Validation of Texas Hepatocellular Carcinoma Consortium Risk Index in the Hepatocellular Carcinoma Early Detection Strategy Study

Abstract

BACKGROUND & AIMS:

We previously developed a hepatocellular carcinoma (HCC) risk stratification model, the Texas HCC Consortium Risk Index (THCC-RI), for patients with cirrhosis. In this cohort study, we aimed to externally validate THCC-RI in a prospective cohort of patients with cirrhosis.

METHODS:

We used data from the Hepatocellular Carcinoma Early Detection Strategy (HEDS) Study, a prospective cohort of patients with cirrhosis enrolled in regular HCC surveillance at 7 centers in the United States. Patients were followed from enrollment until HCC diagnosis, liver transplantation, death, or December 2023.

RESULTS:

The analysis was conducted in 1560 patients who contributed a total of 5051 person-years of follow-up (mean age, 59 years; 47% women; 40% with hepatitis C; 16% alcohol-associated disease; and 22% metabolic dysfunction-associated steatotic liver disease). Over a median follow-up of 2.5 years, 114 patients developed HCC. The THCC-RI had a C-index of 0.77 (95% confidence interval [CI], 0.64–0.85), with area under the receiver operating characteristic curve estimates of 0.77 (95% CI, 0.65–0.85) at 1 year, 0.73 (95% CI, 0.66–0.79) at 3 years, and 0.71 (95% CI, 0.65–0.77) at 5 years. THCC-RI was well-calibrated, with good agreement between observed and predicted risk. Compared with the medium-risk group (deciles 3–8), the high-risk group (deciles 9, 10) had 3.5-fold (hazard ratio, 3.5; 95% CI, 2.4–5.1) higher risk of HCC. THCC-RI had similar discrimination as the Age, Male, Albumin-bilirubin, and Platelets (aMAP) score during the first 2 years of follow-up, but the areas under the receiver operating characteristic curve for aMAP dropped more than those for THCC-RI as the time horizon expanded beyond 3 years.

CONCLUSIONS:

In an independent multi-center cohort, THCC-RI had good performance for predicting the future risk of HCC in patients with cirrhosis, with stable discrimination and calibration over a 5-year follow-up. Implementation of THCC-RI into clinical care pathways requires further research.

The incidence of hepatocellular carcinoma (HCC) in the United States has tripled in the past 3 decades.^1,2 Cirrhosis of the liver is the main risk factor for the development of HCC, with an increasing proportion of cases related to alcohol-related liver disease (ALD) and metabolic dysfunction-associated steatotic liver disease (MASLD).³ However, there is considerable variability in HCC risk among patients with cirrhosis (ranges between 1% and 8% per year).⁴ Better HCC risk stratification in cirrhosis could help in tailoring HCC surveillance intensity or type with patients’ underlying risk.⁵ Risk stratification can also promote shared decision-making between patients and providers by providing a quantifiable personalized HCC risk assessment.^6,7

Using data from a prospective contemporary cohort study, the Texas HCC Consortium Cohort (TCCC), we previously developed the Texas HCC Consortium Risk Index (THCC-RI), that includes the following predictor variables: age, sex, history of smoking, alcohol use, body mass index (BMI), hepatitis C virus (HCV) status, alpha fetoprotein (AFP), albumin, sodium, alanine aminotransferase (ALT) and platelet levels.⁸ The THCC-RI demonstrated good discrimination with an area under the receiver operating curve (AUROC) of 0.75 (95% confidence interval [CI], 0.65–0.85) at 1 year and 0.77 (95% CI, 0.71–0.83) at 2 years, and the model was well-calibrated. In a separate external validation cohort of 21,550 patients with cirrhosis from the United States Veterans Affairs system (2018–2019), AUROC values were 0.70 (95% CI, 0.67–0.74) at 1 year and 0.70 (95% CI, 0.68–0.73) at 2 years, with similarly good calibration.⁸

Despite these promising early data about THCC-RI, further validation across diverse populations and settings remains essential to establish generalizability, reliability, and potential clinical utility. The previous external validation of THCC-RI in the Veteran’s cohort may not fully capture the variability in patient characteristics, liver disease etiology, and clinical practice patterns that exist across different health systems and settings. Therefore, we conducted this external validation study in the multi-center national Hepatocellular Carcinoma Early Detection Strategy (HEDS) cohort to examine the robustness and reproducibility of THCC-RI performance in predicting HCC in patients with cirrhosis. We also compared the performance of THCCRI with Age, Male, Albumin-bilirubin, and Platelets (aMAP),⁹ an HCC predictive score, as well as a simple noninvasive test, Fibrosis-4 index (FIB-4), which has been shown in several studies to be prognostic of long-term outcomes including HCC.^10–12

