Implications of birth cohort patterns in interpreting trends in breast cancer rates.

BACKGROUND: Most investigations of trends in cancer rates are based on a cross-sectional approach, i.e., an examination of trends in rates by year of diagnosis or death. When there are longitudinal effects (i.e., trends in rates with successive birth cohorts), interpretation of cross-sectional trends can be misleading. Based on cross-sectional comparisons, U.S. breast cancer mortality rates have been reported to be decreasing over the last 20 years in younger women but to be increasing during the same period in older women. PURPOSE: To examine the impact of longitudinal effects on the divergence of cross-sectional trends in breast cancer mortality with age, we examined breast cancer mortality rates from 1969 to 1988 by birth cohort for White women in the United States. METHODS: By using a novel, nonparametric, permutational method to analyze 2-year, age-specific mortality rates for women aged 30-89 years, we identified trends in rates with successive birth cohorts. RESULTS: The divergence in trends with age is shown to be consistent with an increase in breast cancer risk with successive birth cohorts from 1900 to 1916 and with a decrease in breast cancer risk with successive birth cohorts beginning around 1926. CONCLUSION: Longitudinal effects have a significant impact on cross-sectional trends in breast cancer mortality. IMPLICATIONS: Continuation of the birth cohort trend in younger women, which could correspond to changes in reproductive patterns accompanying the "baby boom," would result in decreasing cross-sectional trends in women 60-69 years of age over the next decade and in women 70-79 years of age in the subsequent decade. Longitudinal effects must be taken into consideration when monitoring and evaluating the effects of early detection, treatment, and intervention programs using national rates.

Implications of the multistage theory of carcinogenesis applied to occupational arsenic exposure.

The multistage theory of carcinogenesis and its implications for evaluating the effect of exposure to carcinogens in the workplace are described. This theory predicts different relationships between excess carcinogenic risk and duration of exposure, age at initial exposure, and follow-up time since exposure stopped. These relationships are shown to depend on the stage of the carcinogenic process affected by the carcinogen, i.e., action at an early stage or a later stage. The patterns of excess lung cancer mortality were examined for a cohort of copper smelter workers exposed to atmospheric arsenic and other contaminants. Under this multistage hypothesis, the results indicate that arsenic appears to exert a definite effect on a late stage of the carcinogenic process, although an additional effect at the initial stage cannot be conclusively ruled out. Other factors, such as exposure to sulfur dioxide in the environment, calendar year at start of employment, and the potential bias resulting from incomplete exposure histories are also discussed as well as the implications of these results to experimental animal studies.

Analysis of breast cancer mortality and stage distribution by age for the Health Insurance Plan clinical trial.

The Health Insurance Plan (HIP) of Greater New York conducted a clinical trial to determine if screening for breast cancer with mammography and clinical examination would decrease breast cancer mortality. The extent of disease at diagnosis among breast cancers detected by screening and the effect of screening on breast cancer mortality have been evaluated in the cohort of all HIP women diagnosed with breast cancer within 6 years of entry into the trial and followed at least 18 years after trial entry. Six years was the earliest time at which the number of cases diagnosed in the control group was equal to the number of cases diagnosed in the study group. In the cohorts of women 40-49 and 50-64 years of age at entry, shifts were significant to lower stages for screen-detected cases. As a result, the study group women in each age cohort had significantly lower breast cancer mortality than control group women when statistical analyses were restricted to data from cases only. In the 40-49 age-at-entry cohort, the reduced breast cancer mortality in the study group appears to result from lower mortality in stage I cases as well as from earlier case detection, and this may explain differences between the two age-at-entry cohorts in the length of follow-up time required to demonstrate a mortality reduction due to screening.

Birth cohort and calendar period trends in breast cancer mortality in the United States and Canada.

