Cancer statistics, 1996

We present incidence, mortality, and survival statistics to provide a perspective on the patterns of cancer occurrence in the United States population. Estimates of the numbers of new cancer cases and deaths for 1996 are presented according to sex, site, and state. We also present information on cancer and noncancer mortality, the probability of developing cancer at certain ages, and cancer survival in adults and children. Because no nationwide cancer registry exists, there is no way of knowing exactly how many new cases of cancer are diagnosed annually in the United States. We use cancer incidence data collected by the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) program and US population data collected by the Bureau of the Census to estimate the number of new cancer cases that are expected to be diagnosed in the coming year (Table 1).1–3 The SEER program currently tabulates cancer data from nine population-based cancer registries in the following states and metropolitan areas: Connecticut, Hawaii, Iowa, New Mexico, Utah, San Francisco-Oakland, Detroit, Seattle, and Atlanta. These registries, which cover about 10 percent of the US population, are useful for monitoring the occurrence of cancer for the nation.1 We calculated the 1996 estimates of new cancer cases using a two-step procedure. First, we applied incidence rates from the SEER program for 1979–1992 to the US Census Bureau's population projections for the same years to estimate the number of cancer cases diagnosed yearly from 1979 through 1992.1–3 We fitted these annual estimates to a quadratic function that is used to project an estimate of the number of new cancer cases expected to be diagnosed in 1996.4,5 Some additional adjustments were made for selected sites with recent variations in the estimated number of cases diagnosed. Cancer sites that were adjusted include the mouth, prostate, and thyroid in males and the eye, corpus and unspecified uterus, and selected other and unspecified sites in females. The estimated number of new cancer cases for each state was calculated using 1996 estimates of new cancer cases for the United States and 1996 estimates of cancer deaths for each state (Table 3). For each cancer site, state estimates of new cancer cases were calculated by apportioning the number of new cases for the United States as a whole according to the distribution of estimated state cancer deaths in 1996 (see mortality section for additional information on the estimation process). Mortality data on the number of cancer deaths that occur in the United States each year are collected and compiled by the National Center for Health Statistics (NCHS).6 We calculated the US estimates of cancer mortality for 1996 by fitting the number of cancer deaths reported for 1979–1992 to a quadratic function that is used to project an estimate of the number of cancer deaths expected to occur in 1996 (Table 2).4,5 Estimates of state-specific cancer deaths were calculated using the same model fitted with the actual number of reported deaths occurring in each state from 1979 through 1992 (Table 4). Some additional adjustments were made for selected sites with recent variations in the estimated number of deaths. Cancer sites with such adjustments include other and unspecified sites in males and stomach, cervix uteri, and other and unspecified sites in females. The reported number of deaths in 1992 that we present in this report were compiled using the most recent data available from NCHS (Tables 6–9 and 12). Mortality rates were calculated using number of deaths reported to NCHS and US population data from the Bureau of the Census (Tables 7 and 12, Figs. 3 and 4). All rates are standardized to the age distribution of the 1970 Census population. Age standardization is a statistical method used to remove the effects of age differences between populations for comparison purposes.7 We present US mortality data on the 10 leading sites of cancer death for minority populations (Table 10). Reported cancer deaths are based on the underlying cause of death as coded on death certificates for whites, African Americans, Native Americans, Asians and Pacific Islanders, and Hispanics in 1992.6 Cancer deaths among Hispanic persons are presented only for the 48 states that record Hispanic origin on death certificates. In 1990 these states accounted for about 99.6 percent of the Hispanic population in the United States.8 Estimated probabilities of developing invasive cancers at certain ages were calculated by applying age-specific incidence and mortality rates from the SEER program for the years 1990–1992 to a hypothetical group of 10 million persons (Table 5).1,9 For each five-year age interval from ages 0–4 through ages 95 and older, the number of persons developing a specific cancer and the number of persons dying from other causes were calculated. The probability of being diagnosed with a specific cancer during a given age interval was estimated by dividing the number of persons developing cancer in that interval by the number of persons alive and free of that cancer at the beginning of the interval. The lifetime probability of developing a specific cancer was estimated by summing all cancer cases that occurred in the hypothetical group from ages 0 through 95 and older and dividing by 10 million. This procedure for estimating interval and lifetime cancer risk does not assume that all persons live to the end of an age interval or to any fixed age. It also does not take into account individual behaviors and risk factors. For example, the estimated 5.38 percent of females likely to develop lung cancer is a low estimate for smokers and a high estimate for persons who do not smoke. Five-year relative survival rates are commonly used to monitor progress in the early detection and treatment of cancer. The rates are calculated by dividing the survival rate observed among a group of cancer patients by the rate for persons in the general population who are similar with respect to age, sex, race, and calendar year of observation. Five-year relative survival rates are reasonable indicators of the average survival experience of cancer patients in a given population, but they are less informative when used to predict individual prognosis. The interpretation of five-year relative survival rates is difficult because the rates are based on patients whose treatment reflected methods of medical practice at least eight years ago and also because an increase or decrease in survival may be caused by several factors, including changes in early detection techniques and in treatment strategies. We present five-year relative survival rates by site, race, and stage at diagnosis for cases diagnosed during the years 1986–1991 (Fig. 6), as well as trend data from 1960 through 1991 for adults and children (Tables 11 and 13).1,10 The relative survival rates are based on the follow-up of patients through 1993 as reported to the SEER program. International mortality rates were calculated from data made available by the World Health Organization (WHO) (Table 14), age-adjusted to the WHO standard world population.11 Mortality data from China were reported for certain urban and rural areas in the eastern half of the country and included about 10 percent of the total Chinese population. Our estimates of new cases and deaths have limitations and should be interpreted with caution when used to study patterns in the occurrence of cancer. Although the estimates are based on the most recent data, they are computed before the year begins and based on data that are at least three years old. Other sources of cancer data may be helpful for interpreting these data. Incidence and survival statistics are based only on invasive cancers, except for cancer of the urinary bladder, which includes carcinoma in situ.1 The number of deaths for minorities is likely to be underestimated due to underreporting of Asian, Pacific Islander, and Native American race and Hispanic ethnicity on death certificates.8 Figure 1 1996 Estimated New Cancer Cases, United States Percent Distribution of Sites by Sex* Figure 2 1996 Estimated Cancer Deaths, United States Percent Distribution of Sites by Sex* Age-Adjusted Cancer Death Rates* Females by Site, United States, 1930–1992 Age-Adjusted Cancer Death Rates* Males by Site, United States, 1930–1992 Figure 5 Percent of Cancer Cases by Stage at Diagnosis United States, 1986–1991 Five-Year Relative Survival Rates (Percent) by Race and Stage at Diagnosis, United States, 1986–1991