| HOME | HELP | CONTACT US | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Epidemiology and Population Studies: SEER Series |
Surveillance Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, Maryland, USA
Key Words. Cancer statistics • Incidence • Lifetime risk • Multiple primaries • Survival • SEER
Correspondence: Brenda K. Edwards, Ph.D., Division of Cancer Control and Population Sciences, National Cancer Institute, 6116 Executive Blvd., Suite 504, Bethesda, Maryland 20892-8315, USA. Telephone: 301-496-8506; Fax: 301-480-4077; e-mail: edwardsb{at}mail.nih.gov
Received August 8, 2006; accepted for publication September 29, 2006.
Access and take the CME test online and receive 1 AMA PRA Category 1 CreditTM at CME.TheOncologist.com
LEARNING OBJECTIVES
Top
Learning Objectives
Abstract
Introduction
Materials and Methods
Results
Discussion
Disclosures of Potential...
References
After completing this course, the reader will be able to:
|
ABSTRACT |
|---|
|
TopLearning ObjectivesAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosures of Potential...References |
|---|
|
INTRODUCTION |
|---|
|
TopLearning ObjectivesAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosures of Potential...References |
|---|
Trends in population-based cancer rates may be studied by (a) using a comparison of aggregate data for discontinuous time intervals as a snapshot of the U.S. cancer burden or (b) examining yearly rates over a complete time range through modeling or more sophisticated analytical techniques to characterize national trends. Both approaches are used here, with a comparison of conclusions. Past trends in cancer incidence rates, coupled with shifts in the U.S. age structure, are described by changes in the median age of patients diagnosed, which is a summary statistic of age-specific incidence for the population at risk.
The risk of developing cancer during a person’s lifetime and for a fixed period of time starting at selected ages comprises a strategic concept that can affect individual health behavior and medical management of patients undergoing routine clinical care. However, once diagnosed with cancer, prognosis is key with survival influenced by the type of cancer, the extent of disease, and treatment modalities. Improvements in detecting cancer at earlier stages and advances in treatment have yielded an increase in the population of living individuals ever diagnosed with cancer. In addition, with improved outcomes and greater life expectancy, there is an increasing need to consider the development of multiple primary cancers. Long-term high-quality population-based cancer registries such as SEER provide important information on individuals who develop more than one primary cancer during their lifetime.
Projections of the U.S. cancer burden are calculated for 2000 through 2050 and demonstrate that long-term shifts in the age composition of the population will lead to changes in the number of cancer cases diagnosed in various age groups, assuming there are no changes in cancer incidence rates. This growth in the number of future cancer cases affects demands placed on the U.S. health care delivery system.
|
MATERIALS AND METHODS |
|---|
|
TopLearning ObjectivesAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosures of Potential...References |
|---|
presents the central cancer registries that contribute data to the SEER Program, the diagnosis years of participation, the current population size (percentage) of the U.S. population included in SEER geographic regions, and the number of cancer cases diagnosed for 2003. By 1975, all SEER-9 registries were reporting data annually to NCI, and most recent reports for cancer trends are based on diagnosis years 1975–2003.
|
The incidence and death rates presented in this article are expressed as the number of new primary cancers and deaths, respectively, per 100,000 persons at risk per year. Numerators (cases) for incidence rates are derived from the SEER Program, and numerators for death rates are obtained from the National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention [4], which collects data from all state vital health reporting systems. For cancer sites that pertain to one sex only, the population at risk is the sex-specific population (e.g., females for ovarian cancer). Incidence and death rates are age-adjusted according to the 2000 U.S. standard population based on 5-year age groups [5]. Population estimates in the denominator for calculation of incidence and death rates are obtained from the U.S. Census Bureau [6].