BACKGROUND: Previous studies of regional and temporal variation in U.S. breast cancer mortality rates have been confined largely to analyses of rates for white women. PURPOSE: Breast cancer mortality rates from 1969 through 1992 for white women and black women in four regions of the United States and for all women throughout Canada were compared to identify racial, regional, and temporal differences. Differences and trends in the rates were evaluated in view of breast cancer risk factors and relevant medical interventions. METHODS: Age-period-cohort models were fit to the data, and changes in birth cohort trends (suggesting a change in a breast cancer risk factor or protective factor) and calendar period trends (suggesting, in part, the impact of new or improved medical interventions) were examined. RESULTS: Breast cancer mortality rates for white women were significantly higher in the Northeast than in any other region of the United States (two-sided t tests; P<.005); the rates for black women were not. Birth cohort trends for all women were similar until about 1940, with a moderation of mortality risk beginning around 1924. A marked moderation of risk by 4-year birth cohorts was observed for U.S. white women born after 1950, whereas stable or slightly decreasing trends were observed for U.S. black women and Canadian women. For women born from 1924 to around 1938, fertility rates increased for all three groups; after 1950, they declined uniformly. Looking at temporal effects, we found that the slope of the mortality calendar period trend increased in the 1980s compared with the 1970s for all women. In the last calendar period, 1991-1992, a trend of decreasing mortality rates was found for white women in the United States and for Canadian women. IMPLICATIONS: Widespread environmental exposures are unlikely to explain the higher relative breast cancer mortality rates observed for U.S. white women in the Northeast, since the rates for black women in this region were not higher than in other regions. The moderation of breast cancer mortality rates for women born between 1924 and 1938 coincides with increased fertility rates following World War II. Stable or decreasing mortality rates for U.S. women and Canadian women born after 1950 were not expected in view of declining fertility rates, suggesting a change in a breast cancer risk factor or protective factor. The increase in calendar period trend slope in the 1980s likely reflects the coincident rise in breast cancer diagnosis via mammography. The recent decline in calendar period trend for white women in the United States and for Canadian women may be the result of earlier detection and increased use of adjuvant therapy.

Statistical issues in interpretation of chronic bioassay tests for carcinogenicity.

The interpretation of chronic bioassay tests for carcinogenicity requires that the data be appropriately recorded. A "case history" for each animal links the pathology data for each organ of each animal to the length of its life. This information can be used in interpretation of the tumor incidences in light of the survival information. The role of historical controls was discussed, and the use of significance tests in a multidisciplinary approach to the assessment of the pattern of tumor response was suggested. Multiple comparison methods valid for the interpretation of continuous (or measurement) data do not apply to the discrete data analyses used in these studies. The ideas and methods of these studies were applied to an animal study of chloroform.

Variability in the rates of some common naturally occurring tumors in Fischer 344 rats and (C57BL/6N x C3H/HeN)F1 (B6C3F1) mice.

Variability inthe incidence rates of some common naturally occurring tumors for 72 inbred F344 rat and 54 (C57BL/6N x C3H/HeN)F1 (B6C3F1) mouse control groups used in carcinogenesis bioassays was evaluated. Significant heterogeneity of control rates was observed in at least two of six laboratories for lymphomas-leukemias, liver tumors, and pituitary tumors in the male rat and for pituitary tumors and endometrial stromal polyps in the female rat. Significant interlaboratory heterogeneity was observed for several tumor types in the F344 rat. In contrast, control incidence rates for tumors of the lung and liver and lymphomas-leukemias in B6C3F1 mice were relatively homogeneous within four of five laboratories. Significant interlaboratory heterogeneity was observed, however, for these mouse tumors. The causes of significant heterogeneity in naturally occurring tumor incidence rates within and among laboratories are unknown. Although the most appropriate and important comparison of an experimental group is with its matched control, there may be instances in which the historical control rates provide relevant data needed to clearly interpret carcinogenesis bioassay results. With the use of data from bioassays of 4-chloro-m-phenylenediamine and nitrilotriacetic acid, two examples are presented to demonstrate the usefulness of the historical control data.

Analysis of the role of cancer prevention and control measures in reducing cancer mortality.

