The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Lung cancer is the second most commonly diagnosed cancer and the leading cause of cancer-related death in the United States (US). We examined data from 2004 to 2006 for lung cancer incidence rates by demographics, including race and geographic region, to identify potential health disparities.
Data from cancer registries affiliated with the Centers for Disease Control and Prevention's (CDC) National Program of Cancer Registries (NPCR), and the National Cancer Institute's (NCI) Surveillance, Epidemiology, and End Results Program (SEER) were used for this study; representing 100% of the US population. Age-adjusted incidence rates and 95% confidence intervals for demographic (age, sex, race, ethnicity, and US Census region), and tumor (stage, grade, and histology) characteristics were calculated.
During 2004 to 2006, 623,388 people (overall rate of 68.9 per 100,000) were diagnosed with lung cancer in the US. Lung cancer incidence rates were highest among men (86.2), Blacks (73.0), persons aged 70 to 79 years (431.1), and those living in the South (74.7). Among Whites, the highest lung cancer incidence rate was in the South (75.6); the highest rates among Blacks (88.9) and American Indians/Alaska Natives (65.4) in the Midwest, Asians/Pacific Islanders in the West (40.0), and Hispanics in the Northeast (40.3).
Lung cancer is the second most commonly diagnosed cancer in men and women, and the leading cause of cancer-related death in the United States (US).1 As 1 of the most lethal cancer diagnoses, lung cancer has a 5-year survival rate of just 15%.2 Racial and ethnic differences in lung cancer incidence, survival, and mortality have been reported in previous studies.3, 4 Many have shown that US Black populations have higher rates of lung cancer incidence and mortality, compared with US Whites.5-9 Although national cancer registries have consistently reported low lung cancer incidence among American Indian (AI)/Alaska Native (AN) and Asian (A)/Pacific Islander (PI) populations,10 studies have shown that incidence rates in these populations have been underreported, and the rates are disproportionately high compared with the general US population.11, 12 This report is the first to utilize a nationally comprehensive dataset that captures 100% of all lung cancer diagnoses across the US, therefore providing accurate representation for all racial/ethnic groups.
Racial and ethnic differences in lung cancer incidence may be associated with a variety of risk factors and exposures among various subgroups. Although multiple risks have been identified, cigarette smoking (and tobacco use) constitutes the greatest risk for lung cancer onset.13-18 Other risk factors have also been linked to lung cancer, including genetic mutations and environmental exposures, such as radon.19 Radon is a colorless, odorless gas that can become trapped in buildings; it is the second leading cause of lung cancer overall, and the leading cause of lung cancer in nonsmokers.20 A number of risk factors (eg, smoking prevalence, biological effects of carcinogenic exposure, access to appropriate medical care, and disease perception) contribute to disparities in lung cancer incidence and mortality across US subpopulations, and have been shown to disproportionately affect underserved populations.21-30
National datasets have been used to show regional differences in smoking prevalence across the states.31-33 Reports indicate that smoking prevalence among Whites is highest in the South; among Blacks and Hispanics, it is highest in the Midwest; and among A/PI, smoking is highest in the West.34, 35 Regional differences in smoking prevalence may contribute to disparities in lung cancer incidence. Variation in US lung cancer incidence by geographic region has been reported in 38 states,36 as well as within AI/AN and Appalachian subpopulations.11, 37 Here, we present the first comprehensive analysis of lung cancer incidence covering the entire US population. Our findings will help public health practitioners by identifying racial, ethnic, and geographic subgroups that have disproportionately high lung cancer incidence and may particularly benefit from implementation of evidence-based tobacco prevention and cessation efforts.