SEER incidence and U.S. mortality data are available through 2003 diagnoses and deaths, respectively. When SEER cases are reported to NCI in November of each year, data are approximately 98% complete for all cancer sites combined, although some cancer sites, such as melanoma, are less complete. Cancer registries continuously update their databases, and SEER uses statistical modeling to adjust the observed data for underestimation due to reporting delays estimated from the receipt of data files in subsequent years [7]. Incidence, mortality, and survival statistics are presented for four racial groups (white, black, Asian Pacific Islander, and American Indian/Alaska Native) and one ethnic group (Hispanic).
Trends over a given time interval are summarized with the annual percent change (APC) [5]. The APC is obtained by fitting a regression line through the logarithms of the rates for the given time period using weighted least squares. The slope of the line is tested for significant increases or decreases. In addition, the joinpoint regression model [8] is used to characterize changes in cancer rates over time. Each joinpoint denotes a statistically significant change in trend. For the joinpoint analysis, the overall significance level was set to p = .05, and a maximum of three joinpoints and four line segments were allowed.
The SEER software package DevCan was used to calculate the lifetime risk of developing cancer based on age-specific cancer rates for 2001–2003 diagnoses [9–12] from SEER-17 incidence cases. These rates are converted to the probabilities of developing cancer for a hypothetical population [5].
Cancer survival is estimated in different ways, depending on the intended purpose. Relative survival is calculated by comparing observed survival with expected survival from a set of people with the same characteristics as the patient cohort with respect to age, race, sex, and calendar period [13]. Relative survival estimates the effect of the cancer being considered on survival in the absence of other causes of death. It is always larger than or equal to the observed survival.
Multiple primaries describe diagnosis of two or more independent primary reportable neoplasms [14] in an individual. The number of multiple primary cancers observed in the follow-up period after diagnosis of a first primary cancer was divided by the expected number of cancers to produce a standardized incidence ratio (SIR) [15]. The expected number was obtained by applying age-, sex-, and site-specific incidence rates by calendar year to person-years at risk for each respective cancer case. The analysis cohort included patients from SEER-9 areas who were diagnosed from 1973 through 2003 with 18 common cancers for men and women. Excluded from the analyses were all death certificate-only, autopsy, and individual cases with in situ disease as the first cancer diagnosis. Follow-up times, that is, person-years, were computed from 2 months after diagnosis to avoid counting simultaneous cancers and were censored at the time of death, date last known alive, or December 31, 2003, whichever came first.
The number of new cancer cases by sex for selected cancer sites is projected into the future using age-specific population projections, with a baseline assumption that cancer rates remain at their current level. Population estimates for single-year ages were obtained from the U.S. Census Bureau for the years 2000, 2010, 2020, 2030, 2040, and 2050 [16, 17]. The single-year age- and sex-specific estimated cancer counts were calculated using delay-adjusted incidence rates obtained by applying appropriate delay-adjustment ratios from SEER-9 to the SEER-17 data. Estimated single-year counts were then produced for each type of cancer for diagnosis years 1998 through 2002 and summed to calculate average single-year age-specific incidence rates. The single-year age-specific incidence rates for 1998–2002 were applied to the U.S. Census Bureau population estimates to project total cancer case counts. Counts for all sites combined, colon and rectum, and lung and bronchus cancers were estimated for both sexes. Counts for breast cancer and all sites for women and counts for prostate cancer and all sites for men were based on sex- and site-specific incidence rates. The cancer case counts were summed into four age groups: less than 45 years, 45–64 years, 65–84 years, and 85 years and older.
