Liver cancer risk-predictive molecular biomarkers specific to clinico-epidemiological contexts

Abstract

Hepatocellular carcinoma (HCC) risk prediction is increasingly important because of the low annual HCC incidence in patients with the rapidly emerging non-alcoholic fatty liver disease or cured HCV infection. To date, numerous clinical HCC risk biomarkers and scores have been reported in literature. However, heterogeneity in clinico-epidemiological context, e.g., liver disease etiology, patient race/ethnicity, regional environmental exposure, and lifestyle-related factors, obscure their real clinical utility and applicability. Proper characterization of these factors will help refine HCC risk prediction according to certain clinical context/scenarios and contribute to improved early HCC detection. Molecular factors underlying the clinical heterogeneity encompass various features in host genetics, hepatic and systemic molecular dysregulations, and cross-organ interactions, which may serve as clinical-context-specific biomarkers and/or therapeutic targets. Toward the goal to enable individual-risk-based HCC screening by incorporating the HCC risk biomarkers/scores, their assessment in patient with well-defined clinical context/scenario is critical to gauge their real value and to maximize benefit of the tailored patient management for substantial improvement of the poor HCC prognosis.

1. Introduction

Liver cancer is the sixth most common cancer type worldwide with more than 900,000 new cases annually, and the third leading cause of cancer-related death with more than 800,000 deaths each year (Sung et al., 2021). Hepatocellular carcinoma (HCC) is the dominant histological type of primary liver cancer that mostly develops in chronically diseased liver with cirrhosis, the terminal stage of progressive liver fibrosis (Fujiwara, Liu, Athuluri-Divakar, Zhu, & Hoshida, 2019). Chronic infection of hepatitis C virus (HCV) and hepatitis B virus (HBV) has been the major etiologies for HCC along with alcohol abuse, but with the recent remarkable progress in anti-viral drug development and the obesity epidemic, metabolic liver disease (i.e., non-alcoholic fatty liver disease [NAFLD] and newly proposed metabolic dysregulation-associated fatty liver disease [MAFLD]) is sharply emerging as the new predominant HCC etiology globally (J. Liu et al., 2021; Loomba, Friedman, & Shulman, 2021).

Because of the distinctly higher HCC incidence in patients with advanced fibrosis or cirrhosis compared to other cancer types in general population, practice guidelines from professional societies recommend semi-annual HCC screening with abdominal ultrasound with or without alpha-fetoprotein (AFP) in the chronic liver disease patients for detection of early-stage HCC amenable to curative therapies and improved survival (Marrero et al., 2018). However, with the rapidly growing target patient population as well as various provider- and patient-related barriers to implementation of the recommended protocol, the regular HCC screening is astoundingly underutilized in real-world clinical practice (utilization rate < 25%) (Wolf, Rich, Marrero, Parikh, & Singal, 2021). Thus, HCC risk prediction has been increasingly important to mitigate the burden on the HCC screening by identifying a subset of patients who need it most. A Markov-model-based simulation analysis showed that individual-risk-based personalized HCC screening, altering its intensity according to predicted HCC risk, is substantially more cost-effective compared to the current standard-care ‘one-size-fits-all” strategy uniformly applying the screening to all patients with cirrhosis (Goossens et al., 2017).

Clinical studies have shown that HCC risk, measured by annual incidence rate, hugely varies according to multiple clinical factors, particularly liver disease etiology often bound with geographic regions (e.g., HBV in Southeast Asia) (Fujiwara, Friedman, Goossens, & Hoshida, 2018). While annual HCC incidence rate is 3-7% in cirrhosis patients with active HCV infection, the rate in patients with histologically confirmed NAFLD cirrhosis is less than 1% (Sanyal et al., 2021; Simon et al., 2021). Given the recent dynamic and drastic shift of dominant liver disease etiology from viral hepatitis to metabolic disorders due to the widespread use of direct-acting antivirals and the global obesity epidemic, consideration about clinical context becomes more and more critical for cost-effective HCC risk prediction. To date, various clinical and/or molecular factors have been proposed as HCC risk indicators in diverse liver disease patient populations across the world, although none of them has been adopted in clinical practice to date (Kubota, Fujiwara, & Hoshida, 2020). In this chapter, we overview currently available HCC risk indicators in literature according to specific clinical contexts and discuss potential strategies for their clinical deployment and implementation toward the goal to refine management of the patients at risk of HCC development and improve patient prognosis (Figure 1).

Figure 1. HCC risk stratification considering clinico-epidemiological contexts.

2. Clinico-epidemiological contexts relevant to liver cancer risk

Besides liver disease etiology, HCC risk can be affected by various factors as depicted by the diverse HCC incidence across geographic regions despite shared etiology. For example, in Japanese cirrhosis patients with active HCV infection, 5-year cumulative HCC incidence reaches 30%, whereas the incidence is 17% in the Western countries; in the HBV-related cirrhosis patients in the endemic regions, 5-year cumulative HCC incidence is 15%, whereas the incidence is 10% in the Western countries (Fattovich, Stroffolini, Zagni, & Donato, 2004; Mancebo et al., 2013). These observations suggest the presence of multiple factors that influence HCC risk and confound each other, including patient race/ethnicity, exposure to environmental carcinogens, lifestyle-related factors such as diet and drinking, access and pattern of clinical care, among many others. Thus, to enable precise HCC risk assessment, it is ideal to derive and validate HCC risk prediction algorithms, controlling these influential factors. However, this is a challenging task that requires coverage of diverse global patient populations. To date, numerous clinical HCC risk scores have been defined and tested in Asian, European, and American patients (Table 1). Cross-regional validations have been conducted for some of the scores, although the scope of assessing the clinico-epidemiological confounding factors is still limited due to the highly biased distribution of the HCC risk-associated variables. Some scores such as ADRESS-HCC score (Flemming, Yang, Vittinghoff, Kim, & Terrault, 2014), aMAP score (Fan et al., 2020), and Toronto HCC Risk Index (THRI) (Sharma et al., 2017) were developed in multi-etiology cohorts, although variation in etiological composition may limit their general applicability. In contrast, some investigators have developed risk scores tailored for specific liver disease etiology (Ioannou, Green, Kerr, & Berry, 2019). In this section, we summarized recent findings about clinical factors associated with HCC risk according to the major liver disease etiology as the first-order clinico-pathological context that defines HCC risk.

Table 1. Clinical HCC risk scores validated in two or more independent cohorts.

