Temporal trends and racial/ethnic disparities in hepatocellular carcinoma incidence in the US between 2000-2022.

Introduction
Primary liver cancer ranks as the sixth most common cancer and the third leading cause of cancer-related death worldwide.1,2 Although liver cancer is not among the top 10 leading incident cancers in the US, it ranked as the fifth leading cause of cancer-related deaths in men and the seventh in women in 2023.3 Hepatocellular carcinoma (HCC) is the dominant histologic type of liver cancer and accounts for over 80% of all liver cancer cases.4 With an overall 5-year survival rate of about 20% in the US, HCC is also among the deadliest cancers.
Infections with HBV and HCV remain the most important risk factors for HCC. Other risk factors include older age, male sex, obesity, cirrhosis, metabolic dysfunction-associated steatotic liver disease (MASLD), and alcohol-associated liver disease (ALD), with risk varying across racial/ethnic groups.5,6 In the US, there are an estimated 2.2 million adults with viremic HCV7 and 1.59 million adults living with chronic hepatitis B (CHB).8 The prevalence of HCV infection is 1.1–1.52% among non-Hispanic (NH) Black individuals, 0.84% among Hispanic individuals, 0.65–1.1% among NH White individuals, and 0.23% among NH Asian/Pacific Islander (API) individuals.7,9 An estimated 60% of the people living with CHB are NH API and 80% are non-US born.8 The prevalence of MASLD is about 30% in the US, with higher rates in men and Hispanic individuals.10
Previous studies using the National Cancer Institute (NCI)’s Surveillance, Epidemiology, and End Results (SEER)-21 database only extended through 2016,11,12 while SEER-12 analyses (covering only 13% of the US population) ended in 2019,13,14 and a recent analysis using the Centers for Disease Control and Prevention (CDC)’s database extended national trends to 2020.15 However, the SEER-12 and CDC studies reported substantial discrepancies in overall and race/ethnicity-specific trends, likely due to the limited coverage of SEER-12 and potential bias in the latter from COVID-related underdiagnosis. In this study, we address these gaps by excluding 2020 from trend analyses and using the latest SEER-21 data through 2022, which covers approximately 45.8% of the US population (including 39.6% White, 64.9% Hispanic, 43.5% Black, 68.2% Asian, 69.9% Native Hawaiian/Pacific Islander, and 48.3% American Indian/Alaska Natives [AIAN]).16 The goal is to provide an up-to-date and comprehensive assessment of HCC incidence and temporal trends across diverse racial/ethnic groups and age groups, in the era of direct-acting antivirals (DAAs) for HCV treatment and nucleos(t)ide analogues for CHB treatment, to offer new insights to inform public health priorities, guide policy, and support targeted interventions to reduce disparities.

Materials and methods

Data source

Incident HCC cases
Data on incident HCC cases from January 1, 2000 to December 31, 2022 were obtained from the NCI’s population-based SEER∗Stat database, Incidence-SEER Research Plus Limited-Field Data, 21 Registries (SEER-21).17 Due to the extremely low incidence of HCC in individuals aged 0-29 (about 0.1 per 100,000), the analysis was restricted to those aged ≥30 years. HCC was defined by International Classification of Diseases for Oncology, 3rd Edition, using topography code C22.0 and histology codes 8170 to 8175.18 Patients with HCC were classified by sex (male, female), race/ethnicity (NH White, Hispanic, NH Black, NH API, and NH AIAN), and age group (30-39, 40-49, 50-59, 60-69, and ≥70 years).

Population estimates
The county-level bridged-race population estimates, derived from the 2000, 2010, and 2020 US Census and incorporating intercensal estimates (for July 1, 2000-2009, July 1, 2010-2019) and Vintage 2023 estimates (for July 1, 2020-2022), were used as denominators to calculate HCC incidence rates and were obtained from SEER∗Stat.19,20 Age-adjusted HCC incidence rates (AAIRs) were calculated using the 2000 US standard population estimates through SEER∗Stat, grouped in 5-year age intervals.