Complete prevalence is the number of people in a population who are alive on a certain date and who have been diagnosed with cancer at any time in their lives [5]. It differs from incidence in that it considers both newly diagnosed and previously diagnosed persons. Prevalence is a function of both the incidence of the disease and survival.
|
RESULTS |
|---|
|
TopLearning ObjectivesAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosures of Potential...References |
|---|
Table 2 shows the distribution of median age at diagnosis for selected primary cancer sites by 5-year time periods from 1974 through 2003. Comparing the most recent 5-year period (1999–2003) with the earliest period (1974–1978), median age at cancer diagnosis for all sites combined increased by 2 years. Of 44 individual cancer sites, 19 sites (43%) showed an increase in median age at diagnosis of more than 2 years, and five sites (11%) showed a decrease in median age at diagnosis of more than 2 years.
|
Cancers of the cervix uteri, anus, anal canal and anorectum, pleura, and prostate showed a decrease in median age at diagnosis of 4 or more years between the 1974–1978 and 1999–2003 time periods. Kaposi sarcoma showed a dramatic 33-year decrease in median age at diagnosis. Until the early 1980s, Kaposi sarcoma was a rare disease found mainly in older men. During the last 20 years, most cases developed in association with human immunodeficiency virus infection and AIDS. The median age at diagnosis during the time period 1974–1978 was 74 years, compared with a median age at diagnosis of 41 years during the time period 1999–2003.
Age-adjusted incidence rates are shown in Table 3 for 20 cancer types by sex and selected discontinuous 5-year time periods from 1974 to 2003. The description of data in this table is supplemented with conclusions drawn from an examination of yearly data adjusted for reporting delay [7] using joinpoint regression models [8]. Figures 1A and 1B demonstrate the results of the joinpoint regression analysis (i.e., show the years in which rates changed significantly) for the top 10 cancer sites in men and in women. Each bar is divided into the portions of the interval 1975 through 2003 with a change in rates—intervals are color-coded, with red denoting periods with increasing trends, green for decreasing trends, and yellow for periods with stable trends. The significant APCs for each time period are shown within the bar or for the most recent period at the right border of the bar. Joinpoint analysis can be used to quantify changes in incidence rates for specific years, as well as calculate statistically significant trends in time. This method of analyzing the data shows a continuum of change, whereas comparisons of isolated time intervals do not always capture the nature of change over the intervening period.
|
|
shows an increase in the incidence of breast cancer from 1974 to 2003. A more detailed analysis (Fig. 1B) of yearly data indicates a marked increase from 1980 to 1987 (3.7% per year) followed by a lesser increase (0.5% per year) from 1987 through 2001, after which rates began to stabilize from 2001 through 2003. In this example, incidence rates shown in Table 3 present an incomplete picture of the long-term breast cancer trend. Thus, caution should be applied when considering statistics using discontinuous fixed intervals for assessing long-term trends and identifying years in which a change in trends has occurred.
For all sites combined, age-adjusted cancer incidence rates among men increased dramatically until 1992, fell sharply during the next few years, and since 1995 have been stable (Fig. 1A). Among women, however, the cancer incidence rates continued to rise during the period from 1979 to 2003, although the rate of increase diminished after 1987. Incidence rates for leukemia, lung cancer, melanoma, non-Hodgkin lymphoma, and thyroid cancer in women showed an increase during 1975–2003 (Fig. 1B). The analysis for leukemia presented in Table 3 does not convey this long-term increase since adjustment for reporting delay is included in Figures 1A and 1B but not in Table 3 [20]. Delay-adjusted incidence rates include estimates for cancers reported after the initial data collection period for a diagnosis year. Incidence of thyroid cancer began to increase sharply in women early in the 1980s (Table 3; Fig. 1B), and long-term increases in incidence rates have occurred for cancer of the liver and kidney and renal pelvis, as well as Hodgkin lymphoma. Ovarian cancer incidence rates began to decline in 1985 and continued to decline through 2003 (Fig. 1B).
Among men, incidence rates increased (Fig. 1A) during this entire period for cancers of the kidney and renal pelvis, leukemia, and melanoma (which began to stabilize in 2001), but decreased for cancers of the oral cavity and pharynx and for the lung (beginning in 1982). The long-term trends for prostate cancer seem to fluctuate from 1975 to 2003 (Fig. 1A), increasing significantly since the late 1980s and early 1990s (Fig. 1A; Table 3). A sharp decline (Fig. 1A) in prostate cancer incidence occurred between 1992 and 1995, followed by a modest increase in the most recent segment from 1995 to 2003. Increasing trends of liver cancer are at least twice as high in men as in women, whereas the thyroid cancer incidence rate in men is approximately half that observed in women.
