Contemporary epidemiology of hepatocellular carcinoma: understanding risk factors and surveillance strategies

Abstract

The contemporary epidemiology of hepatocellular carcinoma (HCC) shows a shift in the main etiological risk factors from less common but highly virulent (eg, hepatitis C and B) to more common but weak risk factors (eg, alcohol and metabolic syndrome). Therefore, we are in a seemingly paradoxical state of declining overall incidence rates of HCC-related to improved prevention and treatment of viral hepatitis but burgeoning number of people at an elevated risk of HCC. Several geographic regions have reported an increase in HCC attributable to alcoholic liver disease and metabolic dysfunction associated with steatotic liver disease (MASLD). The importance of risk stratification is increasing to allow for targeted prevention and early detection of HCC. Most risk factors predispose HCC through the formation of cirrhosis, which has served as the main risk stratifying factor. However, this scheme is showing cracks at both ends of the spectrum. On one hand, the risk of developing HCC varies widely among patients with contemporary advanced fibrosis or cirrhosis, and on the other hand up to one-third of MASLD-related HCC occurs among patients with no clear evidence of cirrhosis. The use of multidimensional (eg, clinical, epidemiological, and biochemical) predictive algorithms may improve risk stratification efforts. The shift in HCC risk factors also further heightened the importance and limitations of current surveillance practices (eg, reduced performance of ultrasound in MASLD). Therefore, exploring advanced imaging methods, new biomarkers but also existing combinations of biomarkers augmented by clinical factors for HCC early detection is crucial.

Epidemiology

Hepatocellular carcinoma (HCC) is a major contributor to the global cancer burden and is the third most common cause of cancer-related deaths.¹ Worldwide, HCC is the most prevalent form of liver cancer, comprising 75%–85% of all cases.^2,3 Most patients with HCC are diagnosed at an advanced stage with a typically dismal prognosis. Of the few patients diagnosed at an early stage, receipt of curative treatment (ie, liver transplantation, resection, or ablation) and long-term survival can be possible.^3–5 The global incidence of HCC is very heterogeneous due to the varying prevalence of underlying etiological risk factors. Approximately 72% of HCC cases are reported in Asia, with China alone accounting for more than 50%. By contrast, Europe represents 10%, Africa 7.8%, North America 5.1%, and Latin America 4.6% of total HCC cases.⁶ Recently, the overall age-adjusted incidence rates of HCC both globally as well as in some select countries, including China, Japan, and the United States,⁷ have observed a stabilization or decline in rates among specific population subgroups.^8–10 However, while the incidence rates of HCC have started to decline, the frequency of cases continues to increase due to the growing and aging world population.

The risk of developing HCC escalates with advancing age^2,11,12; however, there are pronounced discrepancies between the median age of diagnosis globally. In the United States, the median age of HCC diagnosis ranges from 60 to 69 years, and in China, it is approximately 58 years. Across other African countries, the median age varies from 58 years in Egypt to 46 years in some sub-Saharan African countries.^6,13 A pronounced sex-based difference in HCC incidence and mortality is observed,^12,14 with males experiencing rates 2–4 times higher than females. This disparity is largely due to the higher prevalence of HCC risk factors in men and the potential role of androgens in increasing HCC risk.^2,15,16 Regarding racial and ethnic disparities in HCC within North America, historically, Asians and Pacific Islanders had the highest incidence rates, primarily due to chronic hepatitis B virus (HBV) infection.⁶ However, the implementation of HBV vaccination programs and advancements in antiviral treatments have reduced the population-attributable fraction of HCC from HBV. Concurrently, an increase in metabolic syndrome and alcohol-related liver disease has led to the highest HCC rates now being observed among Hispanics, American Indians and Alaska Natives in the United States . Ongoing monitoring and research into HCC trends and risk factors are crucial for understanding shifting patterns and developing targeted prevention and intervention strategies to mitigate the global impact of HCC.

Risk factors

Almost 90% of HCC cases arise within the context of chronic liver disease, with cirrhosis being the primary risk factor for its development, irrespective of the underlying cause of liver disease.¹⁷ Thus, any agent that induces chronic liver injury leading to cirrhosis is considered a risk factor for HCC. In patients with cirrhosis, annual incidence rates for HCC range from 1% to 8% and are largely dependent on etiologic risk factors for cirrhosis. For example, the incidence rate is 2% in those with HBV and 3–8% in those actively infected with hepatitis C virus (HCV).¹⁸ However, most patients with cirrhosis in contemporary clinical practice have cured HCV, treated HBV with adequate viral suppression, alcohol-associated liver disease (ALD), and MASLD; the risk of HCC among these patients is significantly lower (1-2%) than those reported in outdated series.^19,20Figure 1 shows the changes in HCC etiologies worldwide, between 1990 and 2019.

Figure 1.

The proportion of HCC etiologies globally, between 1990 and 2019. Data from Yang et al²¹ and Toh et al.²². Abbreviations: AHBV, hepatitis B virus, ALD: alcoholic liver disease, HCV: hepatitis C virus, MASLD: metabolic dysfunction-associated steatotic liver disease.