Methods

Study Cohort

The HEDS was initiated as a multicenter study under the National Institutes of Health-sponsored Early Detection Research Network to explore the risk factors, incidence, and disease course of HCC in the United States with a special focus on developing novel biomarkers for early detection of HCC.¹³

The design of the HEDS study and the patient population has been described in detail in previous publications. Patients were enrolled between 2013 and 2021 from 7 centers: Mayo Clinic in Rochester, Minnesota; University of Pennsylvania in Philadelphia, Pennsylvania; Mt Sinai Medical Center in New York City, New York; University of Michigan in Ann Arbor, Michigan; Saint Louis University in Saint Louis, Missouri; Stanford University in Palo Alto, California; and UT Southwestern in Dallas, Texas.¹³ The study has been approved by the Institutional Review Board at all sites and the data coordinating center. The inclusion and exclusion criteria have been described previously.¹³ Briefly, patients were enrolled if they had a diagnosis of cirrhosis based on histology, clinical features (imaging indicating portal hypertension, splenomegaly, and thrombocytopenia in the setting of a chronic liver disease), or noninvasive testing (elastography and blood testing). Patients were excluded if they had a Model for End-stage Liver Disease (MELD) score of >15, clinically significant hepatic decompensation (Grade 3–4 encephalopathy, refractory ascites, Child-Turcotte-Pugh class C), history of cancer within 5 years, an unexplained liver mass, significant comorbid conditions with life expectancy of <1 year, or were listed for liver transplantation (LT).

Longitudinal Follow-up and Risk Factors

Patients who met the inclusion and exclusion criteria were enrolled and followed up prospectively through December 2023 according to the standard of care for cirrhosis at each site. Patients were recommended to be seen every 6 months, and followed up until HCC development, LT, or death. HCC surveillance was done according to each site and provider preference and included ultrasound and AFP, computed tomography, or magnetic resonance imaging with or without AFP.

At baseline, clinical data were collected from patient interviews and electronic medical record [EMR] reviews. These data included history of liver disease and its treatment, medical history, and laboratory data. Validated questionnaires were given to patients to assess alcohol use and smoking. Study coordinators also collected BMI (calculated from height and weight at enrollment) at the time of enrollment. Laboratory and imaging data included ultrasound/magnetic resonance imaging/computed tomography imaging results, AFP, hepatic biochemical tests, and complete blood counts. Research serum and plasma biospecimens were collected at baseline and at each follow-up visit longitudinally.

Demographics included age and self-reported sex, race, and ethnicity. HCV was defined as active (positive for serum HCV RNA) or treated (evidence of sustained virologic response [SVR]) based on data collected from the medical records. Using validated questionnaires, exposure to alcohol was categorized as current (heavy, non-heavy), former, and ever users. We used height and weight values at enrollment to calculate BMI. Etiology of liver disease was determined by the treating physicians at each site. We defined diabetes and dyslipidemia based on patient’s medical history (survey or EMR review). We extracted data for serum levels of AFP, albumin, sodium, ALT, and platelet levels at enrollment.

Primary Outcome

The primary outcome was the development of HCC after enrollment through December 2023, and this was defined using the American Association for the Study of Liver Diseases (AASLD) guidelines.¹⁴ The date of HCC confirmation, either using a lesion biopsy or radiologic interpretation as Liver Imaging Reporting and Data System (LI-RADS) 5 lesion, was used as the HCC diagnosis date. Medical record reviews were performed by each site to capture incident HCC, LT, and deaths. In the analysis of the primary endpoint, LT and death were treated as competing risks.

