Trends analysis of cancer incidence, mortality, and survival for the elderly in the United States, 1975–2020

Abstract Background Cancer burden from the elderly has been rising largely due to the aging population. However, research on the long‐term epidemiological trends in cancer of the elderly is lacking. Methods Registry data of this population‐based cross‐sectional study were from the Surveillance, Epidemiology, and End Results (SEER) database. The study population aged 65 years or more, from geographically distinct regions. Joinpoint regression and JP Surv method were used to analyze cancer trends and survival. Results Mortality rate during 1975–2020 decreased from 995.20 to 824.99 per 100,000 elderly persons, with an average annual decrease of 0.421% (95% CI, 0.378–0.464). While overall incidence increased with no significance. Prostate (29%) and breast (26%) cancer were the most common malignancies, respectively, in elderly males and females, and the mortality for both of the two (prostate 15%, breast 14%) ranked just behind lung and bronchus cancer, which had the highest mortality rates in males (29%) and females (23%). Many cancers showed adverse trends in the latest follow‐up periods (the last period calculated by the Joinpoint method). For intrahepatic cholangiocarcinoma, incidence (male Annual Percentage Change [APC] = 7.4*; female APC = 6.7*) and mortality (male APC = 3.0*; female APC = 3.3*) increased relatively fast, and its survival was also terrible (3‐year survival only 10%). Other cancers with recent increasing mortality included cancer of anus, anal canal and anorectum, retroperitoneum, pleura, peritoneum, etc. Most cancers had favorable trends of survival during the nearest follow‐up period. Conclusion Against the background of overall improvement, many cancers showed adverse trends. Further research for the underlying mechanisms and targeted implements towards adverse trends is also urgent.


| INTRODUCTION
Cancer had been the second cause of death just behind heart diseases in the United States (US) during 2015-2020. 1,2It is estimated that there will be 1,958,310 new cancer cases and 609,820 cancer deaths in the US in 2023, and over half of the newly diagnosis and over 70% deaths are elderly patients (defined as persons aged 65 years or more in this study). 3With the extension of human life expectancy and aging population, the cancer burden would become heavier in this age group, raising growing health concerns and social panic. 4,5In addition, elderly patients were more likely to develop cancer complications, which added to the risk of death. 6Therefore, it is necessary to monitor cancer epidemiology in the elderly and make accurate strategies to reduce future burdens.
As is well known, age is a strongly positive-related hazard factor for most cancers' occurrences. 7,8And there is evidence that elderly patients were often more advanced at diagnosis and also less likely to receive surgical treatment compared to the younger ones. 9Distribution of cancer was also distinct that leading cancers in adolescents and young adults (aged 15-39 years) were thyroid cancer, testicular germ cell tumors, lymphoma, and leukemia, while in the oldest old (aged 85 years or more) were cancer of lung and bronchus, colon and rectum, breast, and prostate. 10For the overall population, previous studies elucidated that cancer incidence and mortality declined from 2013 to 2016 in the US.Specifically, incidence of liver and intrahepatic bile duct cancer increased the fastest in both genders during 2012-2016, while mortality of oral cavity and pharynx cancer in male and corpus and uterus cancer in female increased most rapidly during 2013-2017. 11nd steepest death decline was found in cancer of lung and melanoma (decreased more than 4% per year) during 2015-2019 in total US population. 12o explore cancer epidemiology in the elderly with the updated statistics, we analyzed cancer incidence, mortality, and survival during 1975-2020 in the US.Overall and site-specific cancer data by age, sex, and race/ethnicity were included.

| Data acquisition
The elderly in our study were defined as those aged 65 years or more and divided by 5-year age group (65-69, 70-74, 75-79, 80-84, and 85+ years).Cancer cases diagnosed between January 1, 1975, and December 31, 2020, were extracted from the National Cancer Institute's SEER 8 and SEER 17 (respectively covers 8.3% and 26.5% of the US population) database.Although SEER 8 covers the least population, it has the longest time span from 1975 to 2020, which provides us with data during 1975-1999.And SEER 17 was used for 2000-2020 data. 13Mortality data was extracted from SEER*Stat software 8.4.2 and provided by NCHS, which was based on information from all death certificates filed in the 50 states and the District of Columbia.
SEER*Stat software provides data on incidence, mortality and survival by age at diagnosis/death, year of diagnosis/death, sex, race, stage, site, etc. Stage of blood cancers were excluded for different staging system from solid cancer and SEER data blankage. 14Incidence cases diagnosed from autopsy or death certificate only were excluded.Both incidence and mortality rates were age adjusted to the 2000 US standard population.Relative survival data excluded the last diagnosis year (2020 in our study), which may miss some of the deaths and survival may be overestimated.Relative survival is a net cancer survival using cohort method in the absence of other causes of death, which is defined as the ratio of the proportion of observed survivors in a cohort of cancer patients to the proportion of expected survivors in a comparable set of cancer-free individuals. 15ncidence cases were screened using site recode ICD-O-3 (International Classification of Disease for Oncology, Third Edition)/WHO (World Health Organization) 2008, which was based on ICD-O-3 and updated for hematopoietic codes based on the WHO Classification of Tumors of Hematopoietic and Lymphoid Tissues (2008, Table S1).Mortality data was classified by ICD 8-10 (The International Statistical Classification of Diseases and Related Health Problems 8th/10th Revision).A total of 46 types of cancer were included in this study (Table S2).All cancers' frequencies, age-adjusted incidence/mortality rates, as well as relative survival of the elderly, were extracted, but only 12 cancers with the highest incidence or mortality rates, respectively, in males and females were displayed in the text (data for other cancers were available in Appendix S1).

