 mutation (Easton, 1997). This increase is due chieflyto three specific mutations, two in BRCA1 (185delAG and5382insC) and one in BRCA2 (6174delT).</p><p>Interest in mutation detection is high among affected fami-</p><p>1 Department of Medicine, Division of Genetics, and the 2Department of Psychiatry, University of Rochester School of Medicine, Rochester,NY 14642.</p><p>305</p></li><li><p>306 LOADER ET AL.</p><p>lies (Lerman et al, 1994, 1995) and even in the general public(Andrykowski et al, 1997). Previously, we have assessed in-terest in testing for genetic susceptibility to breast and ovariancancer in Rochester among 498 women visiting their privateobstetrician-gynecologist for any reason and 484 women hav-ing routine mammography in a busy radiologie practice (Cha-liki et al, 1995). These women were unselected for a personalor family history of breast or ovarian cancer. Yet over 90% ineach group said they would want such testing, if free, and 56%provided their name, address, and telephone number for notifi-cation when testing became available.</p><p>The existing information about test receptivity is limited byreliance chiefly on members of single families or small num-bers of families previously recruited for cancer genetic researchwho therefore may not be representative of at-risk families inthe general population (see Discussion). Our study was de-signed to provide more generalizable information and differsfrom previous studies in several ways. First, individuals offeredtesting were not part of families previously identified to havea mutation. Rather, genetic counseling and testing were opento everyone in the county who met the cancer family historycriteria. Second, most individuals coming to inquire about test-ing first learned about it from their own physician, not from aresearch team. Third, to be tested, an individual had to recruita relative willing to be tested. Fourth, all participants expectedto learn their results. Fifth, women who presented themselveswere asked to decide whether or not to be tested on the basisof whether, following detailed pretest education, they expectedthat benefits would outweigh burdens in their own case, andnot in order to help research.</p><p>PATIENTS AND METHODS</p><p>Population studiedTwo populations were studied. One consisted of patients re-</p><p>ferred by their physician because of a family history of breastcancer. Because most of these patients had not had breast can-cer themselves, a second population, women with a family his-tory of breast cancer who had had breast cancer themselves,was obtained from our regional tumor registry.</p><p>To identify the first population, we informed area providersabout the clinical trial at a regular conference of the obstetrics-gynecology staff of each of the five hospitals having such a ser-vice in Monroe County (containing Rochester). Physicianknowledge and attitudes about testing for a genetic suscepti-bility to breast and ovarian cancer as assessed on those occa-sions have been previously reported (Rowley and Loader,1996). All other physicians in the county were invited to par-ticipate via an article in the county medical society newsletter.To be eligible for testing, a woman who had not had breast orovarian cancer, an "at-risk" woman, had to have at least tworelatives with breast or ovarian cancer on the same side of herfamily; one of these had to be a first-degree and the other afirst- or second-degree relative, one of whom had to be living.</p><p>The second population was drawn from all women diagnosedin 1994 with breast cancer at or below age 50 in Monroe County(n = 213) using the Rochester Regional Tumor Registry. Physi-cians of record reported that six had died and did not give uspermission to contact 37 others. We wrote to the remaining 170</p><p>women to inform them of the offer of genetic testing and to in-vite them to contact us if they had a qualifying family history.To qualify, we required a first-degree relative with breast orovarian cancer and an at risk first or second-degree relative sothat these families would be similar to the two-affected-mem-ber families ascertained through physician referral.</p><p>Pretest educationWhen a woman from an affected family telephoned us to in-</p><p>quire about testing, it was determined whether her family metthe criteria stated above. If it did, the woman was sent a set ofquestionnaires, including a cancer family history questionnaire,the Breast Cancer Worry Scale (Lerman et al, 1994), the MillerBehavioral Style Scale (Miller, 1987), the Brief Cope (Carveretal, 1989), the Beck Depression Inventory (Beck et al, 1979),the Life Orientation Test (Scheier et al, 1994), and the Rand36-Item Health Survey 1.0 (SF-36) (Hays etal, 1993). The pur-pose of the Beck Depression Inventory was to exclude from of-fering DNA testing any seriously depressed person, but nonewere detected. Also included was a questionnaire assessingbreast cancer knowledge, attitudes, and surveillance practices.The attitude questions included perceptions about seriousness,susceptibility, curability, and surveillance effectiveness andwere similar in design to those used in previous genetic screen-ing studies of ours (Rowley et al, 1991), and involved ratingusing a Likert scale. As soon as the completed questionnaireswere returned and reviewed, an appointment was made.