Annual Report to the Nation on the Status of Cancer, Part 1: National Cancer Statistics

Abstract Background The American Cancer Society, Centers for Disease Control and Prevention, National Cancer Institute, and North American Association of Central Cancer Registries collaborate to provide annual updates on cancer incidence and mortality and trends by cancer type, sex, age group, and racial/ethnic group in the United States. In this report, we also examine trends in stage-specific survival for melanoma of the skin (melanoma). Methods Incidence data for all cancers from 2001 through 2017 and survival data for melanoma cases diagnosed during 2001-2014 and followed-up through 2016 were obtained from the Centers for Disease Control and Prevention- and National Cancer Institute-funded population-based cancer registry programs compiled by the North American Association of Central Cancer Registries. Data on cancer deaths from 2001 to 2018 were obtained from the National Center for Health Statistics’ National Vital Statistics System. Trends in age-standardized incidence and death rates and 2-year relative survival were estimated by joinpoint analysis, and trends in incidence and mortality were expressed as average annual percent change (AAPC) during the most recent 5 years (2013-2017 for incidence and 2014-2018 for mortality). Results Overall cancer incidence rates (per 100 000 population) for all ages during 2013-2017 were 487.4 among males and 422.4 among females. During this period, incidence rates remained stable among males but slightly increased in females (AAPC = 0.2%, 95% confidence interval [CI] = 0.1% to 0.2%). Overall cancer death rates (per 100 000 population) during 2014-2018 were 185.5 among males and 133.5 among females. During this period, overall death rates decreased in both males (AAPC = −2.2%, 95% CI = −2.5% to −1.9%) and females (AAPC = −1.7%, 95% CI = −2.1% to −1.4%); death rates decreased for 11 of the 19 most common cancers among males and for 14 of the 20 most common cancers among females, but increased for 5 cancers in each sex. During 2014-2018, the declines in death rates accelerated for lung cancer and melanoma, slowed down for colorectal and female breast cancers, and leveled off for prostate cancer. Among children younger than age 15 years and adolescents and young adults aged 15-39 years, cancer death rates continued to decrease in contrast to the increasing incidence rates. Two-year relative survival for distant-stage skin melanoma was stable for those diagnosed during 2001-2009 but increased by 3.1% (95% CI = 2.8% to 3.5%) per year for those diagnosed during 2009-2014, with comparable trends among males and females. Conclusions Cancer death rates in the United States continue to decline overall and for many cancer types, with the decline accelerated for lung cancer and melanoma. For several other major cancers, however, death rates continue to increase or previous declines in rates have slowed or ceased. Moreover, overall incidence rates continue to increase among females, children, and adolescents and young adults. These findings inform efforts related to prevention, early detection, and treatment and for broad and equitable implementation of effective interventions, especially among under resourced populations.

The Centers for Disease Control and Prevention (CDC), the American Cancer Society, the National Cancer Institute (NCI), and the North American Association of Central Cancer Registries (NAACCR) have collaborated annually since 1998 to provide updated information about cancer occurrence and trends by cancer type, sex, age group, and racial/ethnic group in the United States. Part 1 of this report focuses on national cancer statistics and highlights trends in stage-specific survival for melanoma of the skin, the first cancer for which effective immune checkpoint inhibitors were developed (1). Part 2 focuses on the economic burden of cancer in the United States (2).

Data Sources
Cancer Incidence Data. Population-based cancer incidence data by age, sex, and race and ethnicity were obtained from registries that participate in the CDC's National Program of Cancer Registries and/or the NCI's Surveillance, Epidemiology, and End Results (SEER) Program. Only registries whose data satisfied NAACCR's criteria for data quality and completeness were included in this study (3). For rate analyses, 49 states and 1 territory (Puerto Rico) met data criteria for every year during 2013-2017, and for trend analyses, 46 states met data criteria for every year during 2001-2017, representing 99% and 92% of the population of the United States and Puerto Rico, respectively. States included in any analysis in this report are listed in corresponding figure legends and table footnotes.
Anatomic site and histology were coded according to International Classification of Diseases (ICD) for Oncology 3rd edition (4) and categorized according to SEER site groups (5). Only cases of cancer defined as malignant were included in this report, except that in situ and malignant bladder cancers were combined when reporting bladder cancer incidence rates.
Cancer Mortality Data. Cause of death by age, sex, race, and ethnicity (2001-2018) from all 50 states and the District of Columbia was based on death certificate information reported to state vital statistics offices and compiled through CDC's National Center for Health Statistics' (NCHS) National Vital Statistics System (6). The underlying causes of death were selected according to ICD-10, then categorized according to SEER site groups to maximize comparability with ICD-O classifications (5).