Risk score	Variables	Major etiology	Major race/ethnicity	Reference
UM regression model	23 clinical variables	HCV, cryptogenic, alcohol, other	Caucasian, Black, Hispanic	(Singal et al., 2013)
aMAP risk score	Age, sex, albumin-bilirubin, platelets	HBV, HCV, HCV after SVR, non-viral	Asian, Caucasian	(Fan et al., 2020)
			Non-Hispanic
ADRESS-HCC	Age, diabetes, race, etiology, sex, Child-Pugh score	HCV, alcohol, NASH, HBV, other	white, Hispanic/Latino, African American, Asian	(Flemming et al., 2014)
THRI	Age, sex, etiology, platelets	Viral, steatohepatitis, PBC, AIH	n.a.	(Sharma et al., 2017)
Hughes et al	AFP	HCV, HBV	n.a.	(Hughes et al., 2021)
CU-HCC	Age, albumin, bilirubin, HBV-DNA, cirrhosis	HBV	n.a.	(V. W. Wong et al., 2010)
LSM-HCC	Liver stiffness, age, albumin, HBV-DNA	HBV	n.a.	(G. L. Wong et al., 2014)
RWS-HCC	Sex, age, cirrhosis, AFP	HBV	Asian	(Poh et al., 2016)
REACH-B	Sex, age, ALT, HBeAg, HBV-DNA	HBV	n.a.	(H. I. Yang et al., 2011)
NGM1-HCC	Sex, age, family history of HCC, alcohol, ALT, HBeAg	HBV	n.a.	(Yang et al., 2010)
NGM2-HCC	Sex, age, family history of HCC, alcohol, ALT, HBV-DNA	HBV	n.a.	(Yang et al., 2010)
GAG-HCC	Age, sex, HBV-DNA, core promoter mutations, cirrhosis	HBV	n.a.	(Yuen et al., 2009)
Hung et al	Sex, age, ALT, previous liver disease, history of HCC, smoking, HBV/HCV infection	HBV	n.a.	(Hung et al., 2015)
FIB-4	AST, ALT, platelets, age	HBV	n.a.	(Suh et al., 2015)
D2AS risk score	HBV-DNA, sex, age	HBV	Asian	(Sinn et al., 2017)
HCC-RESCUE	Age, sex, cirrhosis	HBV	Asian	(Sohn et al., 2017)
GBM-based model	Cirrhosis, age, platelets, ETV or TDF, sex, ALT, HBV-DNA, albumin, bilirubin, HBeAg	HBV	Asian, Caucasian	(H. Y. Kim et al., 2021)
modified REACH-B	Liver stiffness, sex, age, ALT, HBeAg	HBV treated with NA	Asian	(H. W. Lee et al., 2014)
PAGE-B	Age, sex, platelets	HBV treated with NA	Caucasian	(G. Papatheodorid is et al., 2016)
CAGE-B	cirrhosis, age	HBV treated with NA	Caucasian	(G. V. Papatheodorid is et al., 2020)
SAGE-B	Liver stiffness, age	HBV treated with NA	Caucasian	(G. V. Papatheodorid is et al., 2020)
modified PAGE-B	Age, sex, platelets, albumin	HBV treated with NA	Asian	(Kim et al., 2018)
APA-B	Age, platelets, AFP	HBV treated with NA	Asian	(C. H. Chen et al., 2017)
CAMPAS model score	Cirrhosis, age, sex, platelets, albumin, liver stiffness	HBV treated with NA	Asian	(Lee et al., 2020)
REAL-B	Sex, age, alcohol, diabetes, cirrhosis, platelets, AFP	HBV treated with NA	Asian	(Yang et al., 2020)
AASL-HCC score	Age, albumin, sex, cirrhosis	HBV treated with NA	Asian	(Yu et al., 2019)
CAMD score	Cirrhosis, age, sex, diabetes	HBV treated with NA	Asian	(Hsu, Yip, et al., 2018)
ALT Flare	ALT	HBV treated with NA	Asian, n.a.	(Du et al., 2020)
Ganne-Carri et al	Age, alcohol, platelets, GGT, SVR	HCV	n.a.	(Ganne-Carrié et al., 2016)
REVEAL-HCV	Age, ALT, AST/ALT ratio, HCV-RNA, cirrhosis, HCV genotype	HCV	n.a.	(M. H. Lee et al., 2014)
ADRES score	SVR24, sex, FIB-4, AFP	HCV-SVR treated with DAA	Asian	(Hiraoka et al., 2019)
HCC-SVR score	FIB-4, sex, AFP	HCV-SVR treated with DAA	Asian	(Chun et al., 2020)
Sinn et al	Age, sex, smoking, diabetes, total cholesterol, ALT	non-HCV, HBV, Alcohol	Asian	(Sinn et al., 2020)

AASL, age, albumin, sex, liver cirrhosis; ADRES, after DAAs recommendation for surveillance; ADRESS, age, diabetes, race, etiology of cirrhosis, sex and severity of liver dysfunction; AFP, alpha-fetoprotein; AIH, autoimmune hepatitis; ALP, alkaline phosphatase; ALT, alanine aminotransferase; aMAP, age, male, albumin-bilirubin, platelets; APA; age, platelet, alpha-fetoprotein; AST, aspartate aminotransferase; CAGE, cirrhosis and age; CAMD, cirrhosis, age, male sex, and diabetes mellitus; CAMPAS, cirrhosis, age, male, platelets, albumin, liver stiffness; CU, Chinese University; DAA, direct-acting antivirals; ETV, entecavir; FIB-4, fibrosis-4; FILI, fibrosis improvement after lifestyle interventions; GAG, guide with age, gender, HBV-DNA, core promoter mutations and cirrhosis; GGT, gamma-glutamyltransferase; GMB, gradient-boosting machine; HbA1c, hemoglobin A1c; HBV, hepatitis B virus; HBeAg, hepatitis B e antigen; HCV, hepatitis C virus; HCC, hepatocellular carcinoma; LSM, liver stiffness measurement; n.a., not available/applicable; NA, nucleoside/nucleotide analogues; NASH, non-alcoholic steatohepatitis; NGM1 or 2-HCC, nomogram 1 or 2-HCC; PAGE, platelets, age, gender; PBC, primary biliary cholangitis; REAL, real-world effectiveness from the Asia Pacific Rim Liver Consortium; REVEAL, risk evaluation of viral load elevation and associated liver disease/cancer; RWS, Real-world risk score; REACH-B, risk estimation for HCC in chronic hepatitis B; SAGE, stiffness and age; SVR, sustained virologic response; TDF, tenofovir disoproxil fumarate; THRI, Toronto HCC risk index; UM, University of Michigan; VFMAP, virtual touch quantification, fast plasma glucose, sex, age, AFP.