One goal of the war against cancer is to create declines in cancer mortality rates. A decrease in these rates can only occur in two ways: 1) a decrease in incidence rates and 2) a real increase in overall survival rates. Reductions in incidence rates can be envisioned to occur through three mechanisms (in order of the time course of cancer): 1) reduction or amelioration of environmental or lifestyle risk factors, 2) use of agents that prevent the occurrence of cancer by blocking the progression to cancer, and 3) early detection at a preneoplastic state combined with treatment that prevents or delays progression to invasive cancer. "True" increases in overall survival can occur by two mechanisms (in order of the time course of cancer): 1) early detection of cancer by screening tests and subsequent effective treatment and 2) advancements in treatment. Unique patterns or "fingerprints" of stage-specific incidence and overall incidence and of survival rates characterize the various cancer prevention and control mechanisms that can decrease mortality rates. The rates are presented for five organ sites that have shown reduced cancer mortality. The patterns of rates for breast cancer for women under the age of 65 years were most consistent with early detection. The testicular cancer fingerprints were most consistent with advances in treatment, whereas cervical cancer rates were most consistent with the detection of preneoplastic lesions. The stomach cancer fingerprints were indicative of reductions in lifestyle or environmental risks, and colorectal cancer rates were indicative of a combination of treatment advances and early detection. These fingerprint patterns can be extended to other situations in which mortality trends are changing in order to suggest possible causes of observed changes. Limitations of this model are also discussed.

Recent trends in lung cancer mortality in the United States.

BACKGROUND: Previous age-period-cohort analyses of lung cancer incidence and mortality rates in the United States have demonstrated a decrease in risk by birth cohort through 1950, consistent with declining trends in smoking prevalence. This study was conducted to examine recent lung cancer trends, including trends among the cohorts born after 1950. METHODS: Lung cancer mortality rates from 1970 through 1997 for whites aged 24--83 years and for blacks aged 30--83 years were investigated. Using age--period--cohort analyses with 2-year age and 2-year calendar-period intervals, we examined changes in the slope of the trends in birth-cohort and calendar-period effects. All statistical tests are two-sided. RESULTS: There was an unexpected, statistically significant moderation in the rate of decrease of the birth-cohort trend in lung cancer mortality for whites born after 1950, with a corresponding smaller and statistically nonsignificant moderation for blacks. These data are consistent with smoking initiation rates: Rates of both cigarette and marijuana smoking initiation increased for children aged 12--17 years from 1965 through 1977. There was a statistically significant decrease in the slope of the calendar-period trend for lung cancer mortality in 1990 for both whites and blacks that was observed primarily in people 55 years of age and older. CONCLUSIONS AND IMPLICATIONS: The birth-cohort pattern of lung cancer mortality after 1950 appears to reflect the early impact of teenage cigarette smoking on lung cancer risk in people under the age of 45 years, although a contribution from marijuana smoking cannot be ruled out. This result provides additional support for increasing smoking cessation and prevention programs for teenagers. The calendar-period decrease in lung cancer mortality after 1990 may reflect the long-term benefits of reductions in tobacco carcinogens in cigarettes and increases in smoking cessation beginning around 1960.

Differentiating among proposed mechanisms for tumor promotion in mouse skin with the use of the multievent model for cancer.

Numerous mechanisms for tumor promotion in mouse skin have been recently proposed. These include the multistage, genetic view of Slaga, Furstenberger, and Marks; the multistage, genetic and/or epigenetic view of Parkinson; and the two-stage view of Yuspa and Hennings. With the use of the multievent model, a mathematical model for cancer that can incorporate cell proliferation phenomena, these tumor promotion mechanisms were modeled mathematically. It was found that the models had different predictions about the incidence of papillomas in initiator-promoter-initiator experiments and the incidence of carcinomas in initiator-promoter experiments with varying periods of tumor promotion. Upon analysis, existing data were shown to be in agreement with the Yuspa and Hennings two-stage model. The analysis also supported the view, formulated by Scribner and co-workers, that different types of initiated cells were created, some capable of transformation to carcinomas and some not. Phenomena determined by use of benign tumors as the end point may need to be demonstrated with the use of the carcinoma as the end point, prior to the acceptance of these phenomena being applicable to the human carcinogenic process.

Temporal patterns in colorectal cancer incidence, survival, and mortality from 1950 through 1990.