Survival Data for Melanoma of the Skin. Survival data for cases of malignant melanoma of the skin (melanoma) diagnosed from 2001 through 2014 were from data compiled by NAACCR from registries in 28 states-covering 86% of the US population-considered to have sufficient vital status follow-up to conduct survival analyses during the entire study period (7) meeting the NAACCR standard, that is, death ascertainment through the study cutoff date of December 31, 2016, or follow-up dates on or after January 1, 2017, for a minimum of 90% of patients (8). Cases were censored at an achieved age of 100 years. Cases identified by death certificate or autopsy only or without survival data were excluded from the survival analysis.
Demographic Characteristics. Rates and trends are presented by sex, racial/ethnic group, and age (all ages, children aged 0-14 years, and adolescents and young adults [AYA] aged 15-39 years). Information about race and ethnicity was collected separately and was based on information abstracted from medical records for incidence in cancer registries or death certificates from NCHS for mortality. In this report, information about race and ethnicity was combined to create 5 mutually exclusive racial and ethnic (racial/ethnic) groups: non-Hispanic White (White), non-Hispanic Black (Black), non-Hispanic American Indian or Alaska Native (AI/AN), non-Hispanic Asian or Pacific Islander (API), and Hispanic (any race). Race information for AI/ AN was considered reliable only for geographic areas covered by the Indian Health Service Purchased/Referred Care Delivery Areas; thus, to minimize racial misclassification for AI/AN, incidence and mortality data for this group were based only on counties covered by Indian Health Service Purchased/Referred Care Delivery Areas in states that provided county-level information (11). Persons with other or unknown race or unknown ethnicity were included in overall rates but were not included as separate categories.

Statistical Analysis
Cross-sectional incidence (2013-2017) and death (2014)(2015)(2016)(2017)(2018) rates for all ages combined, children, and AYA by cancer type and racial/ethnic group were calculated using SEER*Stat software, version 8.3.6 (12). All rates were age-standardized to the 2000 US standard population and were expressed per 100 000 standard population. Corresponding 95% confidence intervals (CIs) were calculated as modified gamma intervals and allow for informal comparisons between groups, without specifying a referent group. Rates based on fewer than 16 cases or deaths during the 5-year period were deemed to be statistically unreliable and were suppressed. Because delays in reporting of cancer cases to cancer registries can cause incidence rates to be underestimated, all case counts and incidence rates were adjusted for reporting delay (13). Incidence and death rates and trends are reported for males and females for each cancer type that ranked in the top 15 incident cancers or causes of cancer death for any racial/ethnic group. For children, incidence and death rates and trends are presented for the 3 and 2 most common cancer types, respectively; these numbers for AYA were 6 and 4, respectively.
Temporal trends in delay-and age-adjusted incidence (2001-2017) and age-adjusted death (2001-2018) rates were estimated using joinpoint regression (14). A maximum number of 3 joinpoints (4 line segments) were allowed for both incidence and deaths. Annual percent change (APC) characterizes the slope of a single segment, and average APC (AAPC), a summary measure over a fixed interval. Two-sided statistically significant (P < .05) differences from zero were determined using a t test for the APC and for the AAPC when it laid entirely within the last joinpoint segment and a z-test when the last joinpoint fell within the last 5 years of data. When the slope of the trend (APC or AAPC) was statistically significant, the trend was considered increasing (slope >0) or decreasing (slope <0). Trends based on fewer than 10 cases or deaths in any of the data years were considered statistically unreliable and were suppressed. Corresponding 95% confidence intervals for trends were calculated using the parametric method and allow for informal conservative comparisons between groups.
Two-year age-standardized relative survival for melanoma cases diagnosed during 2001-2014 and followed-up through 2016 was calculated based on complete dates and monthly intervals using the Ederer II actuarial method (8,12). We chose 2-year relative survival to have multiple data points to examine trends in survival following the introduction of new therapies, first of which was approved by the Food and Drug Administration (FDA) in 2011 (15). Annual trends in 2-year age-standardized relative survival and 95% confidence intervals were estimated fitting a proportional hazard joinpoint model to survival data on the log hazard scale using the NCI's JPSurv software (16), with a maximum of 2 joinpoints (3 line segments). All survival estimates were age-standardized using the International Cancer Survival Standards, age standard 2, and age groups 15-44 years, 45-54 years, 55-64 years, 65-74 years, and 75 years and older (8).