MATERIALS AND METHODS
Included in our analyses are data on new lung cancer cases diagnosed from 2004 to 2006 and reported to population-based cancer registries affiliated with the Centers for Disease Control and Prevention's (CDC's) National Program of Cancer Registries (NPCR) and the National Cancer Institute's (NCI's) Surveillance, Epidemiology, and End Results (SEER) program. Both of these registries collect data using uniform procedures and data standards,38 which are then combined to form the US Cancer Statistics (USCS) dataset. Data were reported to NPCR as of January 31, 2010, and to SEER as of November 1, 2009. Data submitted from all 50 states and Washington, DC, met quality standards for publication in the USCS39 for the entire period of 2004 through 2006, and it was sufficient to represent 100% of the US population. We included only cases in which the primary site of cancer was lung or bronchus cancer (C34) (World Health Organization's International Classification of Diseases for Oncology, 3rd edition [ICD-O-3]). Morphology codes for cases diagnosed before 2001 were converted from ICD-O-2 to ICD-O-3.40
We used the following racial and ethnic categories: AI/AN, A/PI, Black, White, Hispanic, and non-Hispanic. Hispanic ethnicity is not mutually exclusive from race categories. The race variable incorporates Indian Health Service (IHS)-linked AI/AN data. IHS patient registration data are electronically linked to identifying patient information in the cancer registry database to reduce racial misclassification of AI/AN persons.41 To identify Hispanic cancer cases, an algorithm developed by the North American Association of Central Cancer Registries (NAACCR) was used. The algorithm combines data items (eg, surname, maiden name, birthplace) to classify persons as Hispanic.42
Because the American Joint Committee on Cancer (TNM) staging system is not available in this dataset, we used the SEER Summary Stage 2000. This staging system is routinely used by cancer registries, and incorporates all information from the medical record including tumor size, nodal involvement and metastasis (http://www.seer.cancer.gov/tools/ssm/). We present the SEER Summary Stage 2000 categories of localized, regional (4 subcategories), distant, and unstaged in this study. Briefly, localized tumors are defined as a single tumor confined to 1 lung, regional is defined as: 1) regional not otherwise specified (NOS), 2) direct extension to proximal organs, 3) ipsilateral lymph nodes, or 4) extension to both proximal organs and ipsilateral lymph nodes. Distant tumors are characterized as tumors that have metastasized or directly extend to distant sites (lymph nodes and/or organs) at the time of diagnosis.
Lung cancer cases were further grouped by the following main histology groups based on ICD-O-3 morphology codes: Nonsmall-cell carcinoma (8011-8015, 8046, 8050-8084, 8140-8384, 8440-8490, 8560), small cell carcinoma (8041-8045), and epithelial carcinoma, NOS (8010, 8016-8040). Nonsmall-cell carcinomas were further categorized as squamous cell carcinoma (8050-8084), adenocarcinoma (8140-8384, 8440-8490, 8560), large cell carcinoma (8011-8015), and nonsmall-cell carcinoma, NOS (8046).
Analyses were conducted using SEER*Stat version 6.6.2 software.43 Frequency counts and percentages are presented for sex, histologic type, age group (<40 years, 40-49 years, 50-59 years, 60-69 years, 70-79 years, and 80+ years), race, Hispanic ethnicity, US Census region, and tumor laterality, stage, and grade. US Census regions are defined by the US Census Bureau.44 With the exception of age-specific rates, rates are age-adjusted to the 2000 US standard population with 19 age groups (Census 25-1130). Ninety-five percent confidence intervals for all age-adjusted rates were calculated with the Tiwari modification.45 Rates, frequency counts and percentages for histologic type, stage, grade, and laterality were limited to microscopically confirmed tumors. The denominators for all rates included estimated yearly population data from the US Census Bureau, which were modified slightly by NCI to enhance population estimates of the Native Hawaiian population in Hawaii.46 Adjustments in population data were made by the US Census Bureau to account for the Gulf Coast population displaced, in 2005, from Alabama, Mississippi, Louisiana, and Texas by Hurricanes Katrina and Rita.47
In the United States, 623,388 people (overall rate, 68.9 per 100,000) received lung cancer diagnoses from 2004 to 2006 (Table 1), of whom 340,457 (86.2) were men and 282,931 (56.2) were women. Incidence rates were highest among persons aged 70 to 79 years (431.1), followed by 80+ years (361.9), 60 to 69 years (250.6), 50 to 59 years (83.1), 40 to 49 years (22.5), and <40 years (0.8). Blacks had the highest incidence rates (73.0), followed by Whites (69.5), AIs/ANs (46.1), and A/PIs (37.1). Compared with non-Hispanics (71.5), Hispanics (36.0) had lower lung cancer incidence rates. Across the US, lung cancer incidence was highest in the South (74.7), followed by the Midwest (72.3), Northeast (68.0), and West (56.5) (Table 1). At time of diagnosis, the majority of lung cancers were diagnosed as distant (30.6). Nonsmall-cell carcinoma (48.7) was the most common histologic type of diagnosed lung cancer, with the most common subtype as adenocarcinoma (22.1) (Table 1).