| Statistical analysis
Temporal trends in incidence and mortality were analyzed by the Joinpoint software (version 5.0.2), which were presented as annual percentage change (APC) and average annual percentage change (AAPC).p-Value is not available for the Joinpoint default method, Empirical Quantile.JPSurve method was used to analyze 1-, 3-, and 5-year relative survival rates by calendar year, and trends in survival were represented as average absolute change in survival (AAC_S). 16

| Elderly cancer demographics
Table S3-S6 presented the demographic characteristics of cancer incidence and mortality for the overarching picture of 1975-2020 and the recent picture of 2000-2020.It was shown that male (54.0%) and white (85.4%) patients accounted for large parts of cases diagnosed during 1975-2020.A similar distribution could be seen in the cohort diagnosed from 2000 to 2020.Males had a larger proportion of incidence and death in each age subgroups than females except for those ≥85 years old (1975-2020 incidence: 45.0% male vs. 55.0%female; 1975-2020 death: 45.2% male vs. 54.8%female).
The proportion of cancer patients incidence peaked at 65-69 year age group (25.7% in 65-69 years; 24.4% in 70-74 years; 21.1% in 75-79 years; 15.6% in 80-84 years; 13.3% in 85+ years; Figure 1).The proportion of cancer patients' deaths increased with increasing age and met a plunge in females aged 75 years and males aged 80 years.Age-adjusted incidence rates peaked at 80-84 years and then decreased as patients aged, while age-adjusted mortality rates always kept a positive correlation with age.

| Elderly cancer distribution
Prostate cancer (29%) and breast cancer (26%) were the most common cancers in males and females, respectively, in while cancer of lung and bronchus was the most common cancer-related cause of death in both men (29%) and women (23%, Figure S1).Table 1 and Table S7 showed 12 cancers' distribution by age and sex with the highest incidence or mortality rate.As people get older, although the proportion of overall incidence decreased, incidence of female cancers of colon and rectum, pancreas, urinary bladder, stomach, and leukemia increased.mortality for females aged 85+ years.Cancers of digestive system (20.7%),respiratory system (16.1%), and male genital system (15.8%,Table S8) accounted for the majority of cancer incidence.Respiratory cancers (31.5%) ranked first in cancer mortality (Table S9).
The 18 cancers with the highest incidence or mortality rates showed similar trends in 1-, 3-, and 5-year relative survival, all of which improved during 1975-2020 (Figure 4).Note that survival for cancer of colon and rectum peaked in