Statistical analysis
Incident HCC case counts and percentage distributions were stratified by patient demographics. Age-adjusted and age-specific incidence rates (ASIR) of HCC, along with their 95% CIs, were estimated using SEER∗Stat. Case counts and rates based on fewer than 16 events were not shown in tables and were excluded from rate calculations, in accordance with US Cancer Statistics suppression guidelines to ensure stable statistical estimates and protect patient confidentiality.21
The percentage changes in incident HCC cases and AAIRs between 2000 and 2022 were used to represent relative changes from the first to the last year of the study period, defined as × 100%. The annual percentage changes (APC) in AAIRs and ASIRs with 95% CIs were calculated to quantify temporal trends from 2000 to 2022. Joinpoint regression analysis was conducted to estimate APCs. A log-linear model was fitted to the natural logarithm of the incidence rates, specified as ln(y) = α + βx + ε, where y denotes the incidence rate and x represents the calendar year. APCs were calculated as 100 × (eβ−1). Following SEER recommendations, data from 2020 were excluded from trend analyses, due to an approximate 10% decline in overall cancer incidence compared to 2019, primarily attributed to COVID-19-related delays in diagnosis and screening.22
All statistical analyses were performed using SEER∗Stat (version 8.4; NCI, Bethesda, MD)23 and the Joinpoint Regression Program (version 5.3; NCI, Calverton, MD). All p values were two-sided, and statistical significance was defined as p <0.05.

Results

Demographic characteristics of incident HCC cases
This analysis included 223,571 incident HCC cases diagnosed among individuals aged ≥30 years in the US between 2000 and 2022. Of these, 109,173 (48.8%) were NH White, 54,523 (24.4%) Hispanic, 29,745 (13.3%) NH Black, 27,643 (12.4%) NH API, and 1,982 (0.9%) NH AIAN individuals (Table S1). Overall, 76.2% of HCC cases occurred in males and 23.8% in females. The highest number of HCC cases was in the 60-69 age group (77,291 cases, 34.6%), followed by those aged ≥70 years (33.3%) and 50-59 (25.4%). Only 6.7% of the HCC cases occurred in individuals aged <50 years. The highest number of incident HCC cases occurred in individuals aged ≥70 years among NH White and NH API, whereas among Hispanic, NH Black, and NH AIAN, the peak was observed in the 60-69 age group (Table S1).
The annual number of incident HCC cases increased steadily from 4,758 in 2000 to 12,201 in 2014, followed by a slower increase to a peak of 13,109 cases in 2019, and then declined to 11,853 cases by 2022, representing a 150% increase compared to 2000 (Table 1, Fig. S1).

Temporal trends in age-adjusted incidence rates of HCC
The AAIRs of HCC increased from 6.8 per 100,000 (11.2 in males vs. 3.2 in females) in 2000 to 12.4 per 100,000 (20.1 in males vs. 5.5 in females) in 2014, followed by a plateau through 2019 before a subsequent decline to 10.5 per 100,000 (16.7 in males vs. 5.1 in females) by 2022 (Figs. 1 & S2, Table 2). Despite the recent declining trend, overall HCC incidence has increased by 54.4% since 2000.
Joinpoint regression showed an initial increasing trend in AAIR of HCC by 5.55% per year (95% CI 4.90%-6.93%) from 2000 to 2007, followed by a slower increase of 3.16% per year (95% CI 2.28%-3.95%) until 2014; AAIR plateaued from 2014 to 2019 (APC = -0.48, p >0.05), before a declining trend by 3.91% per year (95% CI 2.52%-4.83%) from 2019 to 2022 (Fig. 1, Table 2). Similar patterns were observed in males (Table 2, Fig. S3), whose trends closely aligned with the overall population. In contrast, females who have a lower AAIR of HCC exhibited a smaller decline than males from 2014 to 2022 (APC = -0.71, 95% CI -1.46 to -0.05) (Table 2, Fig. S4).