BACKGROUND: Colorectal cancer mortality rates among U.S. white males remained relatively constant from 1950 through 1984 but declined sharply from 1985 through 1990. Those for U.S. white females decreased consistently from 1950 through 1984, with an acceleration of the decline from 1985 through 1990. PURPOSE: A study was planned to investigate patterns in incidence, survival, and mortality rates over time in order to examine possible reasons for the gender difference in mortality trends and for the decrease in the slope of the mortality trends for both males and females in the late 1980s. METHODS: Incidence and survival data from the Connecticut Cancer Registry were examined to investigate the gender differences in mortality rates from 1950 through 1984. Incidence and survival data from the Surveillance, Epidemiology, and End Results (SEER) Program were investigated to examine reasons for the abrupt downturn in mortality rates for both white males and white females beginning around 1985. RESULTS: During the period 1950 through 1984, the colorectal cancer incidence rates in Connecticut increased for males and declined slightly for females. Survival rates were similar for both sexes, increasing on average over 1% per year for both females and males from 1950 through 1984. Examination of SEER data from 1975 through 1990 revealed that for both males and females there were 1) declines in overall incidence and mortality rates beginning in the mid-1980s, 2) steady declines in distant disease incidence rates since 1975, 3) increases in regional disease incidence rates until the early 1980s followed by declines in the late 1980s, and 4) increases in local disease incidence rates until the mid-1980s followed by declines in the late 1980s. Age-period-cohort analyses of mortality rates indicated a statistically significant moderation of colorectal cancer risk with both advancing birth cohorts and recent calendar periods. CONCLUSIONS: The gender differences in colorectal cancer mortality rate trends observed from 1950 through 1984 are due to differences in incidence rate trends between males and females. Declining colorectal mortality rates in the late 1980s for males and females appear to reflect improved early detection. The peaking and subsequent decline of stage-specific incidence rates at later years for successively lower stage indicate sequential stage shifts as cancers are detected increasingly earlier over time. The increased use of sigmoidoscopy and fecal occult blood tests (triggering colonoscopy) appears to have played an important role in reducing colorectal cancer mortality. Improvements in birth cohort trends in risk for colorectal cancer for each sex suggest that lifestyle changes may have also contributed to the steady reductions in colorectal cancer mortality.

Neoplastic and nonneoplastic lesions in aging (C57BL/6N x C3H/HeN)F1 (B6C3F1) mice.

Neoplastic and nonneoplastic lesions in untreated (C57BL/6N x C3H/HeN)F1 (B6C3F1) mice used as controls in carcinogenesis tests were tabulated and evaluated. The most common neoplasms in 2,543 male mice were hepatocellular adenomas and carcinomas. In 2,522 female mice, common tumors were lymphomas, leukemias, pulmonary adenomas and carcinomas, hepatocellular adenomas and carcinomas, and pituitary adenomas. The risk of developing most neoplasms increased with the age of the mouse. Hepatocellular carcinomas metastasized in 12% of the animals with these tumors. Other than lymphomas and leukemias, few other tumors metastasized. Nonneoplastic lesions included cystic hyperplasia of the uterus, nephritis, ovarian and uterine cysts, inflammatory lesions of the lung, mineralization in the brain, and focal hyperplasias in several tissues. The focal hyperplasias in lung and pituitary, adrenal, and thyroid glands were suggestive of the early stages of neoplasia. Comparative aspects of lesions in aging mice and their interpretation in carcinogenesis tests are discussed.

Recent trends in U.S. breast cancer incidence, survival, and mortality rates.