| DISCUSSION
In this cross-sectional study, we found that cancers affecting the elderly were considerably discrepancy from the younger Though overarching trends were favorable, some cancers turned adversely.These findings are important for early cancer detection strategy-making and selection of government investment priorities in the elderly. 18ost previous studies focused on cancer trends in general or more specific populations. 18,19This study was conducted on the background that persons aged 65 years or more were the majority of cancer patients. 3The results in the early follow-up periods of our study were in line with previous studies, which were conducted using data from more limited periods, [20][21][22] and updated data was additionally analyzed in this present study.
Compared with the studies from Global Burden of Disease Study 2019 database, which is the largest and most comprehensive effort to quantify health loss across places and over time, 23 lung cancer had the highest disabilityadjusted life years globally, and it was the most common cancer death in the US. 24For colorectal cancer, incidence and mortality for all-age population increased since 1990 globally.But in this study, incidence and death of colorectal cancer had decreased since 1990 in the elderly people in the US. 25 F I G U R E 3 Annual Percentage Change (APC) of the cancers exhibiting the most rapid shifts in incidence or mortality rates during the recent follow-up period.
Cancer is a complex disease related to many factors, including genetics, lifestyle choices, environmental exposures, and infections.Social determinants of health on cancer risk include socioeconomic status, education, neighborhood and physical environment, employment, and social support networks, as well as access to healthcare. 26The trends in cancer epidemiology often reflect variations in these risk factors.For instance, the incidence of liver cancer rose at first and then began to decline, influenced by advancements in the treatment of hepatitis C virus (HCV).After the World Health Organization (WHO) adopted a strategy in 2014 to eliminate viral hepatitis as a global health threat by 2030, significant progress was made, contributing to the reduction in liver cancer rates. 27he cancer incidence rate in the elderly peaked at the age of 80-84 years, then declined after 85 years old (usually defined as oldest-old).As the poor health conditions and limited life expectancy, physicians usually avoid complicated detection strategies, which may lead to the underestimation of cancer diagnosis in the oldest old. 28,29Elderly males always had a higher proportion of cancer incidence than females, except for those aged 85 years or more, which largely due to the increased cancer of colon and rectum, urinary bladder, stomach, and leukemia in females.The mortality rate in the elderly increased as people aged.Males also had a higher proportion of mortality than females except for those aged 80 years or more, which were attributed to the increased deaths in many cancer types after 85 years old in female, including cancer of breast, colon and rectum, pancreas, etc. Populations aged 85 years or more showed unique cancer distribution, which resulted in them being classified as a special research group. 9,30he cancer incidence rate of elderly males was approximately 1.5 times that of females, similar for mortality, which was greatly attributed to gender differences in genetic predispositions and hazardous lifestyle choices (e.g., tobacco smoking, alcohol consumption). 31Sex disparities became prominent as people got older.Diagnoses of 13 cancers in males and 18 cancers in females, and mortality of 14 cancers in men and 15 cancers in women increased significantly during 1975-2020.Males showed a steeper decrease in both overall cancer incidence and mortality.The gender of patients with non-productive cancers had not been adequately considered during cancer curing until 2016 when the US National Institutes of Health introduced a policy changes towards genders for medical research. 32It is time to optimize cancer treatment strategies in men and women.
Although the overall elderly cancer incidence rate increased with no significance during 1975-2020, many cancer types showed significant increasing trends, including cancer of small intestinal, kidney and renal pelvis, endocrine system, melanoma, non-Hodgkin lymphoma, etc. Females had 18 cancer types with increasing incidence compared to 13 in males.Cancer of peritoneum, omentum, and mesentery in elderly females increased fastest but could play little role in the overall incidence due to its low absolute incidence.Most peritoneum, omentum, and mesentery cancer was metastatic, and most primary sites were ovary and fallopian tubes, which may lead to the increasing incidence in females. 35Our study also found high metastasis rate in ovary cancer, with 68% distant stage at diagnosis. 36herefore, physicians should attach importance to peritoneum, omentum, and mesentery screening after discovering primary malignancies.
Most cancer incidence and mortality rates decreased in the latest follow-up time.8][39] Previous studies also found a greater increase in iCCA versus extrahepatic cholangiocarcinoma during 2001-2017. 40However, misclassification of perihilar cholangiocarcinoma as iCCA should not be ignored in the ICD system, which may cause overestimation of the increasing incidence for iCCA. 40,41ancer of anus, anal canal, and anorectum had similar increasing incidence and mortality rates as iCCA, but survival was relatively well. 42That may attribute to the increasing infection of human papillomavirus due to adverse sexuality (e.g., anal intercourse) and HIV-related immunosuppression. 43Other cancers with increasing incidence in the latest follow-up period include cancer of other digestive organs in both genders and cancer of corpus uteri, other urinary organs, and pancreas in females.
Overall cancer mortality in the elderly decreased in the past two decades.Cancer in males decreased faster than females in the nearest research time.But a total of 18 cancers in elderly males (e.g., liver, anus, bile duct, retroperitoneum, pleura, peritoneum, bone and joints, etc.) and 16 cancers in elderly females (e.g., small intestine, anus, intrahepatic bile duct, retroperitoneum, pleura, bone and joints, other genital organs, etc.) showed upward mortality trend in the latest follow-up periods.][46][47] A steep drop was observed in cancer incidence in 2020, and many cancers showed similar trends, such as cancer of prostate, lung and bronchus, breast, melanoma, etc. Coronavirus disease 2019 (COVID-19) broke out in China at the end of 2019 and immediately spread worldwide.Previous studies indicated that the restriction measure of COVID-19 may lead to the limitation of cancer diagnosis or diagnostic tests. 48Studies also found that cancer patients had a higher risk of COVID-19 infection and developing severe events. 49,50However, we did not observe an increase in mortality or a decrease in survival, likely due to the short interval between COVID-19 diagnosis and the end of the cohort observation period.
Black individuals had a 3% higher proportion of distant stage at diagnosis than White in the elderly during 2010-2020, especially in cancer of oral cavity and pharynx (24% vs. 14%), corpus uteri (17% vs. 9%), and melanoma (14% vs. 5%).Previous research found that African American/ Black individuals had the highest mortality and lowest survival for most cancers compared with any racial/ethnic group, which was driven by the lower socioeconomic status of Blacks. 51,52The elderly cancer patients diagnosed with unstaged stage were similar between White and Black individuals, which partly revealed the efforts made by the US government to reduce racial inequalities. 53ith advances in cancer diagnosis and treatment, the detection of cancers at an early stage has improved, subsequently enhancing survival rates.For instance, endoscopy has become a crucial tool in identifying early-stage cancers of the esophagus and stomach.It is advised that individuals over the age of 40 should undergo biennial endoscopic examinations. 54Additionally, in Japan, insurance policies now cover screening and treatment for H. pylori, a known risk factor for stomach cancer, demonstrating a proactive approach in public health management and cancer prevention. 55here are several limitations of this study.At first, we used the database covering larger population to obtain as accurate statistics as possible, and the mortality data covered a wider US population than incidence data; therefore, there were more deaths than diagnoses during the follow-up periods.Secondly, SEER is a large populationbased database, and misclassification of race/ethnicity is unavoidable.Thirdly, death included data from autopsy or death certificate, which can lead to the misclassification of cancer death, particularly among the elderly population with multiple comorbidities.Forthly, we set the maximum number of joinpoints to 2 instead of the recommending number of 7, because that may cause the difficulty of trend description based on so many types of cancer, and deviation was unavoidable due to this adjustment.And SEER database lacks detailed information such as molecular pathological characteristics, socioeconomic status, and treatment (needs request), which limited our findings.

