Epidemiology

Research Project Awards


top


Breast Cancer Incidence in Women With Abnormal Cytology

Margaret R. Wrensch, Ph.D.
University of California, San Francisco

Breast cancer is a leading cause of death among women in California and the United States. Early diagnosis is considered to be a key factor to minimize mortality; thus, techniques to identify high risk women are essential.

Although methods to identify women with inherited susceptibility to breast cancer are becoming more promising, they are not yet routine and only 5%-10% of breast cancer cases are likely to be due to inherited susceptibility. Breast cancer risks associated with specific histologic findings in benign breast biopsies have been quantified, but fewer than 15% of women who develop breast cancer have had a prior breast biopsy. Mammography has limited predictive value in women under age 50 when used without regard to other risk indicators. Nipple aspiration of breast fluid is a simple, non-invasive way of obtaining fluid from breasts of women who are neither pregnant nor lactating. Breast cells in this fluid can be classified as either normal or as showing various abnormalities, including hyperplasia (an increase in the number of cells in a tissue or organ which results in an enlargement of the area), atypical hyperplasia, and cancer. We previously followed 3170 women who participated in breast fluid studies between 1972 and 1980. Women with proliferative cytology (i.e., hyperplasia or atypical hyperplasia) were significantly more likely to develop breast cancer through April, 1991 than women with normal cytologic findings in nipple aspirates of breast fluid or than women from whom fluid could not be obtained. Results were strongest and most consistent for women under age 55 at time of nipple aspiration. In this study, we will determine the breast cancer incidence and mortality for 3170 women from the original cohort over the period May, 1991 through December, 1996. We will also determine breast cancer incidence, through December, 1996 in 4,080 additional women with known nipple aspirate cytology who participated in breast fluid studies between 1980 and 1991. Information from this new study will be used to augment our initial findings on the prognostic value of cytologic findings in nipple aspirates of breast fluid in identifying women at high risk of breast cancer. Incorporating nipple aspirate cytology findings with other indicators of breast cancer risk represents an innovative approach to identifying women at high risk of breast cancer for either intensive clinical management or for epidemiologic studies of prevention or progression of breast cancer.

This study has immediate relevance to the BCRP priority issue of early detection of breast cancer and to the identification and careful characterization of women at high-risk of breast cancer; it also will provide a data source of risk factors, breast cancer cases, and a large pool of women from which to choose controls for studies of genetic markers of breast cancer susceptibility and progression.

back  top

Case-Control Study of Breast Cancer in Asian-American Women

Anna H. Wu, Ph.D.
University of Southern California

The main objective of this population-based case-control study of breast cancer in Asian-Americans is to investigate the causes of their increased breast cancer rates compared to rates in Asia. The study design and sample size will allow us to test the main hypothesis that high intakes of soy products, vegetables and fruits, dietary fiber, and specific micronutrients reduce the risk of breast cancer and that high intakes of fat increase the risk of breast cancer. We will also examine the association of breast cancer with physical activity, body size, body fat distribution, and weight change.

The proposed three-year investigation will include interviews with 350 Asian-Americans (Chinese, Filipino, Japanese), ages 25 to 74, diagnosed with histologically-confirmed (i.e., confirmed by examination of biopsied breast tissue), primary invasive breast cancer identified in Los Angeles County. An equal number of Asian Americans without breast cancer, matched to cases on age, ethnicity, and neighborhood (the "control" group) will also be interviewed. A standardized in-person interview will be conducted and will include a comprehensive diet history covering usual diet during the year before diagnosis for breast cancer cases and the year before interview for controls.

While this study is a complete, self-standing study, we are very fortunate in being able to conduct this study at this time for two reasons. First, it will allow us to compare our results on physical activity and use of exogenous hormones to those obtained in the Los Angeles component of the nationwide CARE study (which is not covering the Asian-American community or dietary factors). Second, this study will build upon and expand our study on Asian-Americans conducted in the 1980s by adding the important 55-74 age group, information on physical activity, and specifics of dietary factors. We can begin to answer some of the very important new questions which have been raised in the area of physical activity, diet, and cancer during the last decade of research. Thus, with our relatively modest grant, we will be able to gain the advantages of much more costly research conducted with other funds.