HBV infection

Globally, HBV is a leading cause of HCC, accounting for approximately one-third of new cases and deaths. However, the proportion of HCC attributable to HBV differs by region from 60% of cases in Africa and East Asia to 20% or less in Western nations.⁵ By integrating its DNA into host cells, HBV instigates chronic inflammation, fibrosis, and mutations in liver cells, contributing to HCC. Evidence of HBV DNA in non-cancerous cells of HCC patients suggests that genetic integration can precede and potentially initiate cancer development. These data suggest that HCC can occur independently of cirrhosis, but the presence of cirrhosis significantly increases risk.^23–27

For HBV carriers, the lifetime HCC risk is estimated to be between 10% and 25%.²⁴ Risk factors for an increased HCC risk among HBV carriers include demographic characteristics (male sex, older age, Asian or African ancestry, family history of HCC), viral characteristics (HBV DNA levels, genotype C^26–28), and environmental factors (aflatoxin).²⁹ Lifestyle factors such as heavy smoking and alcohol intake significantly multiply the HCC risk in those with HBV by as much as ninefold.³⁰ Metabolic (dysfunction) factors such as diabetes mellitus (DM) are also associated with an additional HCC risk in individuals with HBV. A recent systematic review and meta-analysis of 36 studies indicated that DM was linked to a 36% higher risk for developing HCC in HBV patients,³¹ this risk is higher among patients with longer duration of DM and inadequate glucose control.³² The association between obesity and the risk of HCC in HBV is inconsistent,^33–36 with a recent population-based cohort study in Korean adults with chronic HBV reporting that obesity was linked to a 22% higher risk of HCC in men and a 46% higher risk in women compared to those with a BMI between 18.5 and 22.9.³⁷

There are several validated scoring systems for predicting HCC risk in HBV carriers, their accuracy depends on whether they were developed and tested in patients undergoing antiviral treatment or not.^38,39 Antiviral treatment for HBV infection with nucleos(t)ide analogs (NAs), achieves sustained reductions in HBV DNA levels and improves liver function in most patients, and their use is associated with significant reductions in the short and medium-term risk of HCC.^40,41 The components of these scores include demographic and clinical factors of which the most crucial contemporary factors are the presence of cirrhosis and NA treatment as shown in Table 1.⁴² A meta-analysis of 21 studies (only 2 RCT) published through 2009 revealed that individuals undergoing NA therapy, primarily with the early NAs like lamivudine, had a notably lower incidence of HCC compared with untreated counterparts (2.8% in treated vs. 6.4% in untreated patients) during a 46 (32-108) month period (P = .003).⁴¹ Recent studies are showing similarly significantly lower incidence among those treated with entecavir and tenofovir.^43–45 Some studies reported a greater reduction in HCC risk among those treated with tenofovir than entecavir, but the evidence is not supportive of meaningful clinical differences.^46,47

Table 1.

Summary of hepatitis B virus-hepatocellular carcinoma prediction models.

Risk Score	Author (Year)	Race	Number	Cirrhosis (%)	Prior Treatment	Medication	Risk parameters	Risk score cut-offs
REACH-B	Yang (2011)	Asian (Taiwan)	3584	0	All naïve	Not on Tx	Age, male, ALT, HBeAg, HBV DNA	Low: <10, intermediate: 10–12, High: >12
PAGE-B	Papatheodoridis (2016)	Caucasian	1350	20	33%	ETV/TDF	Age, male, platelet	Low: <10, intermediate: 10–17, high: >17
mPAGE-B	Kim (2018)	Asian (Korea)	2001	19.1	39.5%	ETV or TDF	Age, male, platelet, albumin	Low: ≤ 8, intermediate: 9–12, high: ≥ 13
CU-HCC	-	Asian (China)	1005	38.1	Only 15.1% on Tx	-	Age, albumin, bilirubin, HBV DNA, cirrhosis	Low: <5, intermediate: 5–19, high: ≥ 19
HCC-RESCUE	Sohn (2017)	Asian (Korea)	990	39	All naïve	ETV	Age, male, cirrhosis	Low: ≤ 64, intermediate: 65–84, high: ≥ 85
CAMD	Hsu (2019)	Asian (Taiwan)	23 851	26.4	–	ETV or TDF	Age, male, diabetes, cirrhosis	Low: <8, intermediate: 8–13, high: >13
APA-B	Chen (2017)	Asian (Taiwan)	883	35.9	All naïve	ETV	Age, platelet, AFP (12 months after ETV)	Low: ≤ 5, intermediate: 6–9, High: ≥ 10
REAL-B	Yang (2019)	Asian-Pacific	8048	20.2	20%	Any anti-viral	Age, male, alcohol use, cirrhosis, diabetes, platelet, AFP	Low: <4, intermediate: 4–7, high: 8-13
AASL-HCC	Yu (2019)	Asian (Korea)	1242	39.3	All naïve	ETV or TDF	Age, male, albumin, cirrhosis	Low: ≤ 5, intermediate: 6–19, high: ≥ 20
RWS-HCC	Poh (2016)	Asian (Singapore)	538	14.9	–	Any anti-viral	Age, male, cirrhosis, AFP	Low: <4.5, high: ≥ 4.5

Abbreviations: AFP: alpha-fetoprotein; ETV: Entecavir; TDF: Tenofovir disoproxil.

In addition to those with cirrhosis, select groups of patients with chronic HBV may benefit from regular HCC surveillance. Current guidelines recommend HCC surveillance for those with chronic HBV and no cirrhosis, especially in areas where HBV is typically acquired at birth or in early childhood as these individuals are at increased risk of HCC. The age or circumstances that call for beginning this surveillance can vary. For example, the Canadian Association for the Study of the Liver supports (category 2, grade B) regular screenings for HCC in individuals with non-cirrhotic HBV infection, including Asian men starting at age 40, Asian women starting at age 50, people of African descent beginning at age 20 (due to the prevalence of HCC among non-cirrhotic patients), and individuals with family history of HCC.⁴⁸ These recommendations are driven by the perceived increased magnitude of HCC risk.