Statistical Analyses

We examined the performance of the THCC-RI in distinguishing patients with cirrhosis enrolled in HEDS who developed HCC after enrollment and those who remained without HCC. The THCC-RI uses age (as continuous), etiology (HCV cured or active vs other), history of smoking (current or past vs others), use of alcohol (current heavy use vs other), BMI (continuous), AFP (continuous), serum albumin (continuous), ALT (continuous), and platelet count (continuous) with the outcome being the time to HCC diagnosis (Supplementary Table 1).⁸ The parameters from the THCC-RI were implemented in the HEDS cohort. We assessed model performance through discrimination and calibration. We calculated time-dependent AUROC based on inverse probability of censoring weighting,¹⁵ with controls defined as patients free of any HCC event for up to 5 years of follow-up to evaluate predictive discrimination of the THCC-RI. AUROC or C-index values range from zero to one, with a value of 0.5 corresponding to the performance of a random classifier. A model with a C-index above 0.70 adequately discriminates between risk profiles.¹⁶ We evaluated calibration with visual assessments of calibration curves. We calculated the observed risk using the standard nonparametric (Nelson-Aalen) method. We examined time-specific cumulative risk of HCC in patients with cirrhosis, stratified by deciles of predicted risk scores. We compared cumulative incidence function curves and calculated the hazard ratios (HRs) across e risk groups: the highest 20th percentile (high risk), the lowest 20th percentile (low risk), and an intermediate group. We used 5000 bootstrap samples to estimate the 95% CIs.

We calculated Age, Male, Albumin-bilirubin, and Platelets (aMAP) and FIB-4 and assessed their performance through discrimination, calibration, and time-dependent AUROC as described above.

Results

HEDS included 1723 patients with cirrhosis. Of these, 15 patients had missing data on liver disease etiology; 55 had missing data on smoking, use of alcohol, or BMI; 51 patients had missing values for AFP; and 42 had missing values for serum albumin, ALT, or platelet count, resulting in 1560 patients with complete information to calculate THCC-RI.

The mean age of this cohort was 59 years (standard deviation [SD], 10 years), and 46.8% were female. The cohort included 85% White, 11% Hispanic, and 9% Black patients. HCV was the leading etiological risk factor of cirrhosis (40.4%; 15.1% cured HCV) followed by MASLD (22%) and alcohol-associated liver disease (16.4%) (Table 1). In total, 15% of patients were current tobacco smokers, and fewer than 2% reported current heavy alcohol use. The mean BMI was 31.2 kg/m² (SD, 7.2 kg/m²). Most patients (72.4%) had Child-Pugh Class A and 27.6% had Child Pugh B class at baseline. During a median follow up of 2.5 years (interquartile range [IQR], 1.1–5.2 years), 114 patients developed incident HCC at a cumulative HCC incidence of 1.7%, 3.9%, and 8.5% at 1, 2, and 5 years, respectively. A total of 59 patients received transplantation, and 261 died during follow-up.

Table 1.

Baseline Characteristics of the HEDS Cohort With Cirrhosis

Characteristic	Data
Age at baseline, y	58.9 (10.1)
Sex
Male	830 (53.2)
Female	730 (46.8)
Hispanic or Latino	171 (11.0)
Race
White	1308 (85.1)
Black	108 (7.0)
Asian	33 (2.1)
Other	88 (5.7)
Missing	23 (1.4)
BMI, kg/m2	31.2 (7.2)
Cigarette smoking
Never	673 (43.1)
Former	652 (41.8)
Current	235 (15.1)
Alcohol use
Never	511 (32.8)
Current heavy	29 (1.9)
Other	1020 (65.4)
Cirrhosis etiology
Hepatitis C	630 (40.4)
Hepatitis B	33 (2.1)
Alcohol-related	256 (16.4)
MASLD	343 (21.9)
Cryptogenic	65 (4.2)
Other	233 (14.9)
Child-Pugh class
Class A	1129 (72.4)
Class B	430 (27.6)

NOTE. Data are presented as number (%) or mean (standard deviation).

BMI, body mass index; HEDS, Hepatocellular Carcinoma Early Detection Strategy; MASLD, metabolic dysfunction-associated steatotic liver disease.

Model Performance

The time dependent AUROC values were 0.77 (95% CI, 0.65–0.85) at 1 year, 0.73 (95% CI, 0.64–0.80) at 2 years, 0.73 (95% CI, 0.66, 0.79) at 3 years, 0.72 (0.66, 0.78) at 4 years, and 0.71 (0.65, 0.77) at 5 years, respectively, suggesting good overall discrimination and remined stable over time (Table 2). The model has good performance across subgroups based on sex and etiology of cirrhosis (Table 2). Figure 1 displays the predicted vs observed risk of HCC at 2 and 3 years across deciles of predicted risk. The model was overall well-calibrated across the full range, with good agreement between observed and predicted risk except in the highest decile, where the model underestimated the risk slightly. However, there was good agreement within the expected HCC risk range.

