Validation of a breast cancer risk assessment model in women with a positive family history

Estimation of breast cancer risk by women aged 40 and over: a population-based study

OBJECTIVE

Identify factors associated with knowledge of breast cancer and estimation of risk.

METHODS

Telephone survey of 412 women aged 40 and over, living in Montreal and selected by random digit dialing.

RESULTS

The majority of the respondents had recently been exposed to some information on breast cancer, but only a third quoted the average lifetime probability estimate of about 1 in 10. Older individuals systematically considered themselves at low risk (odds ratio (OR) of perceiving risk as lower than average for women aged 50 or over versus under 50: 2.6, 95% confidence interval: (1.5, 4.6)). In addition, both a first-degree family history of breast cancer (OR: 5.3 (1.7, 17.0)) and a recent mammogram (OR: 3.0 (1.4, 6.2)) were strongly associated with a woman's probability of perceiving herself at high risk.

CONCLUSIONS

Information campaigns should emphasize the frequency of breast cancer in different age groups and the strength of the established associations with specific risk factors. Better knowledge of risk could promote sustained participation in breast screening programs.

Evaluating minority recruitment into clinical studies: how good are the data?

PURPOSE

There has been much publicized concern about difficulty with minority recruitment into research studies, particularly since minority inclusion in randomized clinical trials was mandated by the 1993 National Institutes of Health Revitalization Act. We reviewed recruitment data in published reports from clinical studies to assess the actual degree of success in recruiting minorities versus whites and to identify barriers to recruitment.

METHODS

We abstracted articles published between September 1993 and February 1995 that reported detailed results of participant recruitment for studies conducted in the United States.

RESULTS

Of 65 articles meeting our eligibility criteria (median sample size, 1323), only one (1.5%) reported the racial/ethnic composition of potential study participants. Only two articles (3.1%) provided information about the racial/ethnic composition of eligible subjects, and only one (1.5%) provided information about the racial/ethnic composition of refusing subjects. For enrolled subjects, race/ethnicity was less likely to be reported (58.5%) than were age (90.8%) or gender (80.0%).

CONCLUSIONS

The published literature currently contains insubstantial data to either refute or prove that there are differential recruitment rates among minorities as compared with whites. Changes in reporting will be needed in order to track progress in this area.

Psychological issues in cancer genetics: current research and future priorities

There has been a rapid expansion of genetics research in the field of cancer since cancer predisposing genes are now known to cause a proportion of common cancers as well as rarer cancer syndromes. As a result, the psychosocial impact of being at high risk of cancer has become a focus of evaluation, and studies are being reported which set out to evaluate both the uptake and psychological outcome of genetic counselling, testing and surveillance. Available data concerning psychological aspects are reviewed, including for example, possible implications of genetic testing, attitudes and uptake of breast screening and accuracy of women's risk estimates. Work is in progress to assess the more controversial areas of prophylactic mastectomy, and chemoprevention. Other research examines the longer term impact of belonging to a Cancer Family, and of interventions offered to high risk families. This is crucial since the uptake of counselling and testing is likely to be much greater in cancer prone families than those with other genetic disorders, yet detection and prevention strategies are still unevaluated for important genetically determined cancers such as breast cancer.

Assessing individual risk for breast cancer: risky business

There is increasing demand for prediction of individual women's risk for breast cancer from women, clinicians, researchers, and health planners. Risk assessment for breast cancer is the process of identifying characteristics of an individual woman that are relevant to her risk, and combining those characteristics into a quantitative or qualitative risk profile. This article reviews and compares available methods of predicting risk, discusses benefits and drawbacks to the methods, and compares risk estimates for several hypothetical subjects using the different methods. Current and future uses for risk assessment are described. Risk assessment, while a promising tool for research now, and for clinical areas in the future, is still too imprecise for accurate prediction of breast cancer occurrence in individuals.

The current role of prophylactic mastectomy

Prophylactic mastectomy has a role in preventing breast cancer in the woman at high risk. The rare indications for this operation are based on genetic and histologic factors that affect relative or cumulative risk. Evaluation of women at high risk must draw on multidisciplinary expertise, including genetic counseling. If prophylactic mastectomy is recommended, skin-sparing total mastectomy (not subcutaneous) with autogenous tissue reconstruction is the preferred approach.

Gail model breast cancer risk components are poor predictors of risk perception and screening behavior

The Gail model is being used increasingly to determine individual breast cancer risk and to tailor preventive health recommendations accordingly. Although widely known to the medical and biostatistical communities, the risk factors included in the model may not be salient to the women to whom the model is being applied. This study explored the relationship of the individual Gail model risk factors to perceived risk of breast cancer and prior breast cancer screening among women with a family history of breast cancer. Data from baseline interviews with 969 women found a striking disparity between the objective risk factors included in the model and the accuracy of perceived risk and screening behaviors of this population, particularly among women over the age of 50 years. Risk perception accuracy was unrelated to all of the Gail model risk factors for all age groups. Reported mammography adherence was only associated with having had a breast biopsy in both age groups. Breast self examination (BSE) practice was independent of all measured factors for both age groups. These findings support the need for further research to identify additional determinants of risk perception and motivators of screening behavior.