Cancer Incidence Rates and Trends
The overall cancer incidence rate (per 100 000 population) in 2013-2017 was 448.3, with rates higher in males (487.4) than in females (422.4) ( Table 1). During the most recent 5 years (2013-2017), incidence rates were stable in both sexes combined and among males but increased slightly among females (AAPC ¼ 0.2%, 95% CI ¼ 0.1% to 0.2%). Longer-term trends (2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017) in cancer incidence rates also varied by sex ( Figure 1; Table 2 By racial/ethnic group, overall cancer incidence rates in the most recent 5 years (2013-2017) were stable among White males and Black females; decreased among Black, API, AI/AN, and Hispanic males; and increased among White, API, AI/AN, and Hispanic females ( Table 1). The overall cancer incidence rate was similar among White persons and Black persons, and incidence rates in these 2 groups were higher than rates in other racial/ethnic groups ( Figure 2). The overall cancer incidence rate was higher among males than females in every racial/ethnic group, except API, where the rates were similar. Overall incidence rates were higher among Black males than White males and higher among White females than Black females.
During 2013-2017, incidence rates among males increased for 5 of the 18 most common cancers: melanoma, kidney and renal pelvis (kidney), pancreas, oral cavity and pharynx, and testis; were stable for 7 cancers: liver and intrahepatic bile duct (liver), myeloma, prostate, esophagus, leukemia, non-Hodgkin lymphoma, and thyroid; and decreased for 6 cancers: lung and bronchus (lung), larynx, urinary bladder (bladder), stomach, colon and rectum (colorectum); and brain and other nervous systems (ONS) (Figure 3; Table 1). Trends for the 3 most common cancers among males were similar by racial/ethnic group: prostate cancer incidence rates were stable, whereas lung and colorectal cancer incidence rates decreased in all racial/ethnic groups, with greater declines among Black males (Table 1). Incidence rates for the fourth-most common cancer (bladder) declined in White, API, and Hispanic males, were stable among Black males, and increased among AI/AN males.
Among females, incidence rates increased during 2013-2017 for 8 of the 18 most common cancers: liver, melanoma, kidney, myeloma, corpus and uterus, not otherwise specified (uterus), pancreas, breast, and oral cavity and pharynx; were stable for 4 cancers: cervix, leukemia, stomach, and non-Hodgkin lymphoma; and decreased for 6 cancers: thyroid, ovary, lung, colorectum, bladder, and brain and ONS ( Figure 3; Table 1). However, liver cancer incidence rates among females stabilized during 2014-2017 (Table 2). For the 4 most common cancers among females, during 2013-2017, breast and uterine cancer incidence rates increased in every racial/ethnic group, whereas lung cancer incidence rates decreased among White, Black, and Hispanic females but were stable among API and AI/AN females. Colorectal cancer incidence rates decreased among White, Black, and AI/AN females but were stable among API and Hispanic females (Table 1). Similar to trends among males, declines in lung and colorectal cancer incidence rates were greater in Black females than females in other racial/ethnic groups.

Cancer Death Rates and Trends
The overall cancer death rate (per 100 000 population) in 2014-2018 was 155.5, with a higher rate in males (185.5) than in females (133.5) (Figure 2; Table 3 Figure 1; Table 4).
During 2014-2018, overall cancer death rates decreased in every racial/ethnic group; the average annual pace of decrease ranged from 0.7% among AI/AN persons to 2.1% among API persons. The overall cancer death rate (per 100 000 population) was highest among Black persons (182.5), followed by rates among AI/AN (163.2), White (160.2), Hispanic (110.8), and API persons (98.0) ( Figure 2). Death rates among males during 2014-2018 increased for 5 of the 19 most common cancers (bones and joints, oral cavity and pharynx, soft tissue including heart, brain and ONS, and pancreas); were stable for 3 cancers (liver, nonmelanoma skin, and prostate); and decreased for 11 cancers ( Figure 3; Table 3). Death rates among females increased for 5 of the 20 most common cancers (uterus, liver, brain and ONS, pancreas, and soft tissue including heart), were stable only for oral cavity and pharynx cancer, and decreased for 14 cancers.
The 3 most common cancer deaths among males were lung, prostate, and colorectal cancer in most racial/ethnic groups except API males, in whom lung cancer was the most common cancer death, followed by cancer of the liver, colorectum, and prostate (Table 3). During 2014-2018, lung cancer death rates among males decreased in each racial/ethnic group. Prostate cancer death rates were stable among White, Black, and Hispanic males but decreased among API and AI/ AN males. Colorectal cancer death rates were stable among AI/ AN males but decreased in all other racial/ethnic groups. Pancreatic cancer was the fourth-most common cancer death ARTICLE Table 1. Age-standardized, delay-adjusted incidence rates and fixed-interval trends (2013-2017) for the most common cancers, all ages, by sex, age group, and racial/ethnic group, for areas in the United States with high-quality incidence data a among White and Black males, and death rates increased among White males but were stable in other racial/ethnic groups. Liver cancer was the fourth-most common cancer death among AI/AN and Hispanic males (and second among API males); death rates increased among White and AI/AN males, were stable among Black and Hispanic males, and decreased among API males. Among females, the 3 most common cancer deaths were lung, breast, and colorectal in each racial/ethnic group, except among Hispanic females, in whom breast cancer was the most common and lung cancer was the second-most common ( Table 3). Cancer death rates for these 3 cancers among females decreased during 2014-2018 in every racial/ethnic group, except for breast cancer death rates among API and AI/AN females and colorectal cancer death rates among AI/AN females, in whom rates were stable. Pancreas cancer was the fourth-most common cancer death among females in each racial/ethnic group. Although overall death rates of pancreas cancer slightly increased among females, there was no statistically significant increase in any racial/ethnic group. The largest increases in death rates among females during 2014-2018 were observed for uterine cancer (AAPC ¼ 2.0%), with increasing rates among White, Black, API, and Hispanic females; the AAPC for AI/AN females could not be calculated.