Table 2.

Performance of THCC-RI: Time-dependent AUROC for Predicting HCC

Group (n for cirrhosis, n for HCC)	1 y	2 y	3 y	4 y	5 y
All (1560, 114)	0.77 (0.65–0.85)	0.73 (0.64–0.80)	0.73 (0.66–0.79)	0.72 (0.66–0.78)	0.71 (0.65–0.77)
Female (730, 32)		0.71 (0.49–0.88)	0.69 (0.52–0.82)	0.70 (0.57–0.82)	0.67 (0.54–0.79)
Male (830, 82)	0.72 (0.59–0.81)	0.69 (0.59–0.78)	0.71 (0.63–0.78)	0.71 (0.63–0.77)	0.71 (0.63–0.78)
Viral etiology (663, 62)	0.86 (0.77–0.94)	0.79 (0.64–0.87)	0.78 (0.66–0.85)	0.73 (0.63–0.81)	0.72 (0.62–0.80)
Non-viral etiology (897, 52)	0.76 (0.57–0.86)	0.70 (0.60–0.79)	0.72 (0.62–0.79)	0.73 (0.64–0.81)	0.72 (0.63–0.80)

NOTE. Data are presented as AUROC (95% CI).

NOTE. Data from 1560 patients enrolled in HEDS with information to calculate THCC-RI.

NOTE. AUROC not shown if there were <5 HCC events in the group.

AUROC, area under the receiver operating characteristic curve; CI, confidence interval; HCC, hepatocellular carcinoma; THCC-RI, Texas HCC Consortium Risk Index.

Figure 1.

Calibration plots for 2- (A) and 3-year (B) HCC risks.

We compared THCC-RI with aMAP among 1419 patients who had information on both scores (Table 3). THCC-RI and aMAP had similar discrimination overall, although the performance was more stable over time for THCC-RI than for aMAP. Notably, the AUROCs for aMAP dropped more than those for THCC-RI as the time horizon expanded beyond 3 years. Both models performed better in viral than non-viral etiologies, THCC-RI performed better than aMAP among women, and particularly beyond 2 years, where the AUROC for aMAP dropped as low as 0.55 to 0.60 compared with 0.67 to 0.70 for THCC-RI. The AUROCs for FIB-4 ranged from 0.66 to 0.64 compared with 0.77 to 0.71 for THCC-RI, demonstrating better discrimination offered by THCC-RI (Supplementary Table 2).

Table 3.

Comparative Performance of THCC-RI Time-dependent AUROC for Predicting HCC

Group (n for cirrhosis, n for HCC)	1 y	2 y	3 y	4 y	5 y
All (1419, 101)
THCC-RI	0.77 (0.62–0.85)	0.72 (0.63–0.80)	0.73 (0.65–0.79)	0.72 (0.64–0.78)	0.71 (0.63–0.77)
aMAP	0.80 (0.68–0.86)	0.74 (0.66–0.81)	0.72 (0.64–0.79)	0.70 (0.63–0.76)	0.68 (0.61–0.74)
Female (661, 26)
THCC-RI	–	0.67 (0.41–0.92)	0.67 (0.49–0.81)	0.69 (0.55–0.82)	0.66 (0.52–0.78)
aMAP	–	0.67 (0.33–0.89)	0.55 (0.32–0.72)	0.59 (0.43–0.72)	0.60 (0.44–0.72)
Male (758, 75)
THCC-RI	0.72 (0.57–0.82)	0.69 (0.59–0.78)	0.72 (0.63–0.79)	0.70 (0.62–0.78)	0.71 (0.63–0.78)
aMAP	0.70 (0.57–0.79)	0.68 (0.59–0.76)	0.71 (0.64–0.78)	0.69 (0.61–0.76)	0.68 (0.59–0.75)
Viral etiology (602, 53)
THCC-RI	0.86 (0.74–0.94)	0.78 (0.62–0.88)	0.77 (0.65–0.85)	0.72 (0.60–0.81)	0.71 (0.60–0.80)
aMAP	0.86 (0.73–0.94)	0.81 (0.69–0.89)	0.75 (0.64–0.84)	0.69 (0.58–0.78)	0.66 (0.56–0.76)
MASLD or ALD (817, 48)
THCC-RI	0.75 (0.51–0.86)	0.68 (0.57–0.78)	0.70 (0.60–0.79)	0.72 (0.62–0.80)	0.71 (0.61–0.79)
aMAP	0.75 (0.56–0.85)	0.68 (0.45–0.78)	0.69 (0.56–0.78)	0.71 (0.60–0.79)	0.69 (0.58–0.78)

NOTE. Data are presented as AUROC (95% CI).