2.1. Viral hepatitis

2.1.1. Hepatitis B virus

HBV has been the leading cause of chronic viral hepatitis and HCC, affecting 3.0-5.5% of global population with a wide range of geographic variation; the prevalence of HBsAg positivity among general population is 7.5% in Africa, 5.9% in Western Pacific region, and 3.0% in Southeast Asia, whereas 1.5% in Europe and 0.5% in America (World Health Organization, 2021). National universal vaccination programs have substantially reduced incidence of HBV-related HCC. In Taiwan where the neonatal vaccination was pioneered, HCC incidence per 10⁵ person-years was reduced from 0.92 to 0.23 with the vaccination (Chang et al., 2016). In the current care of HBV-infected individuals typically on treatment with viral-replication-suppressing oral nucleos(t)ide analogues (NAs) such as tenofovir and entecavir, annual HCC incidence rates are 0.0-1.4% in noncirrhotic and 0.9-5.4% in cirrhotic patients (Choi, Choi, & Lim, 2021). The regular HCC screening is particularly effective in patients with chronic hepatitis B to significantly reduce HCC-related mortality, while its clinical implementation needs improvement (Su et al., 2021; Yeo & Nguyen, 2021).

Clinically, co-infection of hepatitis delta virus and aflatoxin exposure are known to elevate risk of HBV-related HCC (Fujiwara et al., 2018). High blood HBV DNA level is an indicator of active viral replication, which is associated with increased incidence of HBV-related HCC and likely related to HBV DNA integration into host genome, enhancing carcinogenesis even in non-fibrotic/cirrhotic livers (C. J. Chen et al., 2006; J. D. Yang et al., 2011). Several clinical-variable-based HCC risk scores have been proposed mainly from the endemic regions, particularly East Asia and Europe to date, in HBV-infected individuals with or without NA treatment (Table 1). These scores consist of similar variables such as patient age and sex, AFP, and other biochemical tests, and their cross-region/population validations have shown generally good performance in predicting HCC risk with AUROCs of 0.70 ~ 0.80 (H. S. Kim et al., 2021; Voulgaris, Papatheodoridi, Lampertico, & Papatheodoridis, 2020). These scores may identify patients who need or do not need regular HCC screening and biomarker-based more refined HCC risk prediction.

2.1.2. Hepatitis C virus

HCV has been a major viral HCC etiology globally, especially in developed countries over the past decades, affecting more than 70 million individuals (1% of global population) (Roudot-Thoraval, 2021). The prevalence of HCV is relatively high in Eastern Mediterranean region (1.6%), Europe (1.3%), moderate in Africa (0.8%), and relatively low in Southeast Asia (0.5%), Western Pacific region (0.5%), and America (0.5%) (World Health Organization, 2021). Recent development of highly effective and safe oral direct-acting antivirals (DAA) has revolutionized the care of patients with chronic hepatitis C, although majority of the patients are still undiagnosed and/or untreated (Yousafzai et al., 2021). HCC risk proportionally increases as liver fibrosis progresses, and approximately 80% of HCC patients have underlying cirrhosis (Fujiwara et al., 2018). HCV viral genetic diversity such as genotypes and quasispecies are known to influence the disease progression toward HCC development (Martinez & Franco, 2020). With the shift of anti-HCV therapies from the interferon-based regimens to DAAs, clinically relevant “high-risk” variants are also shifting (e.g., genotype 1b to 3) according to the therapeutic response. Pharmacological HCV clearance, namely sustained virologic response (SVR), substantially reduces future HCC incidence, although the HCC risk is not eliminated over a decade even after achieving an SVR (Baumert, Juhling, Ono, & Hoshida, 2017; Carrat et al., 2019). Clinical-variable-based HCC risk scores have been developed and/or tested in SVR patients, although their external validation is still limited compared to the HBV HCC risk scores (Table 1). These studies also suggest variation in their performance according to patient race/ethnicity, which should be elucidated in future studies.

2.2. Metabolic disorders

2.2.1. Alcohol abuse

Alcohol-related liver cirrhosis is one of the well-known risks of developing HCC, and the proportion of HCC attributed to alcoholic liver disease is estimated to be consistently 20-25% over the past years (Massarweh & El-Serag, 2017). The HCC incidence rate in patients with alcoholic cirrhosis is reported to be 1.3-3% annually (Morgan, Mandayam, & Jamal, 2004)., and alcohol abuse shows synergistic effects with other etiologies including viral hepatitis on the progression of liver fibrosis and hepatocarcinogenesis (Palmer & Patel, 2012). In fact, several clinical HCC risk scores established in viral hepatitis patients such as NGM-HCC and REAL-B include alcohol abuse as one of their model variables (Ganne-Carrié et al., 2016; Yang et al., 2010; Yang et al., 2020).

2.2.2. Non-alcoholic fatty liver disease

The global prevalence of non-alcoholic fatty liver disease (NAFLD) is estimated to be 25% with an increasing trend (D. Q. Huang, El-Serag, & Loomba, 2021; Younossi et al., 2019). With the newly proposed metabolic dysfunction-associated fatty liver disease (MAFLD), the prevalence is assumed to be even higher (50%) (J. Liu et al., 2021). Both prevalence and incidence of NAFLD-related HCC are predicted to significantly increase over the next decade (Estes, Razavi, Loomba, Younossi, & Sanyal, 2018). The annual HCC incidence rates in histologically-confirmed NAFLD are less than 0.5% in patients with non-cirrhotic fibrosis and less than 1% even with cirrhosis (Sanyal et al., 2021; Simon et al., 2021). HCC development in non-cirrhotic liver is one notable feature of NAFLD, which is observed approximately in 30% of the patients (Mittal et al., 2016). These characteristics, i.e., sharply expanding at-risk population, low incidence rate, and carcinogenesis without involving cirrhosis, highlight urgent need for new strategies of HCC screening based on refined HCC risk prediction to enable cost-effective care of NAFLD patients (Singal & El-Serag, 2021). Several clinical HCC risk scores such as the Age, Diabetes, Race, Etiology of cirrhosis, Sex, and Severity (ADRESS)-HCC score and Toronto HCC risk index (THRI) have been tested in NAFLD patients in the context of multi-etiology patient cohorts (Table 1). There are also HCC risk scores specifically modeled in NAFLD patients (Ioannou et al., 2019)