BACKGROUND: Clinical trials have demonstrated that use of mammographic screening and advances in therapy can improve prognosis for women with breast cancer. PURPOSE: We determined the trends in breast cancer mortality rates, as well as incidence and survival rates by extent of disease at diagnosis, for white women in the United States and considered whether these trends are consistent with widespread use of such beneficial medical interventions. METHODS: We examined mortality data from the National Center for Health Statistics and incidence and survival data by extent of disease from the Surveillance, Epidemiology, and End Results Program of the National Cancer Institute, all stratified by patient age, using statistical-regression techniques to determine changes in the slope of trends over time. RESULTS: The age-adjusted breast cancer mortality rate for U.S. white females dropped 6.8% from 1989 through 1993. A significant decrease in the slope of the mortality trend of approximately 2% per year was observed in every decade of age from 40 to 79 years of age. Trends in incidence rates were also similar among these age groups: localized disease rates increased rapidly from 1982 through 1987 and stabilized or increased more slowly thereafter; regional disease rates decreased after 1987; and distant disease rates have remained level over the past 20 years. Three-year relative survival rates increased steadily and significantly for both localized and regional disease from 1980 through 1989 in all ages, with no evidence of an increase in slope in the late 1980s. IMPLICATIONS: The decrease in the diagnosis of regional disease in the late 1980s in women over the age of 40 years likely reflects the increased use of mammography earlier in the 1980s. The increase in survival rates, particularly for regional disease, likely reflects improvements in systemic adjuvant therapy. Statistical modeling indicates that the recent drop in breast cancer mortality is too rapid to be explained only by the increased use of mammography; likewise, there has been no equivalent dramatic increase in survival rates that would implicate therapy alone. Thus, indications are that both are involved in the recent rapid decline in breast cancer mortality rates in the United States.

False-positive and false-negative rates for carcinogenicity screens.

The implementation of a number of chemical carcinogen screening programs has been accompanied by the observation that some screens might have high false-positive error rates. With designs presently used at the National Cancer Institute and historical spontaneous tumor rates based upon control animals in previous experiments, we compute upper bounds on the false-positive error rates for several screening strategies. False-positive results are much less likely to occur at tissue sites with low spontaneous tumor rates; hence the site at which a significant tumor increase occurs is important. There is danger in relying solely upon the finding of statistical significance without incorporating biological knowledge and corroborative evidence such as the presence of a dose-response relationship or experimentally consistent results in different species or sexes. A report by the National Cancer Institute Carcinogenesis Program demonstrates these concepts.

Tumor variants by hormone receptor expression in white patients with node-negative breast cancer from the surveillance, epidemiology, and end results database.

PURPOSE: Hormone receptor expression (presence-positive or absence-negative) may reflect different stages of one disease or different breast cancer types. Determining whether hormone receptor expression represents one or more breast cancer phenotypes would have important paradigmatic and practical implications. METHODS: Breast cancer records were obtained from the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) database. The study included 19,541 non-Hispanic white women with node-negative breast cancer. Standard tumor cell characteristics and breast cancer-specific survival were analyzed by independent estrogen receptor (ER+ and ER-), independent progesterone receptor (PR+ and PR-), and joint ERPR expression (ER+PR+, ER+PR-, ER-PR+, and ER-PR-). RESULTS: Age frequency density plots by hormone receptor expression showed two overlapping breast cancer populations with early-onset and/or late-onset etiologies. Independent ER+ and PR+ phenotype were associated with smaller tumor sizes, better grade, and better cancer-specific survival than ER- and PR- breast cancer types. Joint ERPR phenotype exhibited biologic gradients for tumor size, grade, and cancer-specific survival, which ranked from good to worse for ER+PR+ to ER+PR- to ER-PR+ to ER-PR-. CONCLUSION: Variations of standard tumor cell characteristics and breast cancer-specific survival by hormone receptor expression in white patients with node-negative breast cancer suggested two breast cancer phenotypes with overlapping etiologies and distinct clinical features.

Pattern recognitiion and structure-activity relationship studies. Computer-assisted prediction of antitumor activity in structurally diverse drugs in an experimental mouse brain tumor system.

This paper reports the application of pattern recognition and substructural analysis to the problem of predicting the antineoplastic activity of 24 test compounds in an experimental mouse brain tumor system based on 138 structurally diverse compounds tested in this tumor system. The molecules were represented by three types of substructural fragments, the augmented atom, the heteropath, and the ring fragments. Of the two pattern recognition methods used to predict the activity of the test compounds the nearest neighbor method predicted 83% correctly while the learning machine method predicted 92% correctly. The test structures and the important substructural fragments used in this study are given and the implications of these results are discussed.

Representation of Asian Americans in clinical cancer trials.