Universal newborn HBV vaccination programs are both crucial and effective for the primary prevention of HCC; although the vaccination schedules still do vary globally. For example, since implementing the national newborn HBV vaccination program in 1984, Taiwan has experienced, as of 2016, a 35.9% (26.8% to 44.4%) drop in HCC incidence in ages <30 years and 51.5% (41.7%–60.1% reduction in HBV-related HCC mortality among age cohort 0–29 years of age.⁴⁹ Other nations like China, Singapore, and Spain have also seen significant decreases in HBV as a contributing factor to HCC following the implementation of newborn HBV vaccination programs and also aflatoxin reduction campaigns.⁵⁰ Evidence suggests that there is a synergistic interaction between aflatoxin and HBV and heightened HCC risk. National food policy reforms in China which have eradicated aflatoxin from the food supply have led to population attributable benefit of 65% for reduced liver cancer mortality, primarily in those with HBV infection.⁵¹

Wider use and increasing indications for NA therapy (eg, Taiwan’s national HBV therapy program launched in November 2003) may have also contributed to the decline in HBV-related HCC. Despite these successes, many HBV-endemic countries have not introduced universal vaccination, leaving vast numbers of people, particularly in Asia and sub-Saharan Africa, at risk for HBV infection and its sequelae.⁵²

HCV infection

Active HCV infection is associated with an approximately 20-fold increase in HCC risk. HCV, an RNA virus, does not integrate into the host’s genome; thus, it is unlikely to be the primary initiator of tumorigenesis per se. In the context of HCV, progression toward HCC stems from prolonged inflammatory damage leading to progressive fibrosis, with cirrhosis being a precursor in approximately 90% of HCC cases.⁵³ The annual incidence of HCC in individuals with HCV-related cirrhosis ranges from 0.5% to 10%, with lower estimates among those cured of the infection.⁵⁴ HCV is the primary cause of HCC in Western European countries, North America, parts of North Africa, and the Middle East. Co-factors that further increase HCC risk among HCV-infected individuals include demographic (male sex, Hispanic ethnicity, and older age), viral (HCV genotype 3, longer infection duration, coinfections with HBV or HIV), clinical (cirrhosis, insulin resistance, obesity, and diabetes), and lifestyle (tobacco use, and alcohol use).^54,55 The importance of most of these factors, except for cirrhosis, has been reduced in people achieving treatment-related sustained virologic response (SVR).

Overall, the incidence rates of HCV-related HCC have started declining in several countries including the United States.^56–58 A meta-analysis of observational studies published through 2012 reported that SVR was associated with reduced risk for HCC (relative risk for all persons, 0.24 [95% CI, 0.18–0.31], and for those with advanced liver disease, 0.23 [CI, 0.16–0.35].⁵⁹ However, the risk of developing HCC in patients with cured HCV remains elevated enough to merit surveillance for at least 4–5 years post SVR in those with advanced fibrosis or cirrhosis as compared with those with no advanced fibrosis.^60–62 For example, a retrospective cohort study by Kanwal et al reported that the annual incidence of HCC in cirrhotic patients with SVR remained higher than in non-cirrhotic patients, with rates of 1.82 versus 0.34 per 100 person-years.⁶³ In patients cured from HCV with directly acting antiviral therapy who have cirrhosis, the cumulative 1-, 2-, and 3-year risks of HCC are 1.1%, 1.9%, and 2.8%, respectively.⁶³ Therefore, continued HCC surveillance is recommended among patients with cirrhosis after HCV cure.⁶⁴ However, recommendations for HCC surveillance in HCV-infected patients with F3 fibrosis vary among guidelines^65–67 and the practice may not be cost-effective.⁶⁸ Other and weaker factors associated with increased risk of HCC following SVR include older age, Hispanic ethnicity, DM, and alcohol use.^69–72 Models incorporating these post-treatment factors have shown high predictive ability for determining de novo HCC risk (C statistic ranging from 0.81 to 0.87). In a cohort of 527 patients with advanced chronic liver disease who achieved sustained virological response (SVR) after interferon-free therapy, post-treatment AFP levels, along with age, liver stiffness measurement, and albumin, emerged as stronger predictors of hepatocellular carcinoma (HCC) development compared to pre-treatment values or treatment-related changes.⁷³ These factors were integrated into models that successfully categorized patients into low- and high-risk groups for HCC development. Approximately 2-thirds of patients were classified as low-risk (<1% annual risk), suggesting that they might not require intensive HCC surveillance, thus potentially optimizing resource allocation. An alternative model that excluded AFP also demonstrated robust performance, reinforcing the utility of post-treatment clinical parameters in effectively stratifying patients by their risk of developing HCC post-SVR.⁷³

Alcohol drinking

Excessive alcohol consumption is a well-established risk factor for HCC^74–77 through the development as well as the progression of cirrhosis. Alcohol-related cirrhosis significantly contributes to the development of HCC, with 20%–25% of HCC cases attributed to alcohol-associated liver disease (ALD).⁷⁸ The annual risk of HCC among patients with alcohol-related cirrhosis ranges from 1.3% to 3%.⁷⁹ In a meta-analysis of 19 cohorts, heavy alcohol consumption (≥3 drinks/day) was associated with a 16% increased risk of HCC. No association was found with lower levels of consumption (<3 drinks/day),⁸⁰ but its impact has not been extensively studied.