In summary, this study will help to provide a better understanding of the role of soy and other modifiable risk factors (i.e., other dietary components, physical activity, body size) in the etiology of breast cancer. Achieving these objectives will have substantial public health significance not only for Asian-Americans, but for all racial/ethnic groups, and will have important implications in the prevention of this disease.

back  top

Susceptibility to Breast Cancer: Dietary and Other Factors

Ralph J. Delfino, M.D., Ph.D.
University of California, Irvine

The proposed study will examine whether an elevated level of exposure to the food-borne carcinogens known as heterocyclic aromatic amines (HAAs) is associated with an increased risk of breast cancer in women. It is biologically plausible that such an association may exist given over 2 decades of experimental research on these potential carcinogens, including several studies showing that they are potent mammary gland cancer-causing agents (carcinogens) in rodents. It is now an appropriate time to examine whether there is any risk of breast cancer in women attributable to exposure to dietary HAAs. These substances are formed when meat is cooked by many common methods at high temperatures, e.g., barbecuing. Simple methods are available to prevent their formation prior to any preferred method of cooking. Therefore, the proposed research could ultimately point to practical approaches to the primary prevention of breast cancer. In addition, the proposed study will examine a well-characterized gene known as NAT2, which is involved in the formation of deleterious mutations caused by HAAs.

The proposed case-control study will recruit 2400 women who are scheduled for a diagnostic breast biopsy to rule out a carcinoma (the most common form of breast cancer), with an expected 600 cases of breast cancer and 1800 controls with benign breast diseases. Since subjects will be interviewed prior to the biopsy, selection and recall bias problems, which are common to case-control studies, will be reduced considerably. The control group will also be divided and compared based upon their biopsy results which are thought to indicate either high, low, or no risk of future breast cancer. An interviewer-administered questionnaire will be used to assess dietary exposures to HAAs. In addition, other important factors, which may influence or interfere with the exposures of interest or interact with them, will be assessed with self-administered dietary and risk factor questionnaires so that they may be taken into account. Diagnoses (from 4 participating breast cancer treatment centers) will be confirmed by the research pathologist. Known or suspected risk factors for breast cancer will be controlled for, including dietary nutrients, body mass, temporal changes in body weight, height, age, menstrual, pregnancy and lactation history, family history of breast and ovarian cancer, alcohol intake, steroid hormone use, socioeconomic status and race/ethnicity.

In summary, the proposed study will extend research on the relationship between breast cancer and diet, by examining specific agents which may explain the inconsistency of other epidemiological studies showing that high levels of meat consumption, and possibly fat, are associated with breast cancer risk. A better understanding of the importance of HAAs and NAT2 genotype can provide important insights into: 1) possible breast cancer-causing mechanisms; and 2) practical approaches to the primary prevention of breast cancer in the general population.

back  top

The Role of Plant Estrogens in Reducing Breast Cancer Risk

Pamela L. Horn-Ross, Ph.D.
Northern California Cancer Center

This study will address two questions: 1) whether estrogens found in plant foods can reduce breast cancer risk, and 2) whether obesity increases breast cancer risk only when a woman's diet does not contain a sufficient amount of these plant estrogens.

The estrogens found in a number of plants are similar to the drug tamoxifen. Women with breast cancer who take tamoxifen are less likely to develop cancer in the other breast and tamoxifen is currently being tested as a way to prevent the development of breast cancer in high risk women. Eating plant foods, such as beans and whole grains, which contain a substantial amount of weak estrogens, may be a natural and cost effective alternative way to reduce breast cancer risk. Among postmenopausal women, those who are overweight have higher estrogen levels and are at greater risk of developing breast cancer. We will test the hypothesis that overweight women who eat a lot of foods which are rich in plant estrogens may lower their increased risk of breast cancer since the weak plant estrogens they are consuming block their own body's strong estrogens from binding with estrogen receptors, thus preventing breast cancer cells from growing. The lower breast cancer rates in Latina women despite their greater obesity may be partly due to their higher intake of plant estrogens from bean sources. If confirmed, this hypothesis could have important implications for reducing breast cancer risk through diet, particularly in White, African American, and Latina women among whom postmenopausal obesity is an important health concern. Our study will include White, African American, and Latina women, age 50-79, residing in two Bay Area counties. 325 women with breast cancer will be identified through the State's cancer database and compared to randomly chosen women without breast cancer. Women will be interviewed in their homes in either English or Spanish about their diet and other breast cancer risk factors; body measurements, including weight and height, will be taken.