NOTE. Data from 1419 patients enrolled in HEDS with information to calculate both THCC-RI and aMAP.

NOTE. AUROC not shown if there were <5 HCC events in the group.

aMAP, Age, Male, Albumin-bilirubin, and Platelets; AUROC, area under the receiver operating characteristic curve; CI, confidence interval; HCC, hepatocellular carcinoma; HEDS, Hepatocellular Carcinoma Early Detection Strategy; THCC-RI, Texas HCC Consortium Risk Index.

Table 4 shows each decile of predicted incidence and mean observed HCC incidence for a given decile in the cohort. There were 3 HCC cases in the lowest decile group. The cumulative incidence of HCC was notably high in the 2 highest deciles. Compared with the medium-risk group (deciles 3–8), the high-risk group (deciles 9 and 10) had a 3.5-fold (HR, 3.5; 95% CI, 2.4–5.1) higher risk of HCC (Figure 2).

Table 4.

Distribution of Observed HCC Risk Based on the THCC-RI in the HEDS Cohort

Decile	Cumulative risk, %
	1 y	2 y	3 y
1	0.7	0.7	0.7
2	0.0	1.6	2.6
3	0.0	1.6	2.5
4	0.7	2.5	3.5
5	0.0	0.9	2.0
6	1.5	4.8	6.8
7	2.9	4.6	6.4
8	1.6	2.6	3.9
9	4.6	6.4	9.4
10	5.3	13.5	18.9

AUROC, area under the receiver operating characteristic curve; HCC, hepatocellular carcinoma; HEDS, Hepatocellular Carcinoma Early Detection Strategy; THCC-RI, Texas HCC Consortium Risk Index.

Figure 2.

Cumulative incidence of HCC according to the THCC-RI in the HEDS cohort.

Discussion

We developed the THCC-RI for risk stratification for HCC in patients with cirrhosis using objective, reliable, and easily measurable demographic and clinical factors. Risk prediction models need to demonstrate reproducibility of discrimination in several external cohorts, but most models have not undergone extensive external validation to support their broad applicability.¹⁷ In this prospective, multi-ethnic, multi-etiology cirrhosis cohort study, measures of THCC-RI performance, including discrimination and calibration, were consistent with the performance in the original development cohort. THCC-RI can provide accurate prognostic estimates for a heterogenous population of ambulatory patients with cirrhosis that may be used to inform decision-making about HCC surveillance and prevention.

Both this validation as well as the original development studies represent a phase 3 biomarker testing study.¹⁸ The prospective uniform specimen collection for patients from several¹⁹ centers mitigates systematic biases by ensuring that information for patients who progressed to HCC and control subjects was collected in the same way.²⁰ Outcomes were ascertained for all subjects in the cohort. The validation cohort included patients with cirrhosis, comprised of different compositions of the sexes, races, and etiological and behavioral risk factors. Specifically, THCCC had a higher proportion of Hispanic patients, and included more patients with MASLD and ALD than HEDS. In contrast, HEDS had a higher proportion of patients with HCV (viremic and post-SVR) than THCCC. Despite these differences, the THCC-RI had similar discrimination and calibration across the 2 cohorts. THCC-RI previously demonstrated good performance in a cohort of United States patients seeking care in the Veterans Health Administration (mean age, 64 years; 61.0% Whites, 25.0% Blacks, and 6.3% Hispanics; etiology, 46% cured HCV, 32% alcohol, 15% MASLD) and in a separate cohort of Asian patients (mean age, 53 years; 100% Asian; etiology: 54.4% alcohol and 30.5% treated HBV,).¹⁹ Together, these data strengthen our evaluation because prediction models that perform well in external cohorts that differ from the derivation cohort provide greater reproducibility (replicating performance in populations that more closely resemble the derivation cohort) and transportability (replicating performance in different populations). Recruitment from different types of health care settings and different regions of the world in these external validation studies demonstrate that THCC-RI is reproducible and likely transportable to other populations with cirrhosis.