2.3. Patient sex and race/ethnicity

2.3.1. Sex

Clinically, it is well known that male sex is associated with two- to three-fold higher HCC incidence and mortality compared to females irrespective of geographic regions, which is interestingly not obvious in intrahepatic cholangiocarcinoma, the second common histological type of liver cancer (Massarweh & El-Serag, 2017). Liver cancer, dominantly HCC, is the third leading cause of cancer-related death in men, whereas it is the sixth leading cause in women (Sung et al., 2021). HCC prognosis is better in women compared to men in patients under 65 years old, while such difference disappears in older patients, suggesting involvement of sex hormones as an underlying biological factor (Rich, Murphy, et al., 2020). In addition to the biological difference, this sex disparity is likely attributable to complex interplays between multiple factors, including biological differences (e.g., sex hormone-related molecular pathways) as well as behavioral and environmental factors (drinking, smoking, dietary habit), and socioeconomic factors associated with patient race/ethnicity (Thylur et al., 2020), to be clarified in future studies. Particularly, there may be specific patient subgroups with unique (including opposite) association of sex with HCC risk. Given the unequivocal association with HCC risk, sex is indeed incorporated in most of the clinical HCC risk scores (Table 1).

2.3.2. Race/ethnicity

Patient race/ethnicity has been increasingly recognized as a confounding factor that influences HCC risk, which is often bound to specific HCC etiology prevalent in certain geographic regions, e.g., HBV infection in Southeast Asia and sub-Saharan Africa, where the patients reside or immigrate from (Fujiwara et al., 2018). Epidemiological studies have shown that the trend of HCC incidence is distinctly different between racial/ethnic groups, reflecting dynamic change in each regional population. For instance, HCC incidence has been decreasing in Asian and Pacific islanders that have been the leading HCC risk populations, whereas increasing in other racial/ethnic groups particularly Hispanics and Blacks compared to non-Hispanic whites over the past several decades in the U.S. (Franco, Fan, Jarosek, Bae, & Galbraith, 2018; Islami et al., 2017). HCC incidence rate in Hispanics is approximately two-times higher (21.2/100,000) compared to non-Hispanic Whites (9.3/100,000) (El-Serag, Sardell, Thrift, Kanwal, & Miller, 2021; Miller et al., 2021). Survival after HCC diagnosis is worse in Blacks likely due to diagnosis at late stages, whereas better in Asians and Hispanics, compared to Whites (Rich, Carr, Yopp, Marrero, & Singal, 2020). Race/ethnicity can be confounded with multiple factors, e.g., relative difference in prevalence of inheritable genetic variants associated with HCC risk, lifestyle (particularly dietary habits), and socioeconomic status influencing access to medical care, in a highly complex manner (Thylur et al., 2020). Given the complexity, no clinical HCC risk score incorporates race/ethnicity as its component. In addition, derivation and validation of clinical risk scores are substantially affected by biased distribution of race/ethnicity-related variables, and therefore generalization of the scores’ performance is compromised. Nevertheless, several attempts to cross-validate some clinical risk scores across geographic regions (e.g., Asia vs. Europe), which may partially address the issue (Voulgaris et al., 2020). More studies at population level will be needed to clarify the influence of race/ethnicity on HCC risk under specific regional and clinical contexts.

2.4. Lifestyle factors and relatively rare contexts of hepatocarcinogenesis

Dietary factors such as Western diet and intake of anti-oxidative substances such as coffee have been associated with reduced HCC risk (Simon & Chan, 2020). Food contamination with carcinogen such as aflatoxin is a well-known HCC risk factor that leads to impaired tumor suppressor genes, e.g., TP53, and is associated with certain DNA mutational signatures (Letouze et al., 2017; McGlynn, Petrick, & El-Serag, 2021). Tobacco smoking is associated with 40-70% increase of HCC risk (McGlynn et al., 2021). Use of generic drugs such as statins, aspirin, and metformin has been associated with lowered HCC risk, indicating their potential utility as chemopreventatives (Athuluri-Divakar & Hoshida, 2019). Other rare suspected habitual and/or dietary HCC risk factors include betel quid chewing, exposure to organic solvents, and heavy metals (e.g., toluene, trichloroethylene, cadmium, and lead) (Thylur et al., 2020). Physical activity is also associated with HCC risk as a potential intervention (Fujiwara et al., 2018). Among these factors, only smoking has been incorporated in clinical HCC risk scores (Table 1). Future studies should explore other factors, particularly modifiable by dietary, pharmacological, and/or physical interventions, to be incorporated in HCC risk scores that could guide therapeutic decision making.

3. Molecular indicators of high-risk liver

Various types of biomolecules such as nucleic acids, proteins, and metabolites, have been actively explored as indicators of elevated risk of HCC development as summarized in Table 2. It has been increasingly recognized that association of these factors with HCC risk is substantially affected by specific clinical contexts such as liver disease etiology as outlined in this section.

Table 2. Molecular HCC risk indicators.