Incidence rates by race and ethnic group based on SEER-17 areas are displayed in supplemental online Table A. Sites selected were the top cancers based on 2000–2003 age-adjusted incidence rates for all races combined for the given sex group (male, female, combined). Blacks had the highest incidence rate (504.4 per 100,000) for men and women for all sites combined. Prostate cancer incidence was 258.3 per 100,000 for black men and 163.4 per 100,000 for white men. Incidence rates for men with cancers of the prostate, lung and bronchus, colon and rectum, and oral cavity and pharynx were higher for blacks, whereas cancers of the urinary bladder, melanoma of the skin, and non-Hodgkin lymphoma had higher incidence rates for whites. American Indian/Alaska Native men had the highest incidence rate of kidney and renal pelvis cancer, at 20.9 per 100,000. Among women, incidence rates for cancers of the colon and rectum and of the pancreas were higher for blacks than whites. White women had the highest incidence rates for breast, lung and bronchus, corpus and uterus not otherwise specified (NOS), non-Hodgkin lymphoma, melanoma of the skin, ovary, and thyroid cancers.
Death rates by race and ethnicity are displayed in supplemental online Table B. The sites selected were the top cancers based on 2000–2003 age-adjusted death rates for all races combined for the given sex group (male, female, combined). Blacks had the highest death rates for men and women for all sites combined. Death rates for cancers of the lung and bronchus, prostate, colon and rectum, pancreas, and esophagus in men were highest in blacks. Leukemia, non-Hodgkin lymphoma, and urinary bladder death rates were highest for whites. Asian Pacific Islanders had the highest mortality for liver and intrahepatic bile duct cancer for both men and women (data not shown). Among women, overall death rates for cancers of the lung and bronchus and ovary, and for non-Hodgkin lymphoma and leukemia, were higher in whites than blacks, whereas female mortality for cancers of the breast, colon and rectum, pancreas, and corpus and uterus NOS were higher in blacks than in whites.
Table 4 displays the lifetime risk of developing cancer, assuming that a person is cancer-free at the starting age. Approximately 45% (one in two) men and 38% (one in three) women will be diagnosed with some form of invasive cancer during their lifetime. For a person 45 years of age, the risk of developing cancer during the next 10 years is 1 in 24 (1 in 26 for men; 1 in 22 for women). One in six persons, aged 65 years, will be diagnosed with cancer by age 75 (one in five for men; one in eight for women). Men have a one in six risk of having prostate cancer during their lifetime, and one in eight women will develop breast cancer. The lifetime risk of developing lung and bronchus cancer is 1 in 12 for men and 1 in 16 for women. Childhood cancers are rare; the risk of a newborn child developing cancer by the age of 10 is 1 in 612. By 30 years of age, 1 in 134 persons will develop some form of cancer. Starting at 65 years of age, approximately 43% (one in two) men and 30% (one in three) women will develop cancer.
|
. The overall 5-year relative survival for individuals diagnosed in 1996–2002 from SEER-17 geographic areas was 89% for female breast, 65% for colon and rectum, and 16% for lung and bronchus. For prostate cancer, relative 5-year survival is 49% at stage IV, although at all stages it is 100%. This can happen because relative survival measures exceed mortality relative to the general population, calculated as a ratio of the observed and expected survival.
|
There is a difference in survival between men and women diagnosed with lung cancer. For persons diagnosed with stage I lung cancer, there is a nine-percentage-point higher relative 5-year survival for women compared with men. The sex difference is smaller for later-stage lung cancer but persists for both white and black patients.