THCC-RI offered stable predictive ability over the 5-year time horizon, hence addressing an important clinical need for long-term prediction. In a comparative analysis, aMAP had similar discrimination to THCC-RI for 1- and 2-year time points, suggesting either tool may be suitable for short-term prediction. However, the predictive ability of aMAP declined beyond 3 years, indicating greater long-term stability of THCC-RI. Similar to this study, Yoo et al recently reported significant predictive power of THCC-RI over an extended period.¹⁹

For a risk model to be clinically relevant, clinicians should be able to use the predicted risk to guide decisions about surveillance or risk reduction. THCC-RI is suitable for implementation in routine clinical practice to identify patients at high risk for HCC who could be candidates for resource intensive HCC surveillance. THCC-RI generates predicted HCC risk, with specific cutoffs or thresholds that separate patients into different risk strata (Table 4). It also holds the promise to further improve risk stratification by identifying low-risk patients, who may be candidates for de-escalation of HCC surveillance. Ideally, a randomized controlled trial should examine the clinical utility of HCC risk stratification in patients with cirrhosis; however, such trials may not be feasible given the large cohort size and lengthy follow-up needed to show a benefit. A decision-analytical approach could help bridge this evidence gap by simulating the clinical course of cirrhosis and quantifying how different risk-stratification models may reduce HCC mortality through improved surveillance and early detection. Future studies will use mathematical modeling or pragmatic trial designs to examine long-term outcomes, harms, and cost-effectiveness of various THCC-RI-stratified screening strategies.

The study had few limitations. Despite the large cohort of 1560 patients with cirrhosis, only 114 patients developed HCC. Therefore, some of the HCC risk estimates have wider 95% CIs than those in the development cohort. We used empirically derived cutoffs (Table 4) to illustrate the full spectrum of HCC risk. However, further research is needed to determine the optimal risk threshold for decision-making, one that accounts for the risks and benefits of HCC surveillance. The discrimination value of the model was “good”; it is possible that adding variables that reflect different aspects of disease progression (eg, genomics, radiomics, proteomics) and incorporating temporal changes in these biomarkers may increase the C-statistic value to >0.8 “strong.” We are currently pursuing this approach.

Conclusion

In summary, this prospective cohort study provided external validation for THCC-RI and further demonstrated its good discrimination and calibration as an HCC risk-stratification tool. Future studies will determine the optimal risk threshold for decision-making while accounting for the risks and benefits of HCC surveillance.

Supplementary Material

Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at https://doi.org/10.1016/j.cgh.2026.01.041.

What You Need to Know.

Background

Risk stratification for hepatocellular carcinoma (HCC) can promote shared decision-making between patients and providers by providing a personalized HCC risk assessment. The Texas HCC Consortium Risk Index (THCC-RI) successfully stratified patients with cirrhosis based on their future risk of HCC, but further validation is needed to establish generalizability and clinical utility.

Findings

In this analysis of 1560 patients with cirrhosis enrolled in a large prospective study, THCC-RI had good performance for predicting the future risk of HCC in patients with cirrhosis, with stable discrimination and calibration over a 5-year follow-up.

Implications for patient care

Future studies will determine the optimal THCC-RI risk threshold for decision-making while accounting for the risks and benefits of HCC surveillance.

Abbreviations used in this paper:

AASLD

American Association for the Study of Liver Diseases

AFP

alpha fetoprotein

ALD

alcohol-related liver disease

ALT

alanine aminotransferase

aMAP

Age, Male, Albumin-bilirubin, and Platelets

AUROC

area under the receiver operating curve

BMI

body mass index

CI

confidence interval

EMR

electronic medical record

FIB-4

Fibrosis-4 index

HCC

hepatocellular carcinoma

HCV

hepatitis C virus

HEDS

Hepatocellular Carcinoma Early Detection Strategy

HR

hazard ratio

IQR

interquartile range

LI-RADS

Liver Imaging Reporting and Data System

LT

liver transplantation

MASLD

metabolic dysfunction-associated steatotic liver disease

MELD

Model for End-stage Liver Disease

SD

standard deviation

SVR

sustained virologic response

TCCC

Texas HCC Consortium Cohort

THCC-RI

Texas HCC Consortium Risk Index