In joinpoint analysis of death rates during 2001-2018, trends for several cancer types changed, notably for melanoma and cancers of the lung, kidney (males only), urinary bladder (males only), colorectum, female breast, prostate, and liver ( Table 4). Some of the recent changes in death rates were favorable. Declines in lung cancer death rates accelerated; among males, Recent changes in death rates for several cancer types were consistent with changes in incidence rates, including decelerated declines in incidence rates for colorectal cancer and stabilized trends for liver cancer and prostate cancer ( Table 2). Deceleration of declines in female breast cancer death rates coincided with a steady, slight increase in incidence rates during 2004-2017, which was preceded by a declining trend during 2001-2004. However, in contrast to declining death rates, melanoma incidence rates increased, and declines in lung cancer death rates outweighed declines in incidence rates. Carolina, South Dakota, Texas, Utah, Vermont, Washington, West Virginia, Wisconsin, and Wyoming. Scattered points were observed rates; lines were fitted rates according to joinpoint regression. a The APC or AAPC is statistically significantly different from 0 (2-sided P < .05), using a t test for the APC and for the AAPC when it laid entirely within the last joinpoint segment and a z-test when the last joinpoint fell within the last 5 years of data. The 95% confidence limits are given in parentheses.

Cancer Among Children
Among children aged 0-14 years, the incidence rate for all cancers combined was 16.8 cases per 100 000 standard population, ranging from 12.6 among AI/AN children to 17.8 among White children (Table 1). Overall cancer incidence rates increased during 2013-2017 (AAPC ¼ 0.7%, 95% CI ¼ 0.5% to 0.9%). The increase occurred in all racial/ethnic groups except among AI/AN children, in whom rates were stable ( Table 1). The most common cancer types included leukemia (5.2 cases per 100 000 standard population), brain and ONS (3.8), and lymphoma (1.6), with increasing trends of 0.7% to 0.8% per year on average for each of these cancers during 2001-2017 (Table 2). Leukemia rates showed the most variability among racial/ethnic groups, ranging from 3.2 cases per 100 000 standard population among Black children to 6.2 among Hispanic children (Table 1). Leukemia incidence rates increased during the most recent 5 years (2013-2017) among White, Black, AI/AN, and Hispanic children but were stable among API children.
The cancer death rate among children was 2.1 deaths per 100 000 standard population, ranging from 1.8 among API children to 2.6 among AI/AN children ( Table 3). The overall cancer death rate among children decreased during 2014-2018 (AAPC ¼ À1.4%, 95% CI ¼ À1.7% to À1.1%). The most common cancer deaths were from brain and ONS cancer (0.7 deaths per 100 000 standard population) and leukemia (0.5 per 100 000 standard population). During 2001-2017, death rates among children for brain and ONS cancer were stable and death rates from leukemia declined an average of 2.9% per year (Table 4).