</p><p>At the first visit, a detailed family history of all types of can-cer was obtained to detect any variant cancer syndrome (e.g.,Li-Fraumeni); none were detected. The patient was given anexplanation of genetic susceptibility for breast and ovarian can-cer and of the significance of mutation detection. Both poten-tial benefits and potential risks and limitations of testing (listedin the Discussion) were enumerated so that she could make adecision for herself; in other words, we merely offered testing,rather than recommending it. It was explained that a negativeresult in an at-risk person had limited value unless an affectedrelative had been shown to have a mutation, because the mu-tation responsible for multiple cases in the family might be onenot detected by current methods or might be in a gene not an-alyzed. For this reason an affected family member had to beshown to have a mutation before an at-risk member was tested.The first visit usually required about 90 min.</p><p>The patient was not asked to make a decision about testingon this first visit, but rather was given an explanatory brochureand a detailed consent/decline form to take home. She was toldthat, if she decided to be tested, she should return with the ap-propriate relative. The appropriate relative for an at-risk womanwas a relative who had had breast or ovarian cancer. If the in-dex patient had had breast or ovarian cancer, we requested thatshe return with a relative who had not had cancer, not becauseit was medically necessary, but because we wished to compareat-risk women who learned about testing from their physicianwith at-risk women who learned about testing from a relativewith breast or ovarian cancer.</p><p>BRCA1 and BRCA2 testingWhen the index patient returned with the appropriate rela-</p><p>tive, any additional questions were answered, informed consentdocumented, and blood drawn on both persons. Mutation de-</p></li><li><p>BREAST-OVARIAN CANCER SUSCEPTIBILITY TESTING 307</p><p>tection for the affected patients was performed by Myriad Ge-netic Laboratories, Inc. (Salt Lake City, UT) by sequencing theentire coding region of BRCA1 and adjacent intronic regionsas previously described (Shattuck-Eidens et al., 1997). Samplesfrom most patients were analyzed independently by OncorMed(Gaithersburg, MD) employing initial screening by antisenseoligonucleotide hybridization for common mutations and byprotein truncation assays for mutations in exon 11 ; these werefollowed by sequence analysis only if an abnormality was de-tected by one of the two screening tests.</p><p>Before results from all the BRCA1 tests had been reported,Myriad Genetics permitted us to offer BRCA2 testing to amember of the 26 families who had submitted samples most re-cently, and whose affected member had tested negative for aBRCA1 mutation. Of these 26 offers, 21 families accepted. Oneindividual each from three other families, who tested negativefor a BRCA1 mutation but were not offered free BRCA2 test-ing, paid for such testing. Thus, of 87 families tested for aBRCA1 mutation, some member of 24 families was also testedfor a BRCA2 mutation. Myriad used the same method to de-tect BRCA2 mutations as for BRCA1 mutations.</p><p>Communication of resultsBefore blood was drawn, it was explained that results were com-</p><p>municated only in person to the index patient and relative together,thus necessitating a third visit. For women found to have a muta-tion, the recommendations for management made by the CancerStudies Consortium (Burke et al, 1997) were communicated.</p><p>Human subjectsThis project was approved by the Research Subjects Review</p><p>Board of the University of Rochester. No charge was made forpretest education, DNA testing, or post-test counseling.</p><p>RESULTS</p><p>Characteristics of index patientsAmong the women who contacted us to inquire about test-</p><p>ing, 140 women ("index patients") met the family history cri-teria and returned questionnaires providing baseline informa-tion. Of these 140 women, 41 women had a personal history ofbreast or ovarian cancer ("affected") and 99 had not ("at risk").These 41 women with a personal history of breast or ovariancancer included 29 patients identified from the Tumor Registry.Of the 140, 111 had families with breast cancer without ovar-ian cancer and 29 had a family history of both breast and ovar-ian cancer. The number of persons with breast or ovarian can-cer per family ranged from two to eleven, but most families(89%) had only two to four such persons. The mean age was44.3 years. The mean number of years of education was 14.9.The proportion married was 74.3%. The proportion employedwas 61.4%. Racial composition was three African-Americans,one Hispanic, and 136 non-Hispanic Caucasians. Nine persons(6.4%) were of Ashkenazi Jewish descent.</p><p>Decision about testingOf these 140 index patients qualifying for testing, 112 came</p><p>for the first visit to receive pretest education, but 28 did not.</p><p>After these 112 index patients had received pretest education,98 chose testing and 14 chose no testing. The acceptance ratewas thus 70% (98/140) for the whole group qualifying for test-ing, but 88% (98/112) for those interested enough to come forpretest education.</p><p>The most common reasons given for choosing testing wereto take extra precautions if a mutation were found (42.9%) andto determine if their offspring were at risk (24.5%). Reasonsfor testing less commonly given were to obtain reassurance orto comply with the recommendation of physician or a relative.</p><p>To those 28 who had qualified for testing, but had never re-turned the set of questionnaires mailed to them, we sent a sec-ond set including a question about their reason for decliningtesting and offered a small payment for completion. Of the 42who did not desire testing (including both the 28 who did notcome for pretest education and 14 who chose no testing aftereducation), 17 provided a reason. Seven were concerned abouttheir emotional reaction to a positive result, five would notchange their surveillance even if their result were positive, threefound that the intended relative was unwilling or had died inthe interim, and two were worried about the effect of a posi-tive result on their insurability.</p><p>Com...</p></li></ul>