The results of this study will provide additional understanding of the role that diet plays in the development of breast cancer and may have important implications for breast cancer prevention.

Innovative, Developmental and Exploratory Awards (IDEAs)


Genetic Susceptibility to Breast Cancer

Brian E. Henderson, M.D.
University of Southern California

A substantial body of experimental, clinical and epidemiologic evidence indicates that hormones play a major role in the etiology of breast cancer. The known risk factors for breast cancer can be understood as measures of the cumulative exposure of the breast to estrogen and, perhaps, progesterone. Exposure to these hormones causes breast cells to grow and divide. We hypothesize that certain genes involved in the metabolism of estrogen in the body, such as those that convert estrogen from less active to more active forms, play important roles in breast cancer. We hypothesize that changes in the estrogen receptor (ER) gene, which works in concert with estrogen and other factors to signal breast cells to divide, are associated with breast cancer risk.

We propose to conduct a study to examine the possible associations between three genes and breast cancer in 400 women from an ethnically diverse cohort in Los Angeles County and Hawaii established in 1993. Fifty breast cancer cases and fifty controls between 45 and 74 years old from each of the following ethnic groups will be included in the study: African-American, Hispanic, Japanese, and White. We have also collected data on family history, diet, and other lifestyle habits from these women via a mailed questionnaire. The questionnaire was translated into Spanish for study participants who were not English-speaking.

To date, very few studies have looked at the possible genetic components of breast cancer between and within different ethnic populations, and no study has focused on these specific genes. The significance of this proposed study lies in its multi disciplinary approach, focus on multiethnic populations and potential future impact. If this work identifies associations between these genes and breast cancer, it could be used to identify possible targets of intervention such as modification of estrogen metabolism within breast cells to slow or stop the proliferation of breast cancer cells.

back  top

Vitamin D & Genetics of Vitamin D Receptor in Breast Cancer

Sue A. Ingles, Ph.D.
University of Southern California

The specific aims of this proposal address the BCRP priority issues of the underlying biological causes of breast cancer (etiology) and the development of the disease (pathogenesis): we plan to test whether genetic variants in the vitamin D receptor (VDR) are related to breast cancer incidence and whether blood vitamin D levels interact with the genetically determined VDR variants to influence the risk of breast cancer. These genetic variants have recently been shown to have a strong effect on bone density; and, since vitamin D plays an important role in cellular growth and differentiation (the process by human body cells become different and specific, e.g., a kidney or breast cell), it is possible that variation in the VDR may affect breast cancer in a similar manner.

To investigate this hypothesis, we propose to conduct a case control study of 200 matched pairs of breast cancer cases and controls (that is, similar women who do not have breast cancer), equally divided among four ethnic groups: African-American, Hispanic, Japanese, and White. DNA from cases and controls will be typed (examined) for the commonly used marker (BSMI) which is tightly linked to the suspected VDR causal mutation. We will also type one of the variants which is a candidate for the causal mutation. Since inconsistencies among previous studies suggest that different BSMI variants may be associated with the causal mutation in different ethnic populations, we will examine linkage disequilibrium between the two loci in each of the four ethnic groups. These data will provide information about the usefulness of BSMI as a measure for the VDR causal mutation.

Given the role that vitamin D is known to play in cellular growth and differentiation, the recent work using vitamin D analogues to suppress the growth of breast cancer cells, and the recent discovery of VDR variants which appear to account for a large proportion of genetic variation in bone density, we feel that there is sufficient evidence to investigate the role of the VDR in breast cancer incidence. A positive result would have implications for breast cancer prevention and treatment: administration of vitamin D or its analogues might be useful for prevention of breast cancer or its progression in individuals with high risk variants. Since the suspected high risk variant is common in the population, a positive result could have a large impact on reducing the incidence of breast cancer in California.