For most cancers, relative survival generally is higher for white than for black patients. For female breast cancer, 5-year relative survival for all stages combined is 90% in white women compared with 78% in black women, with the highest difference occurring in stage III disease (5-year relative survival is 62% in whites vs. 45% in blacks). Similarly, there is a 10-percentage-point difference between races for patients diagnosed at all stages combined of colon and rectum cancer. This survival difference between white and black patients is the greatest for stage I disease in men and stage II disease in women, with smaller differences within other stages. Prostate cancer clinical stage IV survival is five percentage points higher in white males than in black males.
Tables 6, 7, and 8 relate to multiple primary cancers. Each table shows 18 primary sites at which cancer first originated. Table 6 shows the number of patients diagnosed with a first primary cancer, the number and percentage of these patients who then develop multiple primaries (within 2 months and after 2 months of an initial diagnosis), median age at diagnosis based on first cancer, and the 5-year relative and observed survival for these individuals with a first cancer diagnosis. Urinary bladder is the initial (or index) primary cancer site with the highest percentage of individuals with multiple primary cancers (16%), followed by oral cavity and pharynx (15%) and corpus and uterus (11%). When compared to all other primaries, liver cancer showed the fewest multiple primary occurrences (only 1%). As expected, there is an association between the risk of developing subsequent cancers and patient survival, with the exception of thyroid cancer.
|
|
|
for males and in Table 8 for females. The initial primaries shown are those with the largest number of multiple primary events. Reporting rules influence the occurrence of multiple primaries, particularly for same-site tumors and tumor pairs affected by treatment that removes the target organ. Statistically elevated SIRs were found for several initial primary cancer sites. Urinary bladder cancer patients had an elevated risk of subsequent cancers of the kidney and other urinary organs (males: SIR = 11.08, p < .05; females: SIR = 17.79, p < .05). For cancers of the oral cavity and pharynx, the highest SIRs were found for multiple primaries of the same site in both sexes; in addition, there was an increased risk for cancer of the esophagus (males: SIR = 7.67, p < .05; females: SIR = 18.71, p < .05). Kidney and renal pelvis cancer patients had an elevated risk of multiple primary cancers at the same or proximal site, as well as in the bladder. Colon and rectum cancer patients had a significant incidence risk of subsequent cancer of the colon. The risk of melanoma as a multiple primary cancer was highest in melanoma patients, who also had an increased risk of a subsequent thyroid cancer.
The risk of multiple primary breast cancers (SIR = 1.55, p < .05) was significantly elevated in female breast cancer patients. An increased risk of subsequent thyroid cancers (SIR = 1.19, p < .05) was observed after an initial diagnosis of breast cancer. Hormonal factors have been suspected to be relevant to the etiology of thyroid cancer as well as breast cancer [21]. For a first cancer of the corpus and uterus, the risk of multiple primaries of the bladder and other urinary organ diseases was elevated (SIR = 1.41, p < .05). Incidence of thyroid cancer after prostate cancer diagnosis was noted in men (SIR = 1.20, p < .05).
Projections of cancer incidence counts for 2000–2050 are displayed in Table 9, and the pattern of estimated total number of cancers for age groups is presented in Figure 2. As baby boomers (persons born from 1946 to 1964) age and move into the 45–64-age bracket by 2010, more new cancers are expected in this age group if incidence rates remain the same or do not decrease. For example, the group of adults aged 45–64 in the male and female all sites combined category shows an increase from 32% in 2000 to 37% in 2010. A similar increase in this age group can be seen for all of the top cancer sites. By 2030, baby boomers will have aged to between 66 and 84 years, the high-risk age group for cancer. This population growth impact is seen in the increase in the number of cancer patients in the 65–84-year age group in the year 2030. This increase is consistent for all of the top cancer sites.