PURPOSE: The objectives of this study are to analyze the accrual of Asian Americans to National Cancer Institute (NCI)-supported prevention, screening/diagnosis, and treatment trials and to determine if there is proportional ethnic representation. METHODS: Data were obtained on all participants accrued to ongoing prevention and screening/ diagnosis trials and on all patients accrued to treatment trials from 1994 to mid-1998. In the analysis, the percentage of Asian Americans to the total number of trial participants is calculated. For treatment trials, participants were stratified into five age groups: 0-20 years, 21-44 years, 45-54 years, 55-64 years, and 65 or more years. RESULTS: Asian Americans represented 4.8% of subjects accrued in screening/diagnosis trials, 1.8 to 2.2% of subjects in treatment trials, and 0.9% of subjects in prevention trials. Comparison of treatment trial age groups revealed that younger Asian Americans participate significantly more in treatment trials than older Asian Americans. CONCLUSIONS: Asian American accrual in NCI-supported trials is representative of the cancer burden of Asian Americans in the United States. However, Asian Americans 65+ years are underrepresented. Their full participation in cancer trials is justified.

A new method for the analysis of cohort studies: implications of the multistage theory of carcinogenesis applied to occupational arsenic exposure.

Implications of the multistage theory of carcinogenesis for evaluating the effect of exposure to carcinogens in the workplace are described. This theory predicts different patterns of excess risk related to duration of exposure, age at initial exposure, and follow-up time since exposure stopped, depending upon which stage of the carcinogenic process is affected by the carcinogen, i.e., action at an early stage or a late stage. New statistical methodologies are proposed to examine these patterns and are applied to the lung cancer mortality experience from a cohort study of smelter workers exposed to arsenic. Under this multistage hypothesis, the results indicate that arsenic exerts a definite late stage effect though an additional effect at the initial stage cannot be ruled out. The possibilities of biased conclusions resulting from incomplete exposure histories and lack of smoking information are also discussed as well as implications of these results to experimental animal studies.

Review of experimental carcinogenesis by compounds related to vinyl chloride.

The experimental carcinogenesis results in six compounds related to vinyl chloride are reported. Vinylidene chloride, given by inhalation, was carcinogenic in male CD-1 mice, male CD rats, Sprague-Dawley rats and male Swiss mice. Trichloroethylene, given by gavage and inhalation, was carcinogenic in the B6C3F1 mice. When given by gavage, perchloroethylene was carcinogenic in the B6C3F1 mice, and dichloroethane was carcinogenic in Osborne-Mendel rats and B6C3F1 mice. Dibromoethane, given by gavage and inhalation, was carcinogenic in B6C3F1 mice, F344 rats and Osborne-Mendel rats. Finally, epichlorohydrin was carcinogenic in male Sprague-Dawley rats and B6C3F1 mice.

Stage-shift cancer screening model.

A stage-shift cancer screening model is developed in the context of a randomized controlled trial (RCT) of cancer screening. In the model, detection by screening causes the time of diagnosis of the cancer to be advanced so that either the stage at diagnosis is shifted from one stage to the next lower one or the stage of diagnosis is unchanged but the cancer is diagnosed earlier in the stage. These are called external and internal stage shifts, respectively. At each stage the extent of the external and internal shifts and any associated mortality benefits are estimated. Further, the model allows the interrelationships of these benefits within and between stages to be delineated. This then allows us to better understand the results of the RCT. Data from a completed breast cancer screening RCT are used to illustrate the application of the model and its value in improving our understanding of the trial's results.

A method for partitioning cancer mortality trends by factors associated with diagnosis: an application to female breast cancer.

U.S. cancer mortality data derived from information recorded on death certificates are frequently relied upon as an indicator of progress against cancer. A limitation of this measure is the lack of information pertaining to the onset of disease, such as year-of-diagnosis, age-at-diagnosis, stage of disease at diagnosis and histology of lesions. However, population-based cancer registries collect these types of data and allow the calculation of an incidence-file based mortality rate. This incidence-based mortality rate allows a partitioning of mortality by variables associated with the cancer onset. Breast cancer incidence-based mortality measures are created and compared to mortality rates based on death certificates over a comparable time period. Novel mortality measures, such as mortality rates by stage-at-diagnosis, age-at-diagnosis and year-of-diagnosis, are used to illustrate the value of this approach.