Temporal trends in age-specific incidence rates of HCC
The HCC ASIRs in 2000 and 2022 are shown in Table S2, and the trends in HCC ASIR are shown in Fig. 2 and Table S3. There was a steady increasing trend in HCC incidence in individuals aged ≥70 years from 17.0 per 100,000 in 2000 to 32.6 per 100,000 in 2019 (APC2000-2019 = 3.13, p <0.05), which then plateaued after 2019 (p >0.05). HCC incidence among individuals aged 60-69 years began to decline from 33.1 per 100,000 in 2019 to 27.1 per 100,000 in 2022 (APC2019-2022 = -5.99, p <0.05). The younger age groups showed an earlier decline in HCC incidence. Individuals aged 50-59 years had the greatest decline in HCC incidence from 17.5 per 100,000 in 2014 to 8.6 per 100,000 in 2022 (APC2014-2019 = -7.09, APC2019-2022 = -9.99; all p <0.05). Individuals aged 40-49 years showed a declining trend from 3.6 per 100,000 in 2009 to 1.7 per 100,000 in 2022 (APC2009-2022 = -5.32, p <0.05). HCC incidence remained low between 2000 and 2022 in individuals aged 30-39 years. The inflection points for the declining trend in HCC incidence among males aged ≥40 years were generally consistent with the overall trend. Among females aged ≥70 years, the upward trend beginning in 2013 was not statistically significant (APC2013-2022 = 1.45, p >0.05), while the trend plateaued for males aged ≥70 years in 2019 (Table S3).

Racial/ethnic disparities in HCC AAIR trends
In 2000, NH API had the highest AAIR of HCC (18.9 per 100,000), with the highest incidence in both men (29.6 per 100,000) and women (10.0 per 100,000) (Table 1, Fig. 3). After an initial increase, NH API was the first racial/ethnic group to exhibit a declining trend, with an annual decrease of 1.80% beginning in 2007 (95% CI 0.73%-3.08%), nearly 10 years earlier than other racial/ethnic groups (Table 2, Fig. 4). This decline accelerated to 4.76% per year after 2015 (95% CI 3.96%–6.42%). In contrast, AAIR began to decline in NH Black and NH AIAN populations in 2016 by 5.68% (95% CI 4.45%–7.47%) and 4.98% (95% CI 0.97%–11.99%) per year, respectively. Among Hispanic (which plateaued from 2014 to 2019) and NH White populations, AAIR began declining in 2019 by 4.40% (95% CI 1.89%–6.22%) and 3.57% (95% CI 1.52%–5.14%) per year, respectively, marking a consistent decline across all racial/ethnic groups in the most recent period (Table 2, Fig. 4).
By 2022, the AAIR of HCC was highest in the Hispanic (17.7 per 100,000), followed by NH AIAN (16.8 per 100,000), NH API (13.1 per 100,000), NH Black (10.9 per 100,000), and NH White (8.0 per 100,000) populations (Table 1, Fig. 3). The AAIR of HCC in Hispanic individuals surpassed that in NH API starting in 2012. Despite recent declines in AAIRs across all racial/ethnic groups, HCC AAIR increased between 2000 and 2020 by 63.3% in NH White individuals, 38.3% in Hispanic individuals, 26.7% in NH Black individuals, and 90.9% in NH AIAN individuals (Table 1). NH API was the only racial/ethnic group that had a lower HCC incidence (-30.7%) in 2022 than in 2000 (Table 1). The only racial/ethnic groups that did not show a significant declining trend were NH White females (APC2015-2022 = -0.31, p >0.05) and NH AIAN females (APC2014-2022 = -2.13, p >0.05) (Table 2, Fig. S3 and S4).