Table 1. Demographic Characteristics for Lung Cancer Incidence by Sex, United States, 2004-2006a,b,d
68.9 (68.7, 69.1)
86.2 (85.9, 86.5)
56.2 (56.0, 56.4)
Abbreviation: NOS, not otherwise specified.
New cases are diagnosed per 100,000 persons.
dAnalysis limited to microscopically confirmed tumors.
eEthnicity not mutually exclusive of race.
fAll persons are of either Hispanic or non-Hispanic ethnicity; Hispanics might be of any race.
Regional (direct extension and ipsilateral regional lymph nodes)
4.8 (4.8, 4.9)
6.1 (6.1, 6.2)
3.8 (3.8, 3.9)
30.6 (30.4, 30.7)
39.1 (38.9, 39.3)
24.1 (23.9, 24.2)
3.7 (3.7, 3.8)
4.7 (4.6, 4.8)
3.0 (2.9, 3.0)
48.7 (48.5, 48.8)
61.4 (61.1, 61.6)
39.2 (39.0, 39.4)
22.1 (22.0, 22.2)
24.7 (24.6, 24.9)
20.3 (20.2, 20.4)
Squamous cell carcinoma
12.8 (12.8, 12.9)
18.8 (18.6, 18.9)
8.3 (8.2, 8.3)
Large cell carcinoma
2.2 (2.1, 2.2)
2.8 (2.8, 2.9)
1.6 (1.6, 1.7)
Nonsmall-cell carcinoma, NOS
11.6 (11.5, 11.7)
15.1 (14.9, 15.2)
9.0 (8.9, 9.1)
Small cell carcinoma
9.0 (8.9, 9.1)
10.3 (10.2, 10.4)
8.1 (8.0, 8.2)
Epithelial carcinoma, NOS
1.9 (1.9, 2.0)
2.6 (2.5, 2.6)
1.5 (1.5, 1.5)
All other cancers
1.0 (1.0, 1.0)
1.3 (1.2, 1.3)
0.8 (0.8, 0.9)
Overall, Black men had the highest incidence rates (104.5 per 100,000) among men, and White women had the highest incidence rates (57.6) among women (Table 1). Among Black, A/PI, and Hispanic populations, incidence rates in men were approximately twice as high, compared with women (Table 1). Blacks had higher incidence rates for all groups <70 years of age (Table 2). Among individuals 40 to 49 years of age, the highest rates were among Blacks (31.8), followed by Whites (21.7), AIs/ANs (13.4), and As/PIs (10.8). In the same age group, Hispanics (8.5) had lower rates when compared with non-Hispanics (24.4). The proportion of distant lung cancer diagnoses was highest in A/PIs (58.8%), followed by Blacks (54.5%), AIs/ANs (51.4), and Whites (50.0%). Hispanics (55.9%) had a higher proportion of distant lung cancer at time of diagnosis, compared with non-Hispanics (50.4%).
Table 2. Demographic Characteristics for Lung Cancer Incidence by Race/ Ethnicity, United States, 2004-2006a,b
We observed substantial variation in lung cancer incidence by geographic region among racial and ethnic groups. Among Whites, the highest lung cancer incidence rate was in the South (75.6); rates among Blacks (88.9) and AIs/ANs (65.4) were highest in the Midwest, and rates among As/PIs were highest in the West (40.0) (Fig. 1). Incidence rates among Hispanics were highest in the Northeast (40.3) (Fig. 1). Incidence rates for Blacks were substantially higher than all other racial groups in the Midwest and West regions (Fig. 1).