Cancer Among AYA
Among AYA aged 15-39 years, the overall cancer incidence rate was 75.9 cases per 100 000 standard population, ranging from 55.6 among API AYA to 84.4 among White AYA ( Table 1). The most common cancer among AYA was female breast cancer (22.6 per 100 000 standard population), ranging from a rate of 17.7 among AI/AN AYA to a rate of 27.1 among Black AYA. The next most commonly diagnosed cancers were thyroid (12.1 per 100 000 standard population) and testicular cancer (11.0 per 100 000 standard population), with substantial variations in incidence rates by racial/ethnic group, being lowest among Black AYA (5.6 for thyroid and 2.6 for testis cancer) and highest among White AYA (13.8 for thyroid and 13.4 for testis cancer). Overall cancer incidence rates among AYA increased during 2001-2017 (APC ¼ 0.9%, 95% CI ¼ 0.8% to 1.0%), as did incidence rates of testicular cancer, whereas rates decreased for lymphoma and melanoma (Table 2). There were variations in trends during 2001-2017 for cancers of the colorectum, female breast, and thyroid. The annual percent increase in AYA colorectal cancer incidence rates almost tripled from 1.8% during 2001-2011 to 5.5% during   18-year period 2001-2018 for mortality. a AAPCs were statistically significantly different from zero (2-sided P < .05), using a t test when the AAPC laid entirely within the last joinpoint segment and a z-test when the last joinpoint fell within the last 5 years of data, and are depicted as solid-colored bars; AAPCs with hash marks were not statistically significantly different from zero (stable). NOS ¼ not otherwise specified. .31 6.0 (5.9 to 6.2) 1.2 (0.9 to 1.5)

<.001
Brain and other nervous system  .24 Brain and other nervous system .02 ---À1.9 (À2.9 to À0.9) <.001 Brain and other nervous system The cancer death rate among AYA was 8.9 deaths per 100 000 standard population and was highest among Black AYA (11.4 per 100 000 standard population) and AI/AN AYA (10.7 per 100 000 standard population) and lowest among API AYA (6.8 per 100 000 standard population) ( Table 3). The most common cancer deaths among AYA were from female breast (2.2 deaths per 100 000 standard population), brain and ONS (1.0 per 100 000 standard population), leukemia (0.9 per 100 000 standard population), and colorectal cancer (0.9 per 100 000 standard population). Death rates from female breast cancer among Black AYA

Discussion
The decline in overall cancer death rates in the United States has accelerated in recent years, largely driven by accelerated declines in lung cancer death rates. Other favorable recent changes include a decline in melanoma death rates and stabilization (among males) and deceleration (among females) of earlier increasing trends in liver cancer death rates. In contrast, earlier declines in colorectal and female breast cancer death rates have slowed down, and declining trends for prostate cancer have stabilized in recent years. The observed trends in cancer death rates reflect changes in cancer risk factors (notably, cigarette smoking), screening, and treatment (17). In contrast to death rates, overall cancer incidence rates continue to increase among females and AYA and have stabilized among males after earlier declines, largely reflecting changes in cancer risk factors, notably increases in excess body weight (18). Changes in diagnostic practices have also influenced incidence trends for certain cancers, such as thyroid and prostate (19,20).
Death rates for multiple cancer types and overall remain higher among Black persons than in other racial/ethnic groups. The continuing disparity largely reflects a combination of multiple intertwined factors of tumor biology, stage at diagnosis, receipt of timely and effective treatment, and systemic discrimination in cancer care delivery (21,22). Furthermore, largely owing to social determinants of health inequalities, Black persons and individuals of lower socioeconomic groups in general are more likely to have a higher exposure to some cancer risk factors and limited access to healthy food, safe places for physical activity, and evidence-based cancer preventive services (17,23). Broad and multifaceted interventions can help close the racial mortality gap and address system failures across the continuum of care (24)(25)(26).
Historical declines in cigarette smoking have been followed by declines in incidence and death rates for several smokingrelated cancers, including lung, bladder, and larynx. Moreover, as reported in previous studies (27,28) and observed in this study, lung cancer death rates have declined at a faster pace compared with declines in incidence rates since the mid-2000s, and the difference in the pace of decline has become greater in more recent years. This pattern corresponds to the timing of approval of targeted therapies and other advances in care for the most common subtype of lung cancer, non-small-cell lung cancer (NSCLC), which have resulted in increases in survival and accelerated declines in death rates for this subtype (28). In contrast, declines for small-cell lung cancer, with no approved novel treatments during our study period (29), have been similar to declines in incidence rates, with no improvement in survival rates for this subtype (28).
The first targeted therapy for NSCLC was approved by the FDA in 2003 (30), followed by the approval of several other targeted treatments, recommendations for genetic mutation testing of all individuals with nonsquamous NSCLC for relevant genetic targets in 2012 (28), and the approval of 3 immune checkpoint inhibitors for NSCLC in 2015-2016 (31). Increases in the proportion of individuals with NSCLC who receive definitive therapy (32), better understanding of treatment options for older persons with NSCLC (33)(34)(35), other recent advances in care for NSCLC, such as adjuvant therapy (36) and maintenance therapy (37), and increased access to care through the Medicaid expansion (38) may also have contributed to the acceleration in declines in lung cancer death rates. Any contribution of earlier detection through lung cancer screening (39) to declines in lung cancer death rates is likely to be modest, given low use of screening (4.4% of eligible individuals in 2015) (40). Despite substantial declines in death rates, lung cancer is still the leading cause of cancer death in both sexes, accounting for more than 22% of all cancer deaths in the United States (27), underscoring the need for broader implementation of tobacco control interventions.