New Investigator Awards


top


Determinants of Breast-Conserving Surgery and Survival

Cyllene R. Morris, D.V.M., Ph.D.
California Public Health Foundation

Primary therapy for breast cancer in women with localized disease has changed substantially in recent years. Breast-conserving surgery (BCS) with radiation therapy is an equally effective alternative to mastectomy, and was recommended for most women with stage I or II breast cancer by the 1990 National Institutes of Health Consensus Conference. (For this study, stage I or II is defined as localized tumors without lymph node involvement, or tumors up to 5 cm in diameter, with or without extension to axillary lymph node nodes on the same side.) Nevertheless, fewer than 50% of women who are eligible receive BCS in California, even though mastectomies may have profound emotional consequences and do not improve survival rates in women with small tumors. The goal of this study is to determine the reasons for the apparent underutilization of BCS in California. A better understanding of this phenomenon may reduce the human and economic costs of breast cancer. In particular, the physical and psychological impact associated with unnecessary mastectomies represent a human cost that can be minimized by understanding which factors affect the choice of surgical treatment.

The objectives of the project are to: 1) determine statewide and regional trends of BCS utilization and radiation therapy, 2) determine the most significant predictors of choice of surgery for BCS-eligible women, and 3) compare the five-year survival experience of women receiving BCS and BCS-eligible women receiving mastectomies. Data from the California Cancer Registry (CCR), including all reported diagnoses of first primary breast cancer within the state from 1988 to 1995, will be used in the analysis. This study will be the first to address the issue of BCS utilization and survival on a statewide basis. The CCR operates the most complete cancer reporting system in California, and the CCR breast cancer database will be complemented and enhanced through linkages with several additional statewide files. These linkages will enable the analysis of a wide range of factors which may influence the choice of surgical treatment for primary breast cancer, including socioeconomic status, education, insurance coverage, and comorbidities. These factors have not yet been evaluated together with information on clinical factors, treatment, characteristics of the patient, hospital, and county of residence as predictors of surgery received and subsequent survival. The database generated during this study will contain the most comprehensive population-based information on breast cancer in California. Given the large number of breast cancer cases in the state, the proposed study will have enough power to reach a statistically sound conclusion. In addition, the substantial variation in race/ethnicity in California allows for in-depth analysis of possible differences in patterns of care among minority women.

Understanding why BCS has been underutilized in California is an important step towards ensuring that women with breast cancer have an informed choice and receive optimal treatment.

back  top

Exercise, Ovarian Function, and Breast Cancer Prevention

Lisa S. Shames, Ph.D.
University of Southern California

There is substantial evidence to suggest that estrogens play a key role in the biological cause of breast cancer. We recently observed in a large population-based study that exercise, including moderate exercise, may significantly reduce a woman's risk of breast cancer. We hypothesize that the reduction in risk with exercise is due to altered ovarian function, specifically, more frequent anovulatory cycles (i.e., a menstrual cycle without release of an ovum) and/or reduced serum estrogen (and progesterone) levels in ovulatory cycles which likely contribute to a reduced exposure of breast tissue to estrogen and thus reduce the risk of breast cancer. Studies of highly trained athletes show that women who engage in exercise for close to an hour or more per day have a higher frequency of anovulation, lower levels of estradiol (E2 -- the most active form of estrogen) and, in some cases, a shortened luteal phase length with associated lower estradiol levels.

The purpose of this study is to investigate the relationship between exercise and ovarian function across a range of physical exercise activity that non-athletes may reasonably be expected to participate in (i.e., sedentary to less than an hour per day). Specifically we will aim to: 1) determine the relationship between physical exercise activity and frequency of anovulation, 2) determine pre- and post-ovulatory cycle lengths of women exercising at these various levels, and 3) measure serum hormone levels of E2 in these women. We plan to study 300 college students between the ages of 18 and 25 years, who have never borne children, over 2 complete menstrual cycles. A questionnaire consisting of demographic information, exercise, diet and medical history will be administered.