The impact of alcohol on HCC risk is amplified by viral hepatitis and metabolic syndrome.⁸¹ Furthermore, alcohol may have a stronger association with HCC risk among women than men, possibly due to differences in alcohol dehydrogenase activity⁸² or a stronger link between alcohol and cirrhosis among women.⁸³ A meta-analysis of 572 studies published through 2012 examining heavy drinking (>4 drinks/day) found alcohol was associated with an almost four-fold increased HCC risk among women compared with a 59% increased risk among men.⁷⁴

Metabolic syndrome

Growing evidence indicates that metabolic syndrome, characterized by insulin resistance, abdominal obesity, atherogenic dyslipidemia, and hypertension, elevates the relative risk of HCC with and without MASLD. In a 2019 meta-analysis of 11 cohort studies, patients with metabolic syndrome were found to have a 76% increased risk of HCC (relative risk [RR] = 1.76; 95% CI: 1.33–2.33, P = .000, I² = 87.6%).⁸⁴ In a meta-analysis of 17 case-control studies and 32 cohort studies, individuals with diabetes were found to have a significantly higher risk of developing HCC (RR = 2.31, 95% CI, 1.87–2.84)⁸⁵ Additionally, a meta-analysis of 7 cohort studies revealed a statistically significant elevated risk of HCC mortality among individuals with diabetes compared to those without (RR = 2.43, 95% CI: 1.66–3.55).⁸⁵ The risk of HCC increases with BMI in those who are overweight or obese, with HRs of 1.36 (95% CI, 1.02–1.81), 1.77 (95% CI, 1.56–2.01), and 3.08 (95% CI, 1.21–7.86) for BMI >25 kg/m², >30 kg/m², and >35 kg/m², respectively. Moreover, an increase in BMI was associated with increased HCC-related mortality (HR, 1.61; 95% CI, 1.14–2.27).⁸⁶

Metabolic dysfunction-associated steatotic liver disease (MASLD)

A large part of the risk conveyed by metabolic syndrome is mediated by the development of MASLD. MASLD is the most common chronic liver disorder, affecting 20%–30% of people worldwide.⁸⁷ The spectrum of MASLD ranges from steatosis to a more progressive form, metabolic dysfunction associated steatohepatitis (MASH), all of which carry a variable risk for HCC which is largely driven by progression in fibrosis and cirrhosis. Up to one-third of HCCs from MASLD develop in the absence of cirrhosis.⁸⁸ MASLD-related HCC has been reported around the world,⁸⁹ with higher estimates reported in studies from the United Kingdom,⁹⁰ India,^91,92 Germany,⁹³ the United States,⁹⁴ and the Middle East,⁹⁵ and lower estimates(1%–2%) from China and Japan.⁹⁶ Compared with HCC related to HCV or ALD, there is a lower male-to-female ratio in patients with MASLD-related HCC, who are also 5–14 years older.⁹⁴

The overall absolute risk of HCC in MASLD is low. A large cohort study reported an HCC incidence of 0.21 per 1000 person-years among persons with MASLD of any severity, which was 7-fold higher than controls without MASLD.⁹⁷ The main risk factor for MASLD HCC is cirrhosis where the annual incidence rates of HCC range from 0.3% to 4.7%.^88,89,98 HCC incidence in studies of non-cirrhotic MASH is low (0.01%–0.13% per year),⁸⁹ and even lower in non-cirrhotic MASLD population.⁹⁹ Other risk factors for HCC in MASLD include older age, Hispanic ethnicity in the United States, presence of features of metabolic dysfunction especially DM, and possibly high BMI, dyslipidemia, and hypertension, as well as genetic variability.^97,100,101 In European and US cohort studies of individuals with MASLD, diabetes was the strongest independent metabolic risk factor for the development of HCC.^97,102 There is weak evidence that alcohol drinking in low to moderate amounts significantly increases MASLD HCC risk in the presence of advanced fibrosis and cirrhosis.^103,104 PNPLA3 rs738409 single-nucleotide polymorphism is associated with a 67% increased risk of HCC in individuals with MASH or alcoholic cirrhosis in a meta-analysis of 24 observational studies.¹⁰⁵ HCC risk calculators including VA score¹⁰⁶ and NAFLD fibrosis score (NFS)¹⁰⁷ used some of these risk factors and showed modest performance for HCC risk stratification in MASLD. A model incorporating genetic variants in PNPLA3-TM6SF2-GCKR-MBOAT7 and HSD17B13 in a risk score that predicted HCC independently of classical risk factors and cirrhosis in a mostly Caucasian population with an AUROC of only 0.65^[106].

The Global Burden of Disease study highlights burden, temporal, and geographical variations in the causes of HCC.^21,108 Between 1990 and 2019, there was a noticeable shift from chronic viral hepatitis to non-viral causes. In 1990, over half of HCC cases were due to HBV, but this percentage dropped to 42% by 2019. Meanwhile, the incidence of HCC-related to alcohol and MASH rose from 13% to 18% and from 5% to 6%, respectively. In 2019, chronic viral hepatitis remained the predominant cause of HCC in Asia and most African regions, whereas non-viral causes, such as MASH and ALD were the primary contributors to HCC cases in North America, Europe, and Australia.²¹ Due to the high prevalence of metabolic syndrome and despite the low relative risk of HCC, the population-attributable fraction of obesity for HCC cases is estimated to be 16% in Europe.¹⁰⁹ The population-attributable factor of obesity and/or diabetes for HCC cases in the United States is estimated to be 37%.⁸¹ Treating dyslipidemia with statins is associated in observational studies with HCC risk reduction by 37%–42%.^110,111 Diabetes, particularly in males, is linked to a 2–3-fold increased risk of HCC¹¹² with a longer duration of diabetes incrementally raising risk.¹¹³ Treatment with metformin has been associated with a 51% decrease in HCC risk among patients with type 2 diabetes compared with treatment with insulin or sulphonylureas.^114,115 This protective effect is independent of the duration of diabetes, the presence of complications or adequate glycemic control,¹¹⁵ which therefore should remain a top priority in treating patients with diabetes.