In contrast to the rapid decline in incidence rates of smoking-related cancers, incidence rates continued to increase for several cancers, particularly those associated with excess body weight (41), such as cancers of the female breast, corpus uteri, pancreas, kidney, and myeloma (rates rising among females only). The staggering rise of obesity (42,43) and total sitting time (44) continues among both adults and youth. A parallel rise in death rates occurred for pancreatic and uterine cancers, and the earlier declines in death rates slowed for female breast and colorectal cancers, suggesting that for these cancers, increases in incidence rates are of sufficient magnitude to outweigh improvements in survival. Importantly, a recent study of trends by 5-year age groups showed steeper increases in incidence rates among progressively younger ages and successively younger generations for cancers of the corpus uteri, pancreas, kidney, colorectum, and multiple myeloma (18), foreboding of future burden of these cancers in the decades to come.
The declines in colorectal cancer incidence and death rates reflect changes in risk factors (such as reduced cigarette smoking), improvements in clinical management of the disease (affecting mortality only), and increases in colorectal cancer screening (45), which could reduce both mortality and incidence , because precancerous polyps can be detected and removed via colonoscopy (46). Declines in colorectal cancer incidence and death rates accelerated in the 2000s (47), largely owing to a substantial increase in the uptake of colorectal cancer screening among individuals aged 50-75 years, from 33.9% in 2000 to 58.7% in 2010, with colonoscopy use increasing from 19.1% in 2000 to 54.9% in 2010 (48). The deceleration of declines in colorectal cancer incidence and death rates in more recent years coincided with slower increases in uptake of screening since 2010, approaching 59.2% for colonoscopy and 62.4% for any screening test in 2015 (48). Moreover, in contrast to the overall declining trends, both incidence and death rates were rising in adults younger than 50 years, likely reflecting changes in risk factors such as obesity (18). The American Cancer Society lowered its recommended age for initiating colorectal cancer screening for average-risk individuals from 50 to 45 years of age in 2018 (46), and the US Preventive Services Task Force (USPSTF) released a draft statement in October 2020 that included a Grade B recommendation for screening individuals aged 45-49 years (49).
Female breast cancer is the most commonly diagnosed cancer among women and among AYA, and the second leading cause of cancer death among women. The increase of breast cancer incidence is largely driven by hormone-receptor positive cancer (50,51), which may in part reflect continuing reduction of parity rates (52,53), advanced age at first birth (53), the obesity epidemic (postmenopausal breast cancer) (52), high levels of physical inactivity, and increase of alcohol consumption (54,55). As noted in the last year's report (56), the decline of breast cancer death rates overall has slowed since 2007, and even more so since 2014, and rates have stabilized among young women since 2010, halting the progress achieved through early detection and improved treatments. Moreover, substantial racial disparities persist, with death rates 40% higher among Black women than White women despite similar incidence rates.
Prostate cancer incidence rates stabilized from 2014 to 2017 after sharply declining during 2007-2014. These trends coincide with changes in use of prostate specific antigen (PSA) testing following the USPSTF recommendations against PSA-based screening for prostate cancer for men aged 75 years and older in 2008 (57) and for all men in 2012 (58). PSA testing for screening reasons sharply declined from 2007 to 2013 and remained unchanged from 2013 to 2015 (20,59). A recent analysis of prostate cancer incidence by stage at diagnosis, however, showed that rates continued to decline for local-stage disease, whereas rates increased substantially for regional and distant stage diseases (60). The increases in regional-and distant-stage diseases may have contributed to the recent stabilization of death rates after years of declining trends. Indeed, analysis of stage-specific prostate cancer death rates in the SEER areas showed that distant-stage prostate cancer death rates increased between 2012 and 2017, whereas rates for local-stage disease declined (61). Increases in prevalence of obesity (42), a suspected risk factor for fatal prostate cancer (41), may also have contributed to the recent stabilized mortality trend. Effects of the USPSTF recommendation for informed decision for PSA-based prostate cancer screening for men aged 55-69 years in 2018 (62) on prostate cancer incidence and death rates are yet to be determined.
Long-term increasing trends in uterine cancer death rates parallel trends in incidence, although death rates are increasing at a somewhat faster rate. Increasing uterine cancer incidence has been attributed to increasing obesity prevalence and decreased use of combined hormone replacement therapy (63,64).