This study should contribute significantly to our understanding of the relationship between physical exercise activity and ovarian function. With this relationship better understood, it would be possible to provide readily understandable, practical advice to women on how to reduce their breast cancer risk through physical exercise activities.

back  top

Race/Ethnicity and Late Stage Breast Cancer

David J. Delgado, Ph.D., M.P.H.
University of Southern California

Over 20 million Hispanics live in the United States, and more than 50% of them live in California. Between 1974 and 1991, over 50,000 cases of breast cancer have been diagnosed in Los Angeles White and Hispanic women, making it an issue of extreme importance. Breast cancer accounts for the largest number of incident cases of cancer among all ethnic groups in Los Angeles. With the L.A. Hispanic population age distribution reflecting that of third world countries, the issue of age and its impact in early breast cancer detection for the Hispanic population is also important to examine. Incidence rates for racial/ethnic groups have been published, but fail to account for socioeconomic status. Incidence rates by stage-of-diagnosis, socioeconomic status, age, and race/ethnicity are not readily available.

The State of California's Breast Cancer Early Detection Program specifically targets women of lower socioeconomic status and minority racial/ethnic groups in order to improve breast cancer survival and reduce stage-at-diagnosis. In order to measure the progress and success of this program and other similar programs attempting to encourage earlier screening, appropriate baseline data will be needed. In particular, it will be important to be able to confidently examine changes in stage-of-diagnosis over time for subgroups of the population defined by race/ethnicity, age and socioeconomic status. One of the major contributions of this research will be to use various sophisticated statistical techniques (empirical Bayes and constrained empirical Bayes methodology) to provide reliable incidence rates calculated for these sub-populations. This type of methodology will help examine the incidence of late-stage breast cancer as it relates to socioeconomic status, race/ethnicity, and time period, after adjusting for age.

Finally, this methodology will help in studying the combination of all of the above variables in order to better understand the potential impact of early detection interventions. This study can provide a geographically-defined baseline of stage-specific incidence rates from which future calculations of late stage-of-diagnosis can be compared.

We will examine our data for all of Los Angeles, dividing the County for analysis purpose into 86 Medical Service Study Areas (MSSAs) to be targeted for analysis. The results will enable policy makers and community service agencies to target the worst MSSAs with programs for the early detection of breast cancer. We will identify these MSSAs of extreme poverty and estimate baseline information about the incidence of late-stage diagnosis on a racial/ethnic and age-specific basis. Not only should our analysis help to focus future health education/outreach interventions, but will also offer a baseline from which to evaluate the effectiveness of these programs in the future.

back  top

Postdoctoral Fellowship Awards


Alcohol and Risk of Estrogen Receptor-Positive Breast Cancer

Shelley M. Enger, Ph.D.
University of Southern California

In this study, the association of alcohol intake with risk of breast cancer will be assessed separately for estrogen receptor-positive and estrogen receptor-negative tumors, using data from two large epidemiologic studies of risk factors for breast cancer. Estrogen receptors are cellular proteins that bind estrogen, a hormone that stimulates growth of breast tissue. Breast tumor cells with higher levels of estrogen receptors (estrogen receptor-positive tumors) respond better to hormonal therapies than breast tumor cells with low levels of estrogen receptors (estrogen receptor-negative tumors). If alcohol intake is related only to estrogen receptor-positive tumors, this may clarify the role of alcohol in the etiology of breast cancer, and potentially support an estrogen-mediated effect of alcohol on breast cancer. In addition, an association of alcohol with risk of breast cancer that differs for estrogen receptor-positive and estrogen receptor-negative cancers would support the hypothesis that estrogen receptor status defines different types of breast cancer, rather than two stages in the development of one disease.

Subjects are 2,330 women diagnosed for the first time with breast cancer (cases) and 2,330 women without breast cancer (controls) participating in two population-based case-control studies (Studies A and B) of risk factors for breast cancer. All subjects are white (Anglo or Hispanic) female residents of Los Angeles County. Each case subject is individually matched to a control subject on age, parity (Study A only), and neighborhood of residence. Eligible cases in Study A were women in Los Angeles County aged 40 years or younger diagnosed with in situ (localized) or invasive breast cancer from July, 1983, through December, 1988. Eligible cases in Study B were women in Los Angeles County aged 55-64 years diagnosed with in situ or invasive breast cancer from March, 1987, through December, 1989. Alcohol consumption, demographics, and other information were obtained in interview questionnaires. Estrogen receptor status was determined for all case subjects, but for 80 percent of the subjects the original pathology laboratories must be queried for the results. Appropriate statistical analyses will be used to estimate the association of alcohol intake with risk of estrogen receptor-positive and estrogen receptor-negative breast cancer.