Type of biomarker	Biomarkers / scores	Variables	Major etiology	Major race/ethnicity	Reference
SNP	EGF SNP	EGF 61AG (rs4444903, A>G) IFNL3	HBV, HCV	Asian, European, African	(G. Jiang et al., 2015)
	IFNL3 SNP	(rs12979860: C>T, rs8099917: T>G)	HCV, HBV	n.a.	(Qin et al., 2019)
	MICA SNP	MICA (rs2596542, C>T)	HCV, HBV	Asian, European	(X. Luo, Wang, Shen, Deng, & Ye, 2019)
	DEPDC5 SNP	DEPDC5 (rs1012068: T>G)	HCV	Asian	(W. Liu et al., 2019)
	AURKA SNP	AURKA (rs1047972: G>A)	HCV	n.a.	(Farid, Afify, Alsharnoby, Abdelsameea, & Bedair, 2021)
	TLL1 SNP	TLL1 (rs17047200: A>T)	HCV after SVR treated with IFN	Asian	(Matsuura et al., 2017)
	CHI3L1 SNP Intergenic polymorphism	CHI3L1 (rs880633: T>C) intergenic polymorphism (rs597533: A>G) KIF1B or	HCV treated with DAA	n.a.	(Mangoud, Ali, El Kassas, & Soror, 2021)
	KIF1B or 1p36.22 SNP	1p36.22 (rs1740196 6, A>G)	HBV	Asian	(Y. Y. Luo et al., 2019)
	STAT4 SNP	STAT4 (rs7574865, G>T)	HBV	Asian	(L. Zhang et al., 2017)
	HLA-DQB1/HLA-DBA2 SNP	HLA-DQB1/HLA-DBA2 (rs9275319 A>G) Inc-ACACA-1 rs9908998,	HBV	Asian	(D. K. Jiang et al., 2013)
	IncRNA SNP	Inc-RP11-150O12.3 rs2275959, rs1008547, rs11776545	HBV	Asian	(Q. Liu et al., 2021)
	PNPLA3 SNP	PNPLA3 (rs738409: C>G)	NAFLD, alcohol, HCV	Caucasian	(Singal et al., 2014)
	TM6SF2 SNP	TM6SF2 (rs58542926: C>T) HSD17B13	Alcohol	Caucasian	(Tang et al., 2019)
	HSD17B13 SNP	(rs72613567: TA)	Alcohol	Caucasian	(Stickel et al., 2020)
	WNT3A-WNT9A SNP	WNT3A-WNT9A (rs708113: T>A)	Alcohol	n.a.	(Trépo et al., 2021)
	MBOAT7 SNP	MBOAT7 (rs641738: C>T) CELSR2-	NAFLD	Caucasian	(Donati et al., 2017)
	CELSR2-PSRC1-SORT1 SNP	PSRC1-SORT1 (rs599839; A>G) IFNGR1	NAFLD	n.a.	(Meroni et al., 2021)
	IFNGR1 SNP	(rs1327474: G>A)	n.a.	n.a.	(Aref et al., 2021)
Panel of SNPs
	Genetic risk score	SNPs of PNPLA3, TM6SF2, HSD17B13	General population	Caucasian, n.a.	(Gellert-Kristensen et al., 2020)
	Fat-genetic risk score (hepatic fat genetic risk score)	SNPs of PNPLA3, TM6SF2, MBOAT7,	HCV treated with DAA	Italian, Egyptian	(Degasperi et al., 2020)
		GCKR, and hepatic fat content SNPs of PNPLA3, TM6SF2, MBOAT7, GCKR, and hepatic fat content	NAFLD, general population
	Polygenic risk scores	SNPs of PNPLA3, TM6SF2, MBOAT7, GCKR, and hepatic fat + HSD17B13	n.a.	(Bianco et al., 2021)
Tissue transcript ome
	Prognostic liver signature (PLS)	186-gene signature	HCV	Asian, Caucasian	(Hoshida et al., 2013)
	HIR gene signature	233-gene signature	HBV	Asian	(Kim et al., 2014)
	Activated HSC gene signature	37-gene signature	HBV	Asian	(Ji et al., 2015)
	HSC signature	122-gene signature	HCV, HBV	Caucasian, Asian	(D. Y. Zhang et al., 2016)
	Ectopic lymphoid structure signature	12-gene signature	HCV	Asian	(Finkin et al., 2015)
	Immune mediated cancer field signature	172-gene signature	HCV	Caucasian	(Moeini et al., 2019)
Circulating biomolecule
	cfDNA	mutations of 4 genes, HBV integration	HBV	Asian	(Qu et al., 2019)
	miRNA	7/8 miRNAs	HBV	Asian	(C. Wang et al., 2016)
	miRNA	16 miRNAs	HBV, HCV	Asian	(Y. H. Huang et al., 2017)
	DNA methylation	TBX2 hypermethy lation	HBV, HCV, alcohol	Asian	(Wu et al., 2017)
	GlycoCirrhoTest/GlycoHC CRiskScore	serum protein N-glycans	HCV	Caucasian	(Verhelst et al., 2017)
	Serum glycan	M2BPGi	HCV	Asian	(Yamasaki et al., 2014)
	Cytokine	IL-6	HCV	Asian	(H. Nakagawa et al., 2009)
	Cytokine	IL-17	HBV, HCV	Asian	(K. H. Liang et al., 2021)
	Cytokine	IL-27	HBV	Asian	(Yuan et al., 2021)
	Cytokine	Myostatin	Alcohol	Asian	(J. H. Kim et al., 2020)
	Protein; PLSec/Secretome signature	High-risk; VCAM-1, IGFBP-7, gp130, matrilysin, IL-6, CCL-21 Low-risk; angiogenin, protein S	HCV-SVR (DAA), NAFLD/cryptogenic	Caucasian, Asian	(Fujiwara, Kobayashi, et al., 2021)
		IGF-1	HCV	Caucasian	(Mazziotti et al., 2002)
	Protein	OPN	Alcohol, HBV, HCV	n.a.	(Duarte-Salles et al., 2016)
	Protein; LCR1/LCR2	ApoA1, Hp/A2M	HCV, HBV, ALD, NAFLD	Caucasian, Sub-Saharan, North-Africa Middle East, Asian	(Poynard et al., 2019)
	Metabolites	14 metabolites	HBV, HCV, alcohol	n.a.	(Stepien et al., 2021)
	Metabolites; HCC risk score R	2 amino acids (Phe, Gln)	HBV, HCV	Asian	(K. H. Liang et al., 2020)
	Metabolites9	AST and 7 metabolites, including tyrosine, oleamide,	HBV	Asian	(Jee et al., 2018)
		lysoPC 16:1, lysoPC 20:3, 5-hydroxyhex anoic acid, androsterone sulfate, and TUDCA 2 metabolites
	Metabolites	(phenylalanyl-tryptophan, glycocholate)	HBV	Asian	(P. Luo et al., 2018)
	Virome; viral exposure signature	61 viral strains	HBV, HCV, HDV, aflatoxins, alcohol, NAFLD	Caucasian, Black, Asian	(J. Liu et al., 2020)

AFP, alpha-fetoprotein; cfDNA, circulating free DNA; DAA, direct-acting antiviral agent; HBV, hepatitis B virus; HCV, hepatitis C virus; HDV, hepatitis D virus; HCC, hepatocellular carcinoma; HIR, hepatic injury and regeneration gene expression; HSC, hepatic stellate cell; IFN, interferon; M2BPGi, mac-2 binding protein glycosylation isomer; miRNA, microRNA; n.a., not available/applicable; NA, nucleoside/nucleotide analogues; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; SNP, single-nucleotide polymorphism; SVR, sustained virologic response.