Racial/ethnic disparities in HCC ASIR trends
Between 2000 and 2022, HCC incidence in NH API declined across all age groups (Table S3, Fig. S5). Among individuals aged ≥70 years, a significant decline in HCC incidence was observed only in NH API, from 59.2 per 100,000 in 2013 to 40.6 in 2022 (APC2013-2022 = -3.60, p <0.05). Among individuals aged 60-69 years, HCC incidence significantly declined among all racial groups. NH APIs experienced a continuous decline in HCC incidence from 41.1 per 100,000 in 2000 to 28.1 in 2022 (APC2000-2022 = -1.50; p <0.05) (Table S2 and S3, Fig. S5). Declines among Hispanic, NH AIAN, NH Black, and NH White populations emerged later, with significant downward trends beginning sequentially between 2016 and 2019 (all p <0.05). Among individuals aged 50-59 years, there has also been a significant decline in HCC incidence among the various racial groups. NH API and NH Black populations showed an earlier decline in HCC incidence (Table S3, Fig. S5). The NH Black population showed an accelerated decline in HCC incidence starting in 2016, which was steeper than the prior decline from 2009 (APC2016-2022 = -16.96 vs. APC2009-2016 = -5.39; all p <0.05), and NH APIs showed a significant downward trend beginning in 2011 (APC2011-2022 = -7.12, p <0.05). A significant declining trend in HCC incidence was first observed among NH White, Hispanic, and NH AIAN populations, beginning around 2015-2017 (NH White: APC2015-2022 = -8.87; Hispanic: APC2017-2022 = -8.27; NH AIAN: APC2015-2022 = -10.74; all p <0.05). Among individuals aged 30-39 years, only NH API showed a significant declining trend in HCC incidence that began in 2003. Only Hispanic individuals showed a significant increasing trend among individuals aged 30-39 years between 2018 and 2022 (Table S3). Temporal trends in HCC ASIRs among males closely mirrored those in the overall population (Fig. S6), whereas patterns among females differed slightly (Fig. S7).