|
|
). This estimate includes individuals who recently were diagnosed with cancer, those who have survived many years after diagnosis, those with active cancer, and those who are cancer-free. Cancer is more prevalent in women, with an estimated 5.8 million having had a diagnosis of cancer by 2003. Men over the age of 70 constitute more than half of all men diagnosed with cancer. More than a million women in the 50–69-year age group and another million women over the age of 70 have been diagnosed with breast cancer. Approximately 95% of the population previously diagnosed with acute lymphocytic leukemia is under the age of 50. The prevalence of prostate cancer is higher in men over the age of 70: nearly 1.3 of the 1.9 million men previously diagnosed with prostate cancer are at least 70 years old.
|
|
DISCUSSION |
|---|
|
TopLearning ObjectivesAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosures of Potential...References |
|---|
Cancer incidence rates and trends are examined here in three ways: median age at diagnosis over time; comparison of summary incidence rates for four separate 5-year intervals spanning the time over which SEER data have been collected; and analysis of yearly incidence rates through modeling (joinpoint) and the determination of statistically significant APCs for trend segments over time. Median age at diagnosis has shifted slightly between 1974 and 2003, showing a small increase in age at diagnosis of about 2 years for cancer at all sites combined. For cancers common in young adults less than 40 years of age (e.g., bones and joints, Hodgkin lymphoma, acute lymphocytic leukemia, testis, and thyroid), the median age at diagnosis for all five sites has increased from 1974–1978 to 1999–2003 by 2–8 years.
Incidence data are frequently compared between two time periods of several years each to obtain sufficient cases for estimating rates. Although this method allows many aspects of the data to be observed, it does not show the detail available in an examination of yearly incidence rates over time. When adjusted for reporting delays, the age-adjusted incidence rates in men, which are primarily influenced by the rapidly changing trends in the diagnosis of asymptomatic prostate cancer, have been stable since 1995. For women, however, incidence rates continue to rise, influenced largely by changes in trends for cancers of the breast and lung. Detailed examination of breast cancer incidence shows stable rates in the most recent time period (2001–2003), preceded by a deceleration in the rate of increase since about 1987. Factors such as the reduction in use of hormone therapy among postmenopausal women that occurred following the 2002 publication of results from the Women’s Health Initiative [22, 23] and population effects of screening practices with mammography [24] that may explain the recent stabilization and possible decline in breast cancer incidence rates are under investigation. Similarly, although incidence rates for female lung cancer are increasing, the rate of increase is slowing. Increases in the incidence rate for thyroid cancer in women are particularly notable, partially due to occurrence after diagnosis of other primary cancers [25].
Our study shows that women diagnosed with lung cancer have better survival than men, especially those diagnosed with local-stage disease. The reasons for this survival advantage have not been identified, but it is likely due to a variety of factors. A previous study [26] analyzing the SEER database suggests surgery was performed more frequently in women than in men with local-stage disease. Since surgery offers these patients the best chance of long-term survival, this difference in treatment may partially explain the superior survival observed in women. Other concomitant factors such as comorbidity can influence treatment decisions for surgery and adjuvant therapy. In addition, men and women smokers differ in the histological types of lung cancer with which they are diagnosed—women are more likely than men to develop adenocarcinoma [27]. This difference in histology may also influence survival. For patients with local disease, Fu et al. [26] reported a negative impact on survival with any histologic types (small cell, squamous cell, and large cell) other than adenocarcinoma after controlling for demographic, temporal, and treatment effects. Further study is needed to evaluate the biological basis and its impact on sex-specific survival differences in lung cancer patients.