The effect of an educational intervention on the perceived risk of breast cancer

OBJECTIVE

To appraise women's perceived risk of developing breast cancer and the effects of a physician's educational intervention on this perception.

DESIGN

Longitudinal before-and-after study involving four measures of participants risk of developing breast cancer. Eligible women provided the data needed to calculate an objective estimate of their individual risk of developing breast cancer before age 80 using the Gail formula. They also provided a subjective estimate of their individual perceived risk. Then, each participant met with a general internal medicine physician who provided personalized information and education. Immediately after education, and again several months later, we reassessed each woman's perceived risk.

SETTING

Physicians office.

PARTICIPANTS

A convenience sample of 59 women participating in the Tamoxifen Breast Cancer Prevention Trial. Twenty-nine women returned for the follow-up risk assessment.

MEASUREMENTS AND MAIN RESULTS

The median calculated risk of breast cancer before age 80 (by the Gail formula) was 15%, but the median perceived risk before educational intervention was 50%. The perceived risk after educational intervention fell to 25%. At late follow-up, the median perceived risk remained at 25%. The difference between the preeducational perceptions and the calculated estimates was significant (1) < .0001). After educational intervention, perceived risk measures shifted closer to the calculated value, but still remained significantly higher (p <.0001).

CONCLUSIONS

Women often substantially overestimate their chances of getting breast cancer. Educational intervention by a physician, including explanation of an individual's calculated risk, can reduce this error. The effect of education appears to persist at least for several months.

Familial breast cancer

Tremendous progress has been made in recent years in understanding the genetic basis of breast cancer. Epidemiological models have been devised which allow risk estimation for women with one or two relatives with breast cancer. The existence of two autosomal dominant breast cancer susceptibility genes (BRCA1 and BRCA2) have been documented. BRCA1 has been isolated and we are beginning to catalog BRCA1 mutations which appear to be etiologically related to the development of breast cancer in certain families. These BRCA1 mutation studies and the isolation of BRCA2 will allow development of accurate DNA-based testing for the prediction of individual risk in women with a family history of breast cancer. Work by several groups is progressing rapidly on the isolation BRCA2. Ongoing studies investigating the function of BRCA1, and eventually BRCA2, will lead to a better understanding of the role these important genes normally play in preventing the development of breast cancer. Finally, additional breast cancer susceptibility genes are likely to be identified, further expanding the number of tools we have to study, and eventually prevent, breast cancer.

Estimating disease risks for individuals with a given family history in different populations with an application to breast cancer

For many chronic diseases, including most of the common cancers, a family history is known to be a strong independent risk factor. For breast cancer, estimation of risk as a function of family history is considered to provide useful risk assessment for women with a family history of breast cancer. Detailed tables that predict the cumulative risk of breast cancer at specific ages based on various combinations of family history have been constructed for the American white population. Most chronic diseases, however, have a multiple etiology, with multiple genetic and environmental factors. Family history can thus be a result of different susceptibility loci and the aggregation of sporadic cases by chance. Since the relative contribution of the genetic factor may differ in different populations, these tables may not be directly applicable to other populations. We present a method to decompose available estimates of risk based on family history for an arbitrary disease into probabilities for genotype frequency given a particular family history and for disease probabilities by genotype carrier status. These can be reconstructed to obtain risk estimates for different populations. Implicit assumptions are made in the estimation process. These are based on the current state of knowledge and can be updated as further knowledge accumulates.

Using loss of heterozygosity data in affected pedigree member linkage tests

Linkage analysis can be used to test the hypothesis that a marker locus of known location segregates independently from a presumed disease gene. One way to test this hypothesis is to measure the similarity of marker alleles among pairs of relatives affected with the disease. When the disease under consideration is cancer, it is possible to take advantage of the marker alleles in tumors to revise the similarity measure obtained from the observations made in constitutional tissue. Only cancers that arise through the model of recessive oncogenesis are amenable to this revised analysis. This model postulates that cancer is caused by somatic genetic changes which result in the loss of one or both copies of a normal allele at a tumor suppressor locus. If an individual's inherited genotype is heterozygous at the marker locus, the model of recessive oncogenesis suggests that we may observe loss of constitutional heterozygosity at the marker locus in the tumor. In this report, we how how to incorporate this loss of heterozygosity data into affected pedigree member linkage tests. The revised procedure is illustrated using data obtained from relatives with breast cancer. Substantial improvement in the power to reject the different chromosome hypothesis is obtained when loss of heterozygosity is observed in multiple relatives with the same marker alleles retained in the tumors.