Tobacco smoking

A review of 113 studies by the 2014 US Surgeon General’s report revealed that current cigarette smoking was linked to a 70% increased risk of HCC, and former smoking was associated with a 40% increased risk.¹¹⁶ Several epidemiological studies have shown that smokers face an increased risk of developing HCC,⁷⁷ with a meta-analysis of 38 cohort studies and 58 case-control studies published through 2009 reporting an adjusted relative risk of 1.5 (95% CI: 1.37, 1.67) compared with non-smokers.¹¹⁷ Individuals who quit smoking more than 30 years ago had a similar HCC risk to never-smokers.¹¹⁸ There is also evidence that smoking increases HCC risk among patients with cirrhosis irrespective of underlying etiology.¹⁹

Other dietary factors

Coffee consumption has consistently been linked to a decreased risk of HCC.¹¹⁹ A 2017 meta-analysis of 18 cohort and 8 case-control studies found that an additional 2 cups of coffee per day were associated with a 35% reduced risk.¹²⁰ The association is observed mostly with caffeinated coffee. The precise mechanisms behind coffee’s potential protective effects remain unclear, but experimental evidence suggests that caffeine and various other components in coffee, such as diterpenes, could have anti-inflammatory, anti-fibrotic, insulin-sensitizing, and anti-carcinogenic properties.¹²¹ Coffee is also associated with a reduced risk of insulin resistance and diabetes^122,123 and generally lower AST and ALT.¹²⁴ The American Association for the Study of Liver Diseases (AASLD) guidelines support coffee consumption as prevention, however, the recommendation level is weak (level 5).¹²⁵Table 2 provides a summary of current international guidelines for HCC surveillance. People in Sub-Saharan Africa and Southeast Asia often encounter high dietary levels of aflatoxin B1, a mycotoxin from Aspergillus fungi that is linked to an increased risk of HCC, especially when combined with HBV infection.¹²⁷ Studies have also reported an increase in HCC risk with high fat and protein intake^128–130 and a non-significant trend to reduce the risk of HCC with vegetables.^131–133 However, there were no observed associations with fruit consumption.^131,134,135

Table 2.

Comparison of current international guidelines on HCC surveillance.

Guideline	Target population	Screening interval	Screening method	Biomarkers
CASL 126	– Any etiology, Child–Pugh A or B cirrhosis – Any Child–Pugh cirrhosis awaiting LT – Chronic HBV with active hepatitis or family history of HCC, Africans, African Americans, Asian males > 40 yr and Asian females > 50 yr	Every 6 months	US	None
AASLD 67	– Any etiology, Child–Pugh A or B cirrhosis – Any Child–Pugh cirrhosis awaiting LT – Chronic HBV with active hepatitis or family history of HCC, Africans, African Americans, Asian males > 40 yr and Asian females > 50 yr	Every 6 (4–8) months	US	At discretion
EASL 66	– Any etiology, Child-Pugh A or B Cirrhosis – Child-Pugh C awaiting LT – Chronic HBV with ≥10 PAGE-B score – Chronic HCV with bridging fibrosis	Every 6 months	US	None
APASL 65	– Any etiology cirrhosis – Asian males > 40 yr – Asian females > 50 yr – Africans > 20 yr – Family history of HCC	Every 6 months	US	AFP

Abbreviation: AASLD, American Association for the Study of Liver Diseases; AFP, alpha fetoprotein; APASL, Asian Pacific Association for the Study of the Liver; CASL, Canadian Association for the Study of the Liver; EASL, European Association for the Study of the Liver; HCC, hepatocellular carcinoma; LT, liver transplantation; PAGE-B score, patient, age and gender (range 0–18 points); US, ultrasound.

Other HCC risk factors

The annual risk of HCC is generally lower in patients with cirrhosis due to genetic, autoimmune, and cholestatic liver diseases compared to viral etiologies. For example, in a meta-analysis of 8 studies, the annual risk of HCC for patients with hereditary hemochromatosis and cirrhosis was found to be 1.1%.¹³⁶ The overall risk of HCC in hereditary hemochromatosis is significantly higher in patients who were homozygous for the C282Y mutation or had the C282Y/H63D genotype.¹³⁷ In one large cohort study from Sweden, individuals with hereditary hemochromatosis had a 20-fold higher risk of HCC compared to the general population.¹³⁸ Cirrhosis due to alpha-1 antitrypsin deficiency is associated with an annual HCC incidence of approximately 0.9%.^139,140 For cirrhosis due to autoimmune hepatitis, the annual risk ranges from 0.4% to 1.1%.¹⁴¹ Patients with primary biliary cholangitis (PBC) have an annual HCC incidence rate of about 1.3%. In patients with cirrhosis from PBC, the 3-year HCC risk is estimated to be 5.9%.¹⁴² The annual risk of HCC in primary sclerosing cholangitis (PSC) has not been well studied, but the lifetime risk is estimated to be between 0.8% and 2.5%. However, the annual risk of cholangiocarcinoma ranging from 0.5% to 1.5%.¹⁴³ Therefore, routine surveillance imaging is recommended for patients with PSC. Patients with cirrhosis due to congestive hepatopathy, often seen in those with congenital heart disease post-Fontan procedure and chronic heart failure, are also at an elevated risk for HCC¹⁴⁴ but the annual incidence is estimated to be below 0.5%.¹⁴⁴ Although the risk of HCC in some of these conditions, is below the 1.0%–1.5% annual risk threshold for which surveillance might be cost-effective, surveillance is recommended given the limited effort involved with these rare conditions.