The risk of developing cancer is often expressed in terms of lifetime risk, that is, the risk of a newborn developing cancer prior to death. Lifetime risk is confusing to many and assumes current rates hold over a person’s entire lifetime. Shorter risk estimates (e.g., 10 years) conditional on a specific age are often much more understandable and meaningful when communicating risk. Data on risk of developing cancer in future years for men and women at 45 and 65 years of age are likely to be more relevant to patient concerns as they reach ages with established recommendations for cancer screening, as well as to clinicians who provide patient counseling concerning cancer prevention and early detection. In general, the risk of developing cancer increases with age, and differences in short-term risk are more pronounced by age (e.g., at birth, midlife, or retirement) than for long-term risk estimates. The risk of developing invasive cancer is generally higher for men than for women, except for risk associated with cancers occurring during the ages of 20–54 years when women have substantially higher cancer incidence rates than men [5].
Relative survival provides a measure of the excess mortality associated with a cancer diagnosis. The first source of major variation in survival is the particular cancer site. Stage at diagnosis is another major prognostic factor, and the decrease of survival with more advanced disease points to the importance of screening and early detection. However, cancers where early detection is less common, such as liver and pancreatic cancers, are among those with dismal overall relative survival rates. For early-stage lung cancer, a major cancer trial is underway to test the effectiveness of screening with spiral computed tomography imaging to diagnose early-stage disease [28].
In general, black patients have lower relative survival, independent of cancer site and stage at diagnosis. This health disparity may be partially due to variations in the prevalence of risk factors, the use of screening tests for early detection, access to health care services, and/or social and demographic factors [29].
Improvements in cancer survival affect cancer prevalence. As cancer survival improves and more individuals live longer after diagnosis, the prevalent population increases in size. This has led to a greater focus on several areas, such as long-term cancer treatment effects and their complications, issues that come into play in medical follow-up care. In addition, occupational, behavioral, and quality-of-life aspects of cancer survivorship are being addressed [30]. Prevalence provides information on the impact of cancer on the health care system and documents the need for increased research on the needs of survivors.
The SEER Program is the sole source of U.S. population-based data to address issues associated with multiple primary cancers [31], using established rules for collecting and reporting more than one tumor per person. The SEER rules exclude recurrences or tumors of the same histology reporting within 2 months, with exceptions for some tumors such as retinoblastoma, bilateral ovarian tumors, adenocarcinomas of the prostate, certain bladder tumors, and a few others that are reported at one time. Analyses of multiple primaries presented here are based on a total population of over 2 million cancer patients diagnosed with cancer between 1973 and 2003 in the SEER-9 database, yielding approximately 12 million person-years at risk. For some initial cancer sites, a large proportion of multiple primary cancers occur at the same cancer site or organ system as the initial primary. This may be due to exposure, risk factors, or genetic predisposition. A process referred to as "field cancerization" involves effects of carcinogenic exposures or genetic factors over areas of tissues or organs [32]. Statistically significant increased multiple primaries in the same cancer site are reported for oral cavity and pharynx, kidney and renal pelvis, colon and rectum, and melanoma in both sexes, as well as urinary bladder and breast in women. The oral cavity and pharynx site seems to be the most extreme case, with SIRs of 29.5 and 11.5 in women and men, respectively.
Data are presented for various cancer sites, along with the percentage of multiple primaries found among those surviving more than 2 months, in addition to relative and observed 5-year survival among the cohorts. In general, an initial cancer site with a high percentage of multiple primaries must have a relatively high observed survival. For head and neck and urinary tract cancers (including oral cavity and pharynx and urinary bladder, the sites with the highest percentages of multiple primary cancers in Table 6), recent studies have shown evidence of a possible spread of cells from a single clone to multiple sites [33, 34]. Factors involved in multiple primary cancers in separate sites may include the effects of risk factors such as tobacco and alcohol on multiple organs, infections and immunosuppression, genetic predisposition, and treatment effects [32]. Cancer of the thyroid is seen particularly frequently in these data with statistically increased SIRs associated with initial primary cancers of the oral cavity and pharynx, kidney and renal pelvis, melanoma, and prostate in men and of the kidney and renal pelvis, melanoma, and breast in women. As a shared environmental etiology is unlikely for some cancer pairs such as melanoma and thyroid cancer, a likely explanation at present for any relationship between these two tumor types continues to be increased medical surveillance [35]. Information on multiple primaries is important to clinicians and cancer patients during medical management following initial treatment as well as in efforts to minimize iatrogenic effects of cancer therapy identified through analyses of multiple primaries.