Discussion
Based on the NCI’s SEER-21 database that covers almost half of the US population, we found a 54.4% increase in HCC incidence in the past 23 years (2000-2022) among individuals aged ≥30 years. The trend in HCC incidence increased steadily from 2000 to 2007, slowed by 2014, plateaued, and began to decline after 2019. Our overall trends are largely consistent with the report by Yazan Abboud,15 which analyzed 2001-2020 data from the CDC’s United States Cancer Statistics and the National Center for Health Statistics. Our study represents both a continuation and an evolution of prior findings based on the SEER-21 database, which suggested a plateau between 2013 and 2016.12 Our study shows that this plateau persisted through 2019 and was subsequently followed by a decline. However, it differs from studies based on the SEER-12 database that reported a notable decline in HCC incidence during 2015-2019.13,14
CHB is the predominant cause of HCC in NH API, whereas HCV and MASLD are more commonly associated with HCC in other racial/ethnic groups.[24], [25], [26] A study of Medicare recipients (aged ≥65 years) revealed that between 2005 and 2014, HCV-related HCC continued to increase and surpass MASLD-related HCC in 2012, becoming the leading cause of HCC.25 The study found that in 2014, the major causes of HCC were HBV and HCV among API, HCV among Black individuals, and MASLD and HCV among Hispanic and White individuals.25 The earlier decline in HCC incidence observed in this study in NH Black individuals beginning in 2016, compared to the later declines in Hispanic and NH White populations after 2019, may be attributed to differences in predominant etiologies, with the NH Black population primarily affected by HCV, whereas both MASLD and HCV are the most common causes of HCC in Hispanic and NH White populations.25 The smaller increase, plateau, and subsequent decline in HCC incidence in these racial/ethnic groups in recent years appeared to coincide with key milestones in HCV management, including the release of the first guidance on HCV treatment by the American Association for the Study of Liver Diseases (AASLD) and the Infectious Diseases Society of America in 2014; the FDA approval of the highly effective curative DAA treatment for HCV since 2014; and the recommendation in 2012/2013 by the CDC and the US Preventive Services Task Force (USPSTF) to expand HCV risk-based screening to include baby boomers (born 1945-1965) and universal adult (aged 18-79) screening in 2020.27
The CDC estimated that between 2014 and 2020, about 1.2 million people in the US received DAA treatment for HCV infection.28,29 While hepatocarcinogenesis is a multiyear process at the individual level, real-world population-based evidence indicates that DAA therapy can substantially reduce HCC incidence. A large retrospective cohort study using data from the Optum Clinformatics Data Mart database (2010-2021) in the US reported that among patients with cirrhosis, HCC incidence was significantly lower in those treated with DAAs than in untreated patients (20.1 vs. 41.8 per 1,000 patient-years), whereas no significant difference was observed among patients without baseline cirrhosis (3.8 vs. 4.0 per 1,000 patient-years).30 Similarly, Beste et al. reported that the overall HCC incidence among Veterans Health Administration patients declined from a peak of 48.0 per 100,000 in 2015 to 40.3 per 100,000 in 2018.26 This decrease was primarily driven by a significant reduction in HCV-related HCC, which peaked at 31.0 per 100,000 in 2015 and declined by 29.6% to 21.8 per 100,000 in 2018, following the implementation of a large-scale system-wide HCV treatment initiative between 2014 and 2016.26 Comparable population-level declines in HCV-related HCC were also reported in California and Florida between 2014/2015 and 2018.24,31
In contrast, Beste et al. reported that the incidence of non-HCV-related HCC increased after 2015.26 Similarly, Pinheiro et al. reported rising trends in MASLD-related and ALD-related HCC among both males and females in Florida between 2010 and 2018.24 The current high incidence of HCC in Hispanic individuals warrants further studies and targeted interventions, including efforts to increase hepatitis C treatment access and address non-viral etiologies such as ALD, obesity, diabetes, and MASLD. Importantly, substantial heterogeneity exists in Hispanic populations. Puerto Ricans and Cubans accounted for the highest numbers of HCC cases among Hispanic subgroups between 2010 and 2018 in Florida. Puerto Ricans experienced the highest HCV-related HCC burden (10.9 per 100,000 males), whereas Cubans had lower rates (2.8 per 100,000 males) than NH White individuals (4.4 per 100,000 males).24 Recent evidence identified diabetes as the second and obesity as the fourth strongest predictor of HCC risk in a large MASLD cohort.32 The increasing prevalence of obesity and diabetes has contributed to a parallel increase in MASLD,33 fueling the upward trend in MASLD-related HCC. In addition, demographic and nativity shifts, along with competing mortality from non-HCC cancer and cardiovascular diseases,34 may also have contributed to the recent trends.
NH API had a disproportionately high burden of HCC, accounting for 12.4% of incident HCC cases between 2000 and 2022, despite representing only 3.6-6.2% of the adult US population during this period.35,36 In 2000, NH API had the highest HCC incidence among all racial/ethnic groups. Since then, NH API was the only racial/ethnic group to experience a significant decline in HCC incidence across all age groups, occurring as early as 2007 and resulting in a 30.7% drop in incidence in the past 23 years. The declining trend in HCC incidence among NH API since 2007 appears to follow several key milestones in CHB management. These include the release of the AASLD hepatitis B practice guidelines in 2001 (first version), 2004, and 2008; the FDA approvals of oral nucleos(t)ide analogues for CHB treatment, including lamivudine in 1998, adefovir in 2002, entecavir in 2005, tenofovir disoproxil fumarate in 2008, and low-cost generic entecavir and tenofovir in 2014–2017;37 and the recommendations by the CDC in 2008 and the USPSTF in 2014 to screen non-US-born adults from regions with high hepatitis B prevalence, including Asia and Africa. Pinheiro et al. reported that between 2010 and 2018, HBV-related HCC had the highest incidence (7.5 per 100,000) in NH API, followed by MASLD-related HCC (3.9 per 100,000), HCV-related HCC (3.6 per 100,000), and ALD-related HCC (0.8 per 100,000),31 with HCV-related HCC being nearly as common as HBV-related HCC among Vietnamese and Cambodian individuals, which helps explain the accelerated decline around 2015 coinciding with the major milestones in both CHB and HCV management. Notably, substantial heterogeneity also exists in NH API, with HCC incidence ranging from 7.3 per 100,000 in Japanese males to 48.6 per 100,000 in Cambodian males between 2010 and 2018.31 HBV-related HCC remained the predominant etiology in Chinese, Korean, Laotian, Thai, and Hmong males. In contrast, MASLD was the leading cause of HCC (6.2 per 100,000) in Filipino males, followed by HBV-related HCC (3.6 per 100,000). MASLD- and HCV-related HCC accounted for the majority of cases (2.8 and 2.4 per 100,000, respectively) in Japanese males.31
Antiviral treatment particularly among Asian patients with CHB has been found to significantly reduce HCC incidence in both cirrhotic and non-cirrhotic patients.38,39 A study of the Medicare Part D insurance data found that the number of beneficiaries receiving either brand-name or generic tenofovir or entecavir for CHB treatment increased by 54.4%, from 28,446 individuals in 2013 to 43,924 in 2020 (APC2013-2020 = 6.4, p <0.05).37 A sustained decline in HBV-related HCC, especially in NH API, is likely to continue in the future with increasing efforts to screen and treat CHB coupled with a reduction in new CHB cases with improving global infant hepatitis B birth dose and three-dose coverage. In 2023, the CDC recommended universal screening of adults aged ≥18 years. Additionally, a new state law in California, enacted in 2022, mandates that primary care clinics offer adults both hepatitis B and C screening tests and provide evaluation and treatment for those who test positive. Meanwhile, the high MASLD-related HCC incidence among NH API, especially in Filipino and South Asian populations, warrants further epidemiologic investigation and targeted prevention efforts.31
When analyzed by race/ethnicity, there was a significant declining trend in HCC incidence in adults aged 40-49 years, 50-59 years and 60-69 years in all racial/ethnic groups. During 2015–2019, an initial or accelerated decline in HCC incidence was observed among individuals aged 50–59 and 60–69 years, with the greatest decline in the 50–59-year age group, across NH White, Hispanic, NH Black, and NH AIAN populations. No decline was observed among individuals aged ≥70 years in these racial/ethnic groups. The persistently high incidence in those aged ≥70 years may be partly attributable to the increasing contribution of MASLD as the common cause of HCC in this age group.25 Only NH API had a significant declining trend in HCC incidence among all age groups, including those aged 30-49 years and ≥70 years.
This study has several limitations. First, although SEER provides high-quality, standardized incidence data, it lacks information on nativity and individual-level etiologic factors, limiting our ability to comprehensively assess the underlying contributors to racial/ethnic disparities. Second, our trend analysis is ecological in nature and may be affected by unmeasured temporal changes in diagnostic practices, screening uptake, or reporting accuracy. Third, conclusions regarding the effects of specific public health interventions are inferential and cannot be directly tested within this framework. Therefore, our findings should be interpreted with caution, particularly when attributing observed changes in incidence to specific etiological or policy-related factors.
In summary, this study reveals significant variation in HCC incidence trends across race/ethnicity. The earlier and more sustained decline among NH API most likely reflects increased treatment for CHB. Later declines among NH Black and NH AIAN populations, a plateau among Hispanic populations, and an accelerated decline among NH API populations appear to have been influenced by the availability of curative HCV treatment. The declining incidence of HCC suggests sustained efforts in viral hepatitis screening and treatment over the past two decades have achieved meaningful success in the prevention and control of HCC at the population level. Future research should aim to quantify longitudinal etiologic differences across racial/ethnic groups and geographic regions, thereby informing more targeted and proactive prevention strategies.