The results of this research may increase understanding of mechanisms through which alcohol might increase the risk of breast cancer, and the postdoctoral fellow will gain valuable experience working in a renowned breast cancer research program, leading to a career in breast cancer research.

top

Physical Activity: Impact on Hormones and Breast Cancer Risk

Catherine L. Carpenter, Ph.D.
University of Southern California

Evidence from epidemiologic studies suggests that women who engage in strenuous physical activity are at a reduced risk of breast cancer. Possible biological mechanisms have been suggested to explain the influence of physical activity on breast cancer risk. Physical activity is one of several breast cancer risk factors that affect exposure to endogenous hormones.

Strenuous physical activity has been shown to increase the age at menarche, menstrual cycle length, and number of anovulatory cycles (a menstrual cycle without release of an ovum). Furthermore, moderate physical activity has been linked to the reduction of frequency of ovulatory menstrual cycles. Despite the emergence of this relationship, several population subgroups such as postmenopausal women and premenarcheal girls have not been well studied. The objectives of the proposed research are: 1) to assess the effect of lifetime physical activity on breast cancer risk in postmenopausal women, and 2) to further clarify how physical activity influences age at puberty and ovarian hormone patterns in the years immediately following the onset of puberty.

Data from a matched case-control study of hormone replacement therapy and breast cancer risk among postmenopausal Caucasian women age 55 or older will be utilized to address the first objective. A total of 1510 case-control pairs (women who have had breast cancer and women who have not had breast cancer, matched by age, race/ethnicity, and neighborhood) provided detailed information on their lifetime participation in physical activities. Appropriate statistical techniques will judge the possible effect of physical activity on breast cancer risk.

Data from a cohort study of 1200 predominately Hispanic school children from the 4th through the 7th grades will be utilized to address the second objective. Data on physical activity and menarche status is available for 808 girls. Hormonal status and physical activity data has been collected on 345 girls. Again, appropriate statistical techniques will be used to estimate the effect of physical activity on hormonal status.

Results from this study could further clarify the link between lifestyle factors and hormonal events among adolescent girls. Furthermore, identification of physical activity as a possible protective factor among postmenopausal women could provide an intervention tool to assist the effort to reduce breast cancer incidence among women in California.

top

Training Program Awards


USC/Norris Breast Cancer Research Training Program

Ronald K. Ross, M.D.
University of Southern California

This proposal seeks support to establish a Breast Cancer Research Training Program at the University of Southern California USC/Norris Comprehensive Cancer Center (NCCC). The NCCC has a long history of breast cancer research, formalized in recent years as the NCCC Breast Cancer Research Program, but has never had a formal, focused, interdisciplinary graduate research training program devoted to breast cancer.

The proposed program seeks to take advantage of the extraordinary racial/ethnic diversity among the 9 million residents of Los Angeles County, the population served by the NCCC. Our goal is to better understand the reasons for the underlying differences in breast cancer incidence, mortality and survival among the numerically most important of these racial-ethnic groups in the County (African-Americans, Latinas, Japanese, Chinese, Filipinos, Koreans and non-Latina whites). We envision a multi-faceted approach involving epidemiology and prevention scientists, behavioral scientists, tumor biologists, and molecular geneticists, and radiation, surgical and medical oncologists. We will use the numerous patient and data resources available to us to achieve this goal. These include the Los Angeles County SEER cancer registry program; a large registry of breast cancer in twins; dietary and lifestyle data from recently characterized large racial/ethnically diverse cohorts; a large and diverse clinical population seen at NCCC affiliated hospitals; and several large tumor and other biological sample banks for studies of tumor biology and molecular genetics.

We describe and document in this application the commitment of the NCCC to establish this Training Program; we summarize the research resources available for training graduate students for breast cancer research; we describe the evaluation process for applicants; and we overview the planned organizational structure of the Training Program relative to the NCCC overall, including advisory groups and their functions.

top