Projections into the future (through 2050) of the number of cancer cases and percentage of cases in four age groups show the effects of the changing age structure of the U.S. population. In this projection, with the cancer incidence rates held constant to current rates, the percentage of cancer patients for all sites combined in the 65–84-year age group increases to a peak in the year 2030 and then drops, whereas the number and percentage of cancer patients in the oldest age group continually increases. The population 85 years and older is the fastest-growing age sector of cancer patients. Our analysis indicates that, assuming constant cancer incidence rates, the number of cancer patients is expected to more than double from 1.36 million in 2000 to almost 3.0 million in 2050, due to aging and the growing U.S. population. This projection illustrates how cancer counts in a given age range may increase even when the incidence rate remains unchanged, due to the age structure of the population.
Substantial progress is being made against cancer for many, but not all, cancer sites and among many population groups, but not with equal effectiveness. These advances, however, have implications for health services needs in the future. As the prevalent population grows and more individuals live longer after a diagnosis of cancer, their medical care needs change. It is increasingly important to examine long-term treatment effects and to be aware of issues of multiple primary cancers. The population of individuals being treated for cancer is also aging, and projections for U.S. population dynamics will continue to age through the next half century. Thus, the patient population may have increased comorbid conditions that will affect both treatment decisions and healthcare during survivorship. We look to the SEER Program data as a continued source of information to chart these changing patterns as well as other aspects of the U.S. national cancer burden [36, 37].
|
DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST |
|---|
|
TopLearning ObjectivesAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosures of Potential...References |
|---|
|
ACKNOWLEDGMENT |
|---|
|
TopLearning ObjectivesAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosures of Potential...References |
|---|
|
REFERENCES |
|---|
|
TopLearning ObjectivesAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosures of Potential...References |
|---|
This article has been cited by other articles:
 |
|
 |
|
  L. R. Henry, N. P. Solomon, R. Howard, J. Gurevich-Uvena, L. B. Horst, G. Coppit, R. Orlikoff, S. K. Libutti, A. R. Shaha, and A. Stojadinovic The Functional Impact on Voice of Sternothyroid Muscle Division During Thyroidectomy Ann. Surg. Oncol., July 1, 2008; 15(7): 2027 - 2033. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. A. Azim, M. Costantini-Ferrando, and K. Oktay Safety of Fertility Preservation by Ovarian Stimulation With Letrozole and Gonadotropins in Patients With Breast Cancer: A Prospective Controlled Study J. Clin. Oncol., June 1, 2008; 26(16): 2630 - 2635. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Wheatley-Price, K. Ding, L. Seymour, G. M. Clark, and F. A. Shepherd Erlotinib for Advanced Non-Small-Cell Lung Cancer in the Elderly: An Analysis of the National Cancer Institute of Canada Clinical Trials Group Study BR.21 J. Clin. Oncol., May 10, 2008; 26(14): 2350 - 2357. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Leyvraz, S. Pampallona, G. Martinelli, F. Ploner, L. Perey, S. Aversa, S. Peters, P. Brunsvig, A. Montes, A. Lange, et al. A Threefold Dose Intensity Treatment With Ifosfamide, Carboplatin, and Etoposide for Patients With Small Cell Lung Cancer: A Randomized Trial J Natl Cancer Inst, April 16, 2008; 100(8): 533 - 541. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. S LEGG Bladder Cancer Imaging Radiol. Technol., March 1, 2008; 79(4): 333 - 346. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Demichelis and M. A Rubin TMPRSS2-ETS fusion prostate cancer: biological and clinical implications J. Clin. Pathol., November 1, 2007; 60(11): 1185 - 1186. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||