3.1. Dysregulations in hepatic transcriptome

Tissue transcriptome is one of the most widely studied omics data as a source of exploring HCC risk biomarkers (Kubota et al., 2020). One example is Prognostic Liver Signature (PLS) predictive of long-term HCC risk in patients with diverse clinical contexts such as HBV infection, HCV infection with or without pharmacological viral cure (i.e., SVR), alcohol abuse, and NAFLD/NASH, which is already implemented in an FDA-approved diagnostic assay platform as a Laboratory Developed Test (LDT) (Hoshida et al., 2008; Hoshida et al., 2013; King et al., 2015; S. Nakagawa et al., 2016; Ono et al., 2017). Of note, PLS is also implemented in a cell culture-based in vitro system, namely cell-ulture-derived PLS (cPLS), for high-throughput HCC chemoprevention drug screening (Crouchet et al., 2021). A Hepatic Injury and Regeneration (HIR) signature was derived from liver tissues in HBV-related HCC patients to predict de novo HCC recurrence (Kim et al., 2014). Transcriptomic signatures can be defined to detect pathogenic status of hepatic cell types such as hepatic stellate cells, the driver of liver fibrogenesis and carcinogenesis in viral hepatitis and NAFLD patients (Higashi, Friedman, & Hoshida, 2017; Ji et al., 2015; D. Y. Zhang et al., 2016). Gene signatures of histological architecture can also be trained; a 12-chemokine-gene signature identified presence of HCC-risk-associated lymphocyte aggregate, ectopic lymphoid structure (ELS) in mainly HCV-infected patients (Finkin et al., 2015).

3.2. Dysregulations in hepatic proteome and secretome

Hepatic proteomic dysregulations that lead to alterations of secretome can have a potential as non-invasive biomarkers in circulation to monitor molecular pathogenic status of the liver. One classical example of such molecule is alpha-fetoprotein (AFP), which is an oncofetal protein and used as an HCC tumor marker. Interestingly, elevation of AFP can be observed years before clinically detectable HCC development, suggesting that it captures carcinogenesis-prone status of diseased liver tissue microenvironment (Hughes et al., 2021). Serum cytokines, including IL-6, IL-17, and IL-27, as well as other types of serum proteins such as laminin γ2 monomer and IGF-I, were reported to be associated with HCC risk (J. H. Kim et al., 2020; K. H. Liang et al., 2021; Mazziotti et al., 2002; H. Nakagawa et al., 2009; Yamashita et al., 2021; Yuan et al., 2021) (Table 2). The hepatic-transcriptome-based PLS was translated into a serum-protein-based surrogate marker, Prognostic Liver Secretome signature (PLSec), by utilizing a computational pipeline translating tissue transcriptome to secretome (Translation of tissue gene expression to SECretome; TexSEC, www.texsec-app.org) (Fujiwara, Kobayashi, et al., 2021). PLSec in combination with AFP (PLSec-AFP) was validated for association with long-term HCC risk and hepatic decompensation in HCV-infected patients who achieved SVR and cirrhosis patients with mixed etiologies (Fujiwara, Fobar, et al., 2021; Fujiwara, Kobayashi, et al., 2021). The TexSEC pipeline also identified a secretome signature predictive of survival in severe alcoholic hepatitis patients (Fujiwara, Trépo, et al., 2021).

3.3. Metabolic dysregulations

Since liver is the major organ responsible for the essential metabolic functions in human body, measuring dysregulations of metabolites is a sensible approach to monitor molecular status of the liver at risk of carcinogenesis (Beyoğlu & Idle, 2020). Global metabolite profiling has been performed by using mass spectrometry (MS) and/or nuclear magnetic resonance spectroscopy (NMR) (Beyoğlu & Idle, 2013). Serum NMR analysis of European Prospective Investigation into Cancer and Nutrition (EPIC) cohort revealed that circulating metabolites were correlated with various lifestyles or environmental exposures associated with HCC risk (Assi et al., 2015). Unbiased liquid-chromatography-MS study of EPIC cohort identified 14 metabolites associated with long-term HCC risk, including 9 high-risk (N1-acetylspermidine, isatin, p-hydroxyphenyllactic acid, tyrosine, sphingosine, L,L-cyclo(leucylprolyl), glycochenodeoxycholic acid, glycocholic acid and 7-methylguanine) and 5 low-risk (retinol, dehydroepiandrosterone sulfate, glycerophosphocholine, γ-carboxyethyl hydroxychroman and creatine) metabolites (Stepien et al., 2021). In a Korean cohort, seven metabolites (tyrosine, oleamide, lysophosphatidylcholines 16:1, lysophosphatidylcholines 20:3, 5-hydroxyhexanoic acid, androsterone sulfate, and tauroursodeoxycholic acid) were identified by using non-targeted liquid-chromatography-MS (Jee et al., 2018). Plasma phenylalanine and glutamine levels were associated with HCC incidence in Asian patients with mainly viral hepatitis (K. H. Liang et al., 2020). Phenylalanyl-tryptophan and glycocholate were also identified as a serum metabolite biomarker panel in combination with AFP, which predicts presence of preclinical HCC before clinical diagnosis (P. Luo et al., 2018).

3.4. Other liver-derived biomolecules

Serum Mac-2-binding protein glycosylation isomer (M2BPGi) is a serum glycan biomarker that can predict the development of HCC in the patients with various etiologies including HCV and HBV treated with oral antiviral therapy (Hsu, Jun, et al., 2018; Shinkai et al., 2018; Tseng et al., 2020; Yamasaki et al., 2014). GlycoCirrhoTest, a serum glycomics biomarker which could distinguish chronic liver disease patients with compensated cirrhosis from those with earlier stage of fibrosis, was reported to be associated with future HCC risk in cirrhotic patients (Verhelst et al., 2017). Circulating nucleic acids, e.g., cell-free (cf) DNA and RNA, are another class of non-invasive biomarker candidates. For example, TBX2 hypermethylation in plasma was associated with increased HCC risk in an Asian cohort (Wu, Yang, Wang, Chen, & Santella, 2017). Hepatocellular carcinoma screen (HCCscreen) score, which is composed of genetic alterations in cfDNAs (TP53, CTNNB1, AXIN1, and TERT) as well as serum AFP and des-γ-carboxy prothrombin levels, was associated with HBV-related HCC occurrence within 6-8 months (positive predictive value of 0.17) (Qu et al., 2019). Circulating micro-RNA (miRNA) was also reported to be associated with HCC risk. An miRNA-based risk score consisting of 15 miRNAs with AFP was associated with 5-year HCC incidence in HBV-positive individuals with Korean ancestry (C. Wang et al., 2016). Another 16-miRNA-based risk score was also associated with HCC occurrence in HBV- or HCV-related cirrhotic patients (Y. H. Huang et al., 2017). Nucleic acids and proteins loaded in extracellular vesicle (EV), a lipid bilayer-delimited particle secreted by liver cells, may serve as a new class of biomolecules to predict future HCC risk (Adeniji & Dhanasekaran, 2021).