This study includes cancer registry data from all 50 states and Washington, DC, and represents 100% of the entire country. We present the most recent data on lung cancer incidence, tumor characteristics, sex, race, and ethnicity. Overall, we found a disproportionate lung cancer burden in the Black population. Blacks had higher incidence rates and were diagnosed at younger ages compared with other racial groups. We also found regional variation in lung cancer incidence rates among racial/ethnic groups. Rates for all races/ethnicities combined were highest in the South, and geographic regions with the highest incidence rates varied across racial/ethnic populations.
The increased burden of lung cancer in minority populations may be linked to a variety of factors. The higher lung cancer incidence rates observed among Blacks may be due to increased susceptibility to smoking-induced lung cancer,48 disadvantaged access to health care services,49 and lower success rates in smoking cessation attempts,50 compared with the general US population. In concurrence with prior reports on lung cancer incidence among AI/AN populations,51, 52 we found relatively low lung cancer incidence rates among these populations, despite reports of high prevalence of cigarette use.53, 54 The discrepancy in smoking prevalence and lung cancer incidence may be due to cultural preference for tobacco use, racial misclassifications, or incomplete registry data.11 Our findings are in agreement with a previous study that reported low lung cancer rates among As/PIs.12 Low lung cancer incidence among As/PIs may be partially explained by the low national smoking prevalence among this population.34 However, although lung cancer incidence is generally low in A/PI populations, studies have shown variation in smoking prevalence and cancer incidence rates among A/PI subgroups.12, 55 The relatively low lung cancer incidence rates in Hispanics may be partially associated with low smoking prevalence among Hispanic populations.34
Although lung cancer incidence has been shown to vary by geographic region in Appalachian37 and AI/AN populations,11 we report a variation in lung cancer incidence for Black, White, AIs/ANs, As/PIs, and Hispanic populations across US Census regions. Regional variations in lung cancer incidence may be explained by many factors, including regional differences in smoking prevalence,31, 34 environmental exposures,16 socioeconomic status,56 and missed opportunities to address tobacco use in state or tribal comprehensive cancer control plans.57 Our findings are consistent with published reports that smoking prevalence varies considerably across the US. There is evidence that smoking prevalence is highest among Whites in the South, among Blacks, Hispanics, and AI/AN in the Midwest, and among As/PIs in the West.11, 34, 35 These regional and racial variations in tobacco use may explain some of our lung cancer incidence findings. Among Blacks, regional variations in smoking prevalence have been associated with residential migration, social stress, racism, targeted tobacco advertisements and marketing ploys, cultural influences, and community structure.35 Differences in state-specific tobacco control efforts36, 57, 58 may also contribute to regional variation in smoking prevalence.
This study is subject to at least 3 limitations. First, information about smoking status is not consistently collected among the cancer registry data; therefore, we are unable to attribute lung cancer incidence to smoking tobacco within our dataset. Second, racial and ethnic data in registries are of varying quality for AIs/ANs and Hispanics.59 Therefore, persons of AI/AN or Hispanic descent may not be accurately reported herein. Finally, because of the lag time between exposure and lung cancer, and the mobility of our population, persons exposed in 1 US census region may be diagnosed in another, thus, displacing the cancer burden from region to region.
Despite these limitations, the monitoring of incidence rates for lung cancer has improved. Conducting analyses using data from both NCPR and SEER extends cancer surveillance to provide a complete picture of lung cancer in the US and has facilitated calculation of reliable national incidence rates for all racial and ethnic minority groups. Our findings highlight the need to extend evidence-based cancer control efforts to people with or at risk of lung cancer, including tobacco control and other methods that will decrease lung cancer disparities.
To reduce the overall lung cancer burden in the US, we must improve access to and affordability of health care, while increasing research funding to address lung cancer and related health disparities.60 In addition to implementing population-based interventions, further surveillance on lung cancer incidence and smoking prevalence within subpopulations in various geographic regions of the US is warranted. Understanding of regional and racial-ethnic differences in lung cancer incidence may contribute to more effective preventive and treatment strategies that will ultimately reduce the disproportionate burden of lung cancer in the US.
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
All authors have read and approved the manuscript, and there are no financial disclosures, conflicts of interests, and/or acknowledgements necessary.