In our study, and as previously observed (63), incidence rates are similar for Black and White women, whereas uterine cancer death rates are twice as high in Black women. Similar to many other cancers, the disparity in death rates is associated with later stage at diagnosis and poorer quality of care (65) as well as higher frequency of nonendometrioid subtypes (which are more aggressive and have poorer prognosis) among Black women (66,67). Lack of correction for variation in hysterectomy rates can obscure overall trends and differences in incidence by racial/ethnic group, because the prevalence of hysterectomy in the United States has changed over time (eg, from 27.3% in 2000 to 23.9% in 2015) and varies by race/ethnicity (eg, 27.6% among Black women and 15.3% among Asian women in 2015) (67). A recent study, which analyzed hysterectomy-corrected uterine cancer incidence rates, found that rising rates are largely a result of the rapid increase in nonendometrioid subtypes among all racial/ethnic groups, which may in part explain slightly faster increases in uterine cancer death rates, although incidence rates of endometrioid subtypes, which account for about 75% of all uterine cancer cases, have also increased among Black, API, and Hispanic women (66). In addition to interventions to help women achieve and maintain a healthy body weight and physical activity level, promoting awareness of the importance of timely evaluation of abnormal vaginal bleeding has been shown to decrease the proportion of women diagnosed at more advanced stages (63).
Increasing trends in kidney cancer incidence rates in part reflect the obesity epidemic (18) and an increase in incidental detection of indolent or early-stage kidney cancers following a rise in the use of advanced imaging (68,69). Increases in the detection of indolent or early-stage cancer may also have contributed to declining trends in kidney cancer death rates, because prognosis for cancers confined to kidney is favorable, with a 5-year relative survival of 93% during 2009-2015 (27). Progress in treatment, such as targeted therapy for advanced kidney cancer, the first of which was approved by the FDA in 2005, may also have contributed to declining kidney cancer death rates (70). However, it is unclear why kidney cancer death rates declined in an accelerated rate during 2015-2018 among males. Similarly, bladder cancer death rates declined at an accelerated rate during 2013-2018 only among males. Declines in smoking, a main risk factor for bladder cancer (71), likely explain the slight, steady decline in bladder cancer death rates among females, because the declines parallels the incidence trends. The first new treatment for bladder cancer in almost 3 decades was approved by the FDA in 2016, likely with limited impact on survival in our study period (72).
As noted in the last year's report (56), earlier increases in liver cancer death rates have decelerated among females and stabilized among males in more recent years. However, whereas incidence rates of liver cancer were increasing in last year's report, rates have stabilized in both sexes in this year's report. Because more than two-thirds of liver cancer cases are attributable to potentially modifiable risk factors (71), trends in liver cancer occurrence may largely reflect changes in risk factors. Historically, the birth cohort born during 1945-1965 has had substantially higher prevalence of hepatitis C virus infection, notably among men, and higher age-specific liver cancer death rates compared with other birth cohorts (73,74). Increases in liver cancer incidence and death rates in the past several decades coincided with the aging of the 1945-1965 birth cohort and a rise in obesity and type 2 diabetes (75). Several factors may have contributed to the stabilization or deceleration of increases in liver cancer incidence and death rates, including continued decline in liver cancer rates in API population (76), the introduction and dissemination of treatment for chronic hepatitis C virus infection over the past decade (77), and advances in liver cancer treatments (affecting mortality only) (78). However, use of these interventions and their impact varies across subpopulations, for example, by race/ethnicity or socioeconomic status (79,80), and the overall 5-year relative survival for liver cancer remains poor (33% for those diagnosed during 2009-2015) (27).
Last year's report found that 5-year incidence trends for thyroid cancer had stabilized among both males and females after increasing for several decades. This year, for the first time, 5year incidence rates are statistically significantly decreasing 2.0% per year among women of all racial/ethnic groups. Thyroid cancer incidence rates among AYA, which had been increasing, have now stabilized. However, incidence rates of advancedstage thyroid cancer (81) and larger papillary thyroid cancers of classical variant (size !1 cm) have slightly increased in recent years (82), likely due to the obesity epidemic. As discussed in last year's report, declines in overall thyroid cancer incidence are likely attributable to changes in diagnostic practices for lowrisk tumors (19). A small proportion of the decline during 2015-2017 has been attributed to diagnostic coding changes for follicular variant of papillary thyroid carcinoma (82).
There are some other notable trends in this study. Reasons for declining incidence rates but increasing deaths rates for cancers of the brain and ONS are unclear and need further research. For cancer of the oral cavity and pharynx, incidence rates in both sexes and death rates among males are increasing. As shown in previous studies, the increasing incidence trends are limited to cancers in subsites with a strong association with human papillomavirus infection (such as tonsil and oropharynx), likely due to changes in sexual practices, whereas rates are declining in other subsites, largely due to declines in smoking prevalence (83). Similarly, incidence rates of anal and vulvar cancers-both associated with human papillomavirus-are increasing (84,85). In contrast, because of widespread dissemination of screening using the Papanicolaou test, the incidence rate of cervical cancer in the United States has been declining for several decades (85,86). However, studies by histological subtype have indicated that the decline has mainly been limited to cervical squamous cell carcinoma, the most common type of cervical cancer, whereas the incidence of cervical adenocarcinoma, which is less likely to be detected by Papanicolaou test compared with squamous lesions, has shown increasing trends among White women aged 30-59 years in recent years (87). Increases in incidence of childhood cancers overall and for major cancer types may be in part due to increases in exposure to risk factors, including obesity and radiation (88)(89)(90), although changes in classification of diseases or diagnostic procedures may also have contributed to increasing trends in some pediatric cancers (89).