4. Host and systemic factors associated with liver cancer risk

The above-mentioned clinical contexts are assumed to be accompanied with underlying various host genetic variants/aberrations and systemic molecular dysregulations that often represent cross-organ interactions to develop cancer-permissive hepatic tissue microenvironment. These factors in literature are overviewed in this section.

4.1. Germline and somatic genetic variants/polymorphisms

Inheritable genetic polymorphisms, particularly single nucleotide polymorphisms (SNPs), have been heavily studied given the easy access as biomarkers via readily available specimens such as buccal swab in daily clinical practice, and numerous HCC-risk-associated SNPs were reported (Kubota et al., 2020) (Table 2). Prevalence of risk alleles/genotypes is often biased across different racial/ethnic groups, suggesting their influence in determining race/ethnicity-specific susceptibility to HCC in response to clinical and environmental risk factors. Magnitude of association with HCC risk for these SNPs is generally modest, e.g., odds ratio of 1.5 or less, and multi-SNP “polygenic” scores have been explored (Bianco et al., 2021; Fujiwara et al., 2018). While germline DNA variants are easy to assay, they are unmodifiable by therapeutic intervention like sex.

Several SNPs were discovered and validated in patients with specific etiology. SNPs in DEPDC5A and TLL1 genes were identified as HCC risk indicators in HCV-infected patients, whereas SNPs in KIF1B and STAT4 genes were reported in HBV-infected patients (W. Liu et al., 2019; Y. Y. Luo, Zhang, Huang, & Hu, 2019; Matsuura et al., 2017; L. Zhang, Xu, Liu, & Chen, 2017). SNPs in MICA, IFNL3, and EGF genes were associated with HCC risk among individuals with either HCV or HBV infection (G. Jiang et al., 2015; Qin et al., 2019). SNPs in PNPLA3, TM6SF2, and HSD17B13 that were initially discovered for association with NASH, were also associated with HCC risk among NAFLD, HCV-SVR, and alcoholic hepatitis patients (Degasperi et al., 2020; Gellert-Kristensen et al., 2020; Kozlitina et al., 2014; Singal et al., 2014; Tang et al., 2019). Recently, a SNP in WNT3A-WNT9A has been identified as a risk factor of alcohol-related HCC (Trépo et al., 2021). Polygenic scores consisting of multiple SNPs (e.g., SNPs in PNPLA3, MBOAT7, TM6SF2, and GCKR) were currently actively explored in HCV-SVR and NAFLD patients (Bianco et al., 2021; Degasperi et al., 2020). Somatic DNA mutations in PKD1, KMT2D, and ARID1A genes in cirrhotic liver may have protective role from carcinogenesis (Zhu et al., 2019).

4.2. Systemic inflammation and disorders

Metabolic disorders, often associated with obesity and NAFLD, are well known drivers of systemic and hepatic inflammation mediated by various cytokines/chemokines such as interleukin-17 that lead to hepatocarcinogenesis, which could involve interaction between multiple organs/tissues such as adipose tissues, brain, and gut. Obesity, diabetes, and metabolic syndrome are associated with two- to three-fold increase of HCC incidence (Fujiwara et al., 2018). Use of drugs to treat diabetes (e.g., metformin) and hyperlipidemia (e.g., lipophilic statins) is associated with reduced HCC risk, suggesting that biomarkers related to response to these drugs may have roles in HCC risk prediction and monitoring (Athuluri-Divakar & Hoshida, 2019).

Accumulation of ionic metal such as iron is known to increase oxidative stress and genetic damage that lead to carcinogenesis. Hereditary hemochromatosis (HH) is a genetic disorder with excessive iron absorption mostly caused by germline variants in HFE gene, and a known HCC risk condition, whereas iron overload can also elicit similar risk (Fujiwara et al., 2018). Autoimmune liver diseases such as autoimmune hepatitis (AIH) and primary biliary cholangitis (PBC) develop progressive liver fibrosis, and are at risk of developing HCC and/or cholangiocarcinoma once cirrhosis is established (Y. Liang, Yang, & Zhong, 2012; Marrero et al., 2018). However, their incidence and prevalence are low (annual incidence of up to 1% or far less) because of the small patient population. In contrast, despite the similarly low incidence rates, prevalence is high in patients with NAFLD and HCV-SVR due to the vast size of patient population.

Alpha1-antitrypsin deficiency, caused by mutation in SERPINA1 gene, is also associated with increased risk of developing both cirrhosis and HCC along with pulmonary emphysema (Strnad, McElvaney, & Lomas, 2020). Porphyrias, caused by dysfunctional heme biosynthesis, are associated with elevated risk of primary liver cancer (B. Wang, Rudnick, Cengia, & Bonkovsky, 2019). Other rare genetic diseases associated with HCC risk include glycogen storage disease type I and hereditary tyrosinemia type I (Dragani, 2010).

Genetic integration of adeno-associated virus (AAV), a virus belongs to the parvovirus group, into several cancer driver genes was identified in a small subset of HCCs, however, the potential risk of AAV on hepatocarcinogenesis is yet to be established (Nault et al., 2015; Schäffer et al., 2021).

4.3. Extrahepatic microorganisms

4.3.1. Gut microbiome

Intestinal bacterial flora (or gut microbiota) consists of more than 10¹⁴ microorganisms, representing more than 10⁴ bacterial species, and changes in its composition (i.e., dysbiosis) are expected to serve as a source to identify HCC risk biomarkers as well as therapeutic targets to alter the risk (Schwabe & Greten, 2020; Zhou et al., 2020). Preclinical studies in rodent models have identified several cellular signaling such as toll-like receptor (TLR) pathway linking gut dysbiosis and HCC risk, which could be disrupted by gut sterilization, as well as physiological pathway such as bile acid metabolism that shape the gut-liver axis, affecting HCC risk (Thilakarathna, Rupasinghe, & Ridgway, 2021). Multiple studies in human HCC patients with various etiologies have shown that there are certain protective or harmful bacterial species with regard to HCC risk and biological/clinical consequence (Xu et al., 2021). Panels of gut bacteria (e.g., Enterococcus, Limnobacter, and Phyllobacterium) were shown to be associated with presence of HCC specific to or regardless of etiology (Hernandez et al., 2021; Ren et al., 2019; Zheng et al., 2020). Oral cyanobacteria was associated with HCC risk in patients with liver disease from mixed etiologies (Hernandez et al., 2021). Gut dysbiosis and its HCC risk association are likely influenced by specific dietary habits, host genetic factors, and geographic setting the patients reside, among many other factors, and therefore their characterization is expected to contribute to refining clinical-context-specific HCC risk prediction and chemopreventive intervention.