Surveillance strategies for early HCC detection

Importance of early detection and evidence for current HCC surveillance practice

The key distinction between screening and surveillance lies in their objectives and target populations. Screening involves testing asymptomatic individuals within the general population to detect early signs of a particular disease, aiming to identify cases early for timely intervention. Conversely, surveillance focuses on monitoring individuals at high risk of developing a specific disease, aiming to detect disease progression or recurrence.¹⁴⁵

Optimal screening tests are designed to detect asymptomatic or subclinical diseases and must meet several criteria, including high sensitivity, cost-effectiveness, and availability. Diseases suitable for screening include those that are of high burden in selected populations with a proven treatment strategy and outcomes that improve with early detection.¹⁴⁶ Screening tests at regular intervals in at-risk populations are called surveillance.¹⁴⁷ The AASLD recommends offering surveillance when the risk of HCC is at least 1.0%–1.5% per year. This recommendation encompasses patients with cirrhosis and high-risk non-cirrhotic hepatitis B carriers.¹²⁵ With recent improvements in diagnostic and treatment modalities, the threshold of annual HCC risk at which surveillance is cost-effective could be as low as 0.8%.¹⁴⁸

The 1-year and 3-year survival rates in patients with advanced HCC are low at 36% and 17%, respectively.¹⁴⁹ Early detection presents the opportunity for curative treatment options, such as surgical resection, transplantation, and percutaneous ablation.^149–151

The only randomized controlled trial supporting HCC surveillance was conducted in over 18 000 Chinese patients.¹⁵² Patients were randomly assigned to a surveillance group (n = 9373) or control group (n = 9443). Surveillance in this study entailed the measurement of serum alpha-fetoprotein test (AFP) levels and ultrasound imaging every 6 months. There was a significant reduction (37%) in HCC-related mortality in the surveillance group. Despite being a landmark study, it had limitations, including the absence of outcome data other than death, missing information regarding all-cause mortality, and failure to account for clustering, potentially leading to misleading statistical significance. Also, this trial was conducted in an untreated HBV-infected patient population and lacks generalizability to contemporary HBV populations on treatment. It is also uncertain if these results are generalizable to other populations with chronic liver disease where there could be a higher competing risk of liver-related mortality. Most of the evidence for evidence of HCC surveillance efficacy and effectiveness comes from observational studies.^153–157 A meta-analysis of 47 studies, involving 15 158 patients, found that 6284 (41.4%) had HCC detected through surveillance. The analysis revealed that HCC surveillance was associated with improved early-stage detection (odds ratio [OR] 2.08, 95% CI, 1.80–2.37) and curative treatment rates (OR 2.24, 95% CI, 1.99–2.52). Additionally, HCC surveillance was linked to significantly prolonged survival (OR 1.90, 95% CI, 1.67–2.17), even after adjusting for lead-time bias in some studies.¹⁵³

Surveillance rates were higher in studies conducted in the United States (51%) and Europe (45%).¹⁵³ More recent studies with long-term prospective follow-up of patients with compensated viral cirrhosis have shown that those who adhere to the recommended 6-month screening interval have a higher proportion of HCC detected at an early stage, leading to a survival benefit because of access to curative procedures.¹⁵⁸ Limitations include lead-time bias, in which part of the survival benefit could be attributed to earlier diagnosis due to surveillance, and time-length bias in which tumors that were diagnosed early differ in prognosis from those diagnosed later.¹⁶ Nonetheless, the consistency of the data has supported guidelines for regular HCC surveillance.¹⁵³

Western recommendations typically advocate for a 6-month screening interval for HCC surveillance based on the estimated volume doubling time of HCC, which is approximately 6 months.¹⁵⁹ Japanese guidelines propose a 3-month interval for high-risk groups with cirrhosis from HBV or HCV.¹⁶⁰ Previous research indicated that HCC surveillance is advantageous for patients with Child–Pugh A or B cirrhosis. However, it may not provide substantial survival benefits for those with Child–Pugh C cirrhosis, unless they are candidates for liver transplantation. Additionally, patients with short-life expectancy from significant medical comorbidities, advanced age, or poor performance status, also lack a clear survival benefit from surveillance.¹³⁶

Negative viewpoints on screening highlight issues related to suboptimal screening tests coupled with low adherence or poor implementation of screening procedures.¹⁶¹ The effectiveness of surveillance programs could be compromised by inadequate frequency of screenings, substandard imaging techniques, or insufficient follow-up, rather than inherent flaws in the concept of surveillance itself.¹⁶ Therefore, efforts to optimize the implementation of surveillance programs could be key to realizing their potential benefits in improving patient outcomes.