As noted in the last 2 annual reports and in other studies, death rates for cutaneous melanoma have declined rapidly in recent years following introduction of new therapies, including targeted and immune checkpoint inhibitors, the first of which was approved by the FDA in early 2011 (15,56,91,92). In this year's report, we present trends in stage-specific survival from 28 states and find increases in 2-year relative survival beginning in 2009 for distant-stage disease (3.1% per year). An increase in survival for distant-stage melanoma has also been documented in the CONCORD-3 study, in which 1-year net survival for cases diagnosed during 2001-2010 in the United States was stable at 43%, but it started to increase from 2010, reaching 56.6% for cases diagnosed in 2013 (93). The increase in late-stage melanoma survival in the United States slightly preceded the FDA approval of new therapies, likely because of the administration of these therapies through clinical trials and the FDA expanded access programs before the approval (93). A populationbased study of treatment patterns and survival among patients in Ontario, Canada, documented increasing use of systemic therapy and new drugs from 2007 to 2015 (94). Two-year survival increased from 15% among those diagnosed in 2007-2009 to 35% for those diagnosed in 2014-2015; survival gains were most marked for patients treated with systemic therapy, which increasingly involved new drugs over the time period. Among these patients, 2-year survival increased from 16% in 2007-2009 to 44% in 2014-2015 (94).
The gains in survival for patients with metastatic melanoma are important indicators of potential progress in cancer treatment by introduction of new targeted and immune therapies. Recent research in melanoma treatment has focused on combination immunotherapies (95), optimal sequencing and combination of BRAF-MEK and immune checkpoint inhibitors, understanding the mechanisms of primary and acquired resistance to these therapies, and biomarkers for treatment selection and monitoring (96). Although advances in treatment of latestage melanoma are promising, continuing efforts in primary prevention can reduce melanoma incidence rates, which are increasing among males and females. Recent analyses of incidence trends among non-Hispanic White persons show decreasing trends among younger adults, likely reflecting public health and policy interventions (97).
Results of this study represent national cancer trends, because mortality data cover the entire US population, incidence data cover 99% for incidence rates during 2013-2017 and 93% for incidence trends during 2001-2017, and melanoma survival data cover 86%. However, the study has several limitations. First, cancer rates for AI/AN, API, and Hispanic populations may be underestimated due to misclassification of data on race and ethnicity (98), and single-race population estimates derived from the original multiple-race categories may lead to some uncertainties about the population estimates (10). Moreover, data generally are available only for broad and heterogeneous racial/ethnic groups, whereas these groups may include subpopulations with very distinct cultural and health profiles. Second, although temporal trends for some cancer types may vary by histological or molecular subtype (50,51,99), we did not examine these patterns because they are beyond the scope of this report. Third, state cancer registries may have different methods of follow-up; however, these differences are unlikely to have substantially affected the findings in melanoma survival trends.
Cancer death rates in the United States continue to decline overall and for many cancer types. Furthermore, declines in death rates have accelerated for lung cancer and melanoma, likely owing to advances in treatment. However, earlier progress has slowed for several common cancers, death rates are increasing for several other cancers, and incidence rates are increasing slightly among females and in younger ages. These results inform ongoing and future efforts in prevention, early detection, and treatment and broad implementation of effective interventions, especially among under resourced populations.

Funding
This work was supported by the American Cancer Society, the Centers for Disease Control and Prevention, the National Cancer Institute, and the North American Association of Central Cancer Registries. The American Cancer Society is a not-for-profit public health organization that receives support from the public through fundraising and direct contributions. The Society also receives a small portion of support from corporations and industry to support its mission programs and services.

Notes
Role of the funder: The funding institutions had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclosures: FI, AJ, JM, HS, KRY, and JZ are employed by the American Cancer Society, which receives grants from private and corporate foundations, including foundations associated with companies in the health sector for research outside the submitted work. The authors are not funded by any of these grants and their salary is solely funded through American Cancer Society funds, and they have nothing else to disclose. The other authors have no conflicts of interest to disclose.
Author contributions: All authors: Conception and design. All authors: Data analysis and interpretation. FI, EMW, HS, and AJ: Drafted initial manuscript. All authors: Drafting, revising, editing manuscript.
Disclaimer: The findings and conclusions in this article are those of the authors and do not necessarily represent the official positions of the authors' agencies (the American Cancer Society, the Centers for Disease Control and Prevention, the National Cancer Institute, or the North American Association of Central Cancer Registries).