Abbreviations
AAIR, age-adjusted incidence rate; AASLD, American Association for the Study of Liver Diseases; AIAN, American Indian/Alaska Native; ALD, alcohol-associated liver disease, APC, annual percent change; API, Asian/Pacific Islander; ASIR, age-specific incidence rate; CDC, Centers for Disease Control and Prevention; CHB, chronic hepatitis B; DAA, direct-acting antiviral; NCI, National Cancer Institute; NH, non-Hispanic; SEER, Surveillance, Epidemiology, and End Results; USPSTF, US Preventive Services Task Force.

Authors’ contributions
WJ acquired and analyzed the data, performed data visualization, and drafted the manuscript. SS conceptualized the study, reviewed the results, and contributed to interpretation. HP reviewed the results and contributed to interpretation.

Data availability
Data used in this study are publicly available at www.seer.cancer.gov. SEER∗Stat Database: Incidence - SEER Research Plus Limited-Field Data, 21 Registries, Nov 2024 Sub (2000-2022).

Declaration of generative AI and AI-assisted technologies in the writing process
During the preparation of this work, WJ used ChatGPT in order to improve language and readability of our manuscript. We have reviewed and edited the content as needed and take full responsibility for the content of the publication.

Financial support
Supported by a gift to the Stanford Asian Liver Center by Dr. CJ and Ha Lin Yip Huang and the John C. Martin Foundation.

Conflict of interest
The authors declare no conflicts of interest that pertain to this work.
Please refer to the accompanying ICMJE disclosure forms for further details.