HCC surveillance tests

Imaging tests

Ultrasonography is an economical and non-invasive method for surveillance, devoid of risks or radiation exposure for patients. However, its efficacy in detecting HCC in a cirrhotic liver is influenced by various factors, including hepatic characteristics, such as abnormal liver texture; patient traits, like obesity; and technical constraints, such as the quality of ultrasound and the expertise of the operator.¹⁶² Its overall sensitivity for detecting HCC at any stage is 84%. This value is also affected by the traits outlined above as well as the severity of liver disease,^163–165 so sensitivity can be inconsistent between different centers and patients. Also, one study indicated that 1 in 5 HCC surveillance exams in the United States are inadequate.¹⁶⁵ This inadequacy was particularly pronounced in obese patients and those with cirrhosis from alcohol or MASLD, likely because of compromised ultrasound visualization due to subcutaneous fatty tissue present alongside hepatic steatosis.¹⁶⁵ Consequently, small or early-stage HCC nodules can go under-detected.^165,166 Increasing the surveillance frequency to every 6 months instead of annually boosted sensitivity to 70% for early-stage HCC detection.¹⁶⁷

Given the limitations of ultrasound for HCC surveillance, CT and MRI are increasingly being used in clinical practice.¹⁶⁸ However, there is limited data supporting the routine use of thiscross-sectional imaging for HCC surveillance. A small, single-center randomized trial comparing CT and ultrasound-based surveillance among patients with cirrhosis failed to find a significant difference in early detection (62.5% vs 55.5%, respectively, P = .93) or HCC-related mortality (8.8% vs. 6.0%, respectively, P = .46).¹⁶⁹

Another prospective cohort study comparing MRI- and ultrasound-based surveillance among patients with predominantly HBV-related cirrhosis found that MRI-based surveillance had significantly higher sensitivity for early HCC detection than ultrasound (83.7% vs 25.6%, respectively).¹⁷⁰ Further data are needed regarding MRI performance in non-HBV patients and its cost-effectiveness before routine adoption in clinical practice. Concerns about potential physical harm, financial costs, and limited radiologic capacity also restrict the routine use of CT or MRI for HCC surveillance in all patients with cirrhosis.¹⁶ While the AASLD practice guidelines acknowledge the limited reliability of ultrasound in patients with truncal obesity or marked parenchymal heterogeneity, CT or MRI is not recommended as the primary imaging technique for HCC surveillance. However, CT or MRI could be employed in select patients with inadequate ultrasound visualization or at high risk for inadequate ultrasound.¹³⁶

Future directions in imaging surveillance include abbreviated MRI protocols that focus on T1-weighted imaging post-gadoxetate disodium injection supplemented with T2-weighted imaging and diffusion-weighted imaging. These protocols show promising sensitivity (80%–90%) and specificity (91%–98%) in small cohort studies.^170–173 Future research should evaluate the cost-effectiveness and safe implementation of MRI, particularly in situations where ultrasound reliability is compromised, such as in individuals with truncal obesity or significant parenchymal heterogeneity associated with cirrhosis.

Laboratory tests

AFP is an extensively studied blood-based biomarker that has completed all 5 phases of biomarker validation but has shown limited efficacy when used alone. Serum AFP values above 20 ng/mL demonstrate approximately 70% sensitivity and 90% specificity for any HCC detection.¹⁷⁴ Meta-analyses comparing ultrasound alone versus ultrasound plus AFP for early HCC detection have demonstrated that combining ultrasound with AFP improves sensitivity for detecting early-stage HCC to 63% (95% CI, 48%, 75%) vs 45% (95% CI, 30%, 62%), respectively, but with a decrease in specificity to 84% vs 92%.¹⁷⁵

Several approaches have been proposed to improve AFP performance for HCC surveillance. AFP is conventionally interpreted using a cutoff of 20 ng/mL for all patients with cirrhosis regardless of underlying etiology.¹⁷⁶ One strategy is to evaluate AFP values trend over time rather than using a single-test result with a fixed threshold.^177,178 or using different cutoffs based on the etiology of cirrhosis.¹⁷⁹ AFP-L3 measures a specific subfraction of AFP and is more specific but less sensitive than AFP.¹⁸⁰ Des-gamma-carboxy prothrombin (DCP), also known as a protein induced by vitamin K absence/antagonist-II, is a variant of prothrombin produced at high levels by some HCCs.^181–184 DCP and AFP-L3, have shown promise in phase II biomarker studies, but their individual efficacy is not clear in phase III cohort studies.¹⁸⁵ These biomarkers are currently approved by the US Food and Drug Administration for risk stratification, but not for HCC surveillance.