4.3.2. History of viral exposure

A Viral Exposure Signature (VES), composed of 61 viral strains determined using VirScan, a virome technology for detecting the exposure history to all known human viruses, was associated with future HCC development (J. Liu et al., 2020). Interestingly, this study revealed that some viruses such as human herpesvirus 5 were associated with future HCC development as previously unknown HCC risk factors (Fujiwara & Hoshida, 2020).

5. Clinical implementation of molecular liver cancer risk assessment

HCC risk assessment should ultimately be incorporated in the algorithm of regular HCC screening to improve early tumor detection. Even if research studies confirm clinical utility of new HCC risk biomarkers and/or scores according to specific clinical context, there are multi-fold obstacles to hamper their clinical translation as outlined in this section.

5.1. Physicians’ and patients’ perspectives

Even the guideline-recommended use of imaging- and blood-based biomarkers (i.e., ultrasound and AFP) is poorly utilized in real-world clinical practice (utilization rate < 25%) due to multiple provider- and patient-related factors (Wolf et al., 2021; Zhao & Nguyen, 2016). Obstacles on the physicians’ side include limited knowledge about HCC screening particularly in primary care providers, logistical barriers such as time constraint in clinic, and the overwhelmingly large target patient population to regularly monitor for HCC development (Dalton-Fitzgerald et al., 2015; Dirchwolf et al., 2021; Fujiwara et al., 2018; Kubota et al., 2020; Simmons, Feng, Parikh, & Singal, 2019; Wolf et al., 2021). On the other hand, encouragingly, physicians are willing to tailor HCC screening tests if the probability of future HCC can be reliably estimated up-front (N. J. Kim et al., 2020). To facilitate use of new clinical HCC risk scores, web-based tools could be an option to consider: an example of such tool, HCC Risk Calculator (www.hccrisk.com), is available for representative HCC etiologies for clinical-context-specific risk prediction. On the patients’ side, prohibitive factors include lack of knowledge about HCC screening, high costs for the tests, and physical access to the tests (Farvardin et al., 2017; Singal et al., 2021). Nevertheless, cirrhosis patients prioritize early HCC detection over potential screening-related harms or inconvenience (Woolen et al., 2021). Thus, education and outreach are critical in overcoming the issue for both physicians and patients to improve intake to the screening, considering clinical contexts that are often bound to racial/ethnic and socioeconomic disparities. Affordable and non-invasive tests will also lower the bar for clinical utilization of HCC screening.

5.2. Practical and regulatory issues for clinical implementation

A prior Markov-model-based simulation analysis showed that individual-risk-based personalization of HCC screening algorithm and modality is substantially cost-effective compared to the current “one-size-fits-all” strategy uniformly offering semi-annual screening with ultrasound and AFP (Goossens et al., 2017). Thus, once clinical utility of HCC risk markers and/or scores is confirmed, they may be considered for incorporation to develop such personalized HCC screening strategy (Figure 1). However, assessment and determination of their clinical utility can be challenging due to heterogeneous performance of the biomarkers/scores across various clinical contexts/scenarios. For example, association of an HCC risk biomarker with magnitude of the risk (e.g., hazard ratio in Cox regression) hugely varies according to liver disease etiology (Fujiwara, Kobayashi, et al., 2021; S. Nakagawa et al., 2016). This raises a question: to what extent should HCC risk biomarkers/scores be tailored to each specific clinical context/scenario. If a risk biomarker/score is defined in more specific clinical context, its performance is expected to be improved, while its generalizability will be compromised. In addition, clinical validation of candidate HCC risk biomarkers may become more challenging due to the confined patient population by clinical context, especially given the requirement of rigor in the phased step-wise biomarker validation for regulatory approval of the assays implemented in clinical diagnostic platform (Fujiwara et al., 2019). Smaller target patient population to apply the biomarkers may also diminish commercial interest and incentive for clinical assay development.

6. Conclusion

In this chapter, we overviewed clinical and molecular indicators of HCC risk in patients with chronic liver diseases with specific consideration about clinical context that compromises development and application of new HCC risk biomarkers and scores across heterogeneous patient populations. Given the complexity of the heterogeneity, future studies should clarify performance and utility of candidate HCC risk biomarkers and/or scores with proper characterization of target patient population to gauge their real value under well-specified clinical context/scenario to ultimately improve the poor prognosis of HCC patients. Furthermore, clinical utility of the biomarkers/scores should be revisited and redefined along with the ever-evolving clinical demographics of liver disease patients with the progresses in development of new therapeutics such as direct-acting antivirals and chemopreventive agents.

List of Abbreviations

AAV

adeno-associated virus

AFP

alpha-fetoprotein

AIH

autoimmune hepatitis

AUROC

area under the receiver operating characteristic

cPLS

cell-culture-derived Prognostic Liver Signature

DAA

direct-acting antivirals

ELS

ectopic lymphoid structure

EPIC

European Prospective Investigation into Cancer and Nutrition

FDA

U.S. Food and Drug Administration

HBV

hepatitis B virus

HCC

hepatocellular carcinoma

HCV

hepatitis C virus

HH

hereditary hemochromatosis

HIR

Hepatic Injury and Regeneration

LDT

laboratory developed test

M2BPGi

Mac-2-binding protein glycosylation isomer

MAFLD

metabolic dysregulation-associated fatty liver disease

MS

mass spectrometry

NA

nucleos(t)ide analogues

NAFLD

non-alcoholic fatty liver disease

NASH

non-alcoholic steatohepatitis

NMR

nuclear magnetic resonance spectroscopy

PBC

primary biliary cholangitis

PLS

Prognostic Liver Signature

PLSec

Prognostic Liver Secretome signature

SNP

single nucleotide polymorphisms

SVR

sustained virologic response

TexSEC

Translation of tissue gene expression to SECretome

TLR

Toll-like receptor

VES

Viral Exposure Signature