| CONCLUSIONS
The elderly populations are the majority of cancer patients, while long-term epidemiology research was limited.In this cross-sectional study, the rate of cancer incidence increased with no significance while mortality rate declined by 0.42% annually from 1975 to 2020.
Prostate and breast cancer were the most common malignancies in elderly males and females, respectively.And cancer of lung and bronchus showed the highest mortality rate.Overall incidence, mortality and survival improved, while incidence of intrahepatic cholangiocarcinoma and anus, anal canal, and anorectum cancer increased fast in the last period calculated by the Joinpoint method.It is time to optimize targeted cancer-curing medical strategies towards site-specific cancers and demographic-specific populations in the elderly.
Prostate cancer (n = 3,411,06) surpassed lung and bronchus (n = 269,768) cancer to be the leading cancerous cause of death in males aged 85+ years.Cancer of colon and rectum (n = 284,128) surpassed cancer of lung and bronchus (n = 255,017) and breast (n = 236,849) in F I G U R E 1 Overall cancer case distribution and average annual incidence (A) and death (B) rate per 100,000 persons by age and sex in the elderly population, United States, 1975-2020.

F I G U R E 2
Trends in incidence and mortality rates per 100,000 persons (1975-2020, United States, Age-Adjusted to the 2000 US standard population): (A) Overall cancer incidence trend by sex; (B) Overall cancer mortality trend by sex; (C) Incidence trend for cancers of the top 12 incidence in male; (D) Mortality trend for cancers of the top 12 mortality in male; (E) Incidence trend for cancers of the top 12 incidence in female; (F) Mortality trend for cancers of the top 12 mortality in female.

T A B L E 3
Annual Percentage Change (APC) and Average Annual Percentage Change (AAPC) for 12 cancers of the leading mortality rates by sex, United States, 1975-2020.-2020.17Overall cancer incidence rate increased during 1975-1991 and decreased during 2008-2020.While overall mortality rate had decreased since 1975.And relative survival of most cancers increased steadily during 1975-2019.

F I G U R E 4
Trends in 1-, 3-and 5-year relative survival for all-sites and 18 cancers with top 12 incidence or mortality in male or female, United States, 1975-2019.F I G U R E 5 Stage distribution for all-sites and cancers with top 12 incidence or mortality in male or female, United States, 2010-2020, excluding blood cancers.| 11of 14 XU et al.
Cancer incidence distribution by sex and age, United States, 1975-2020.a T A B L E 1 a Incidence data for 1975-1999 are from the Surveillance, Epidemiology and End Results (SEER) program: Incidence-SEER Research Data, 8 Registries, Nov 2022 Sub (1975-2020)-Linked To County Attributes.Incidence data for 2000-2020 are from the SEER program: Incidence-SEER Research Data, 17 Registries, Nov 2022 Sub (2000-2020)-Linked To County Attributes.b Ranked by decreasing incidence for males and females.The 12 sites listed respectively for male and female are those with the highest incidence (1975-2020).c Liver excludes intrahepatic bile duct.