Multidimensional biomarker scores for HCC surveillance

Due to intra-tumoral heterogeneity observed in HCC, a single biomarker may not be sufficient for optimal sensitivity in early HCC detection. Consequently, the combination of multiple biomarkers is increasingly used to enhance sensitivity. One well-studied biomarker panel is GALAD, which incorporates gender, age, AFP-L3, AFP, and DCP.¹⁸⁶ In a multinational phase II study involving 6834 patients (2430 with HCC and 4404 with chronic liver disease), GALAD achieved sensitivities ranging from 60% to 80% for early HCC detection.¹⁸⁵ In a phase III biomarker study, GALAD demonstrated a notable enhancement in sensitivity for detecting HCC.¹⁸⁷ However, this improvement was accompanied by an increase in false-positive results. Overall, the performance of GALAD was modest and comparable to using AFP-L3 alone or the HES algorithm.¹⁸⁷ Similarly, the hepatocellular carcinoma early detection screening (HES) algorithm was developed and included various components such as current AFP levels, age, platelet count, and ALT levels, along with longitudinal changes in AFP.¹⁸⁸ This updated algorithm showed significant improvement compared to using AFP alone in predicting the 6-month risk of HCC in the derivation sample. Notably, the sensitivity of the algorithm within 6 months prior to HCC, corresponding to a 10% false positive rate, was 61.37%, which represented a 3.84% absolute improvement over using AFP alone.¹⁸⁹ In a study analyzing 4804 cases of HCC in a cohort of veterans with cirrhosis over a median follow-up time of 3.12 years, the HES algorithm demonstrated improved sensitivity compared to the AFP assay alone. The HES study estimated that for every 1000 imaging analyses, the HES algorithm detected 198.57 cases of HCC compared to 185.52 cases for the AFP assay alone, resulting in the detection of 13 additional cases of HCC (P < .0005).¹⁹⁰ HES V2.0 incorporates AFP-L3% and DCP in addition to changes in AFP level, AFP-L3%, and DCP level in the prior year, combined with age, ALT, platelets, and etiology (HCV, alcoholic liver disease). HES V2.0 was developed and tested in a large prospective cohort phase III biomarker study involving a well-defined cirrhosis population undergoing HCC surveillance.¹⁹¹ Data from 2331 patients were analyzed, among whom 125 developed HCC (71% early stage). HES V2.0 had a significantly higher performance for identifying new HCC, including early stage, than GALAD. For any HCC, HES V2.0 demonstrated a higher true positive rate (TPR) than GALAD overall (+7.2%), at 6 months (+3.6%), at 12 months (+7.2%), and 24 months (+13.0%) prior to HCC diagnosis. For early HCC, HES V2.0 showed higher sensitivity/TPR than GALAD overall (+6.7%), at 12 months (+6.3%), and 24 months (+14.6%), but not at 6 months (+0.0%).¹⁹² (Figure 2). These approaches hold promise for improving early detection of HCC and warrant further validation.

Figure 2.

Components of GALAD, HES, and HES V2.0 and comparison of true positive rates.

Novel HCC biomarkers

There are also ongoing efforts in the development of novel cancer biomarker assays. These assays include tests for cancer-specific DNA mutations, differentially methylated regions of DNA, microRNAs, long noncoding RNAs, native and post-translationally modified proteins, and biochemical metabolites. A methylated DNA panel exhibited a high c-statistic of 0.96 (95% CI, 0.93, 0.99) with a sensitivity exceeding 90% for early HCC detection in a phase II study involving 146 patients (95 with HCC and 51 with cirrhosis).¹⁹³ Recent findings indicate differential expression of various biomolecules in exosomes released from tumor cells compared with those from normal cells.¹⁹⁴ Another promising biomarker is cell-free DNA released from tumor cells, which can be detected in peripheral blood samples. However, its utility for surveillance is still in the early phases of evaluation, as it is uncertain if cell-free DNA is present in sufficient quantities in patients with early-stage tumors.¹⁹⁵

The National Cancer Institute of the US National Institutes of Health created the early detection research network. The early detection research network has provided a guide for the clinical advancement of surveillance efforts. They advocate for a PRoBE, prospective-specimen-collection, retrospective blind evaluation, and design to test new biomarkers as clinical tools before recommending their use.¹⁹⁶ To address these gaps, investigators are building large prospective cohorts, such as the early detection research network’s hepatocellular cancer early detection study and the Cancer Prevention Research Institute of Texas’s HCC Consortium cohort, with stored biobanks of longitudinal serum and plasma samples to facilitate phase III validation of HCC biomarkers.¹⁹⁷

Conclusion

The risk of developing HCC varies among patients with cirrhosis, making it challenging to assess individual risk levels accurately.¹⁹⁸ Despite participation in surveillance programs, some cirrhotic patients are diagnosed with HCC at an advanced stage due to the limitations of ultrasound sensitivity. To address this challenge, exploring advanced imaging like MRI and new biomarkers for HCC detection is crucial. However, costly surveillance may not be justified for cirrhotic patients with controlled HBV/HCV who have low HCC incidence.^41,199–201 HCC scoring systems based on routine clinical features could allow personalized assessment of HCC risk and refinement of screening policies. However, the effectiveness of these systems and their applicability to non-cirrhotic MASLD patients remain uncertain.^202–204 Stratifying patients into groups of low-, intermediate-, or high-risk for developing HCC could optimize cost-effectiveness and resource allocation in screening strategies. Further research and validation are required to enhance the clinical utility of these approaches.

Supplementary material

Supplementary material is available at Journal of the Canadian Association of Gastroenterology online.

Author contributions

F.J., G.C., H.E.S.: conception, design, literature review, writing of first draft, approval of last draft.

Funding

This study was conducted with partial support from the National Institutes of Health (NIH) U01 CA230997 (PI Kanwal), P01 CA263025 (PI El-Serag), P30 DK056338 (PI El-Serag) and Cancer Prevention and Research Institute of Texas RP150587 (PI El-Serag), RP190641 (PI El-Serag), and RP220119 (PI El-Serag). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or the Cancer Prevention and Research Institute of Texas. Hashem El-Serag had grant funding from Gilead, Merck, Wako and Glycotest. The funding agencies had no role in design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation of the manuscript.

Conflicts of interest

The authors declare no conflicts of interest.

In addition to this COI statement, ICMJE disclosure forms have been collected for all co-authors and can be accessed as supplementary material.

Data availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.