Residual risk of hepatocellular carcinoma development for chronic hepatitis C patients treated by all oral direct-acting antivirals with sustained virological response

Abstract

Background:

The treatment of chronic hepatitis C (CHC) infection underwent a significant transformation with the introduction of all-oral direct-acting anti-virals (DAAs). These medications offered a high success rate in treatment, shorter duration, good tolerability, and expanded treatment options. However, a residual risk of hepatocellular carcinoma (HCC) development remained for a few patients even after achieving sustained virological response (SVR). To date, there is a lack of real-world data on evaluating risk factors associated with de novo HCC in CHC patients post-SVR, particularly in Taiwan.

Methods:

Between January 2017 and December 2019, a total of 671 consecutive CHC patients who achieved SVR after receiving DAAs were included for analysis. Patients with a history of HCC or liver transplantation prior to DAAs, a short follow-up period (<1 year), or treatment failure with DAAs were excluded. The primary outcome was the development of HCC following the initiation of DAAs. Variables associated with the primary outcome were assessed using multivariate Cox proportional hazards models.

Results:

The mean age of the enrolled patients was 65.1 ± 12.8 years, with 39.6% of them being male. Among the patients, 30.6% had advanced (F3-4) fibrosis, and the median follow-up period was 2.90 years. The cumulative incidence of HCC in CHC patients post-SVR12 was 1.6% at 1 year, 4.4% at 2 years, 4.8% at 3 years, 5.3% at 4 years, and 6.1% at 4.8 years, respectively. Variables independently associated with de novo HCC were advanced liver fibrosis (hazard ratio [HR] = 6.745; 95% CI = 1.960-23.218; p = 0.002), end-of-treatment 12 weeks (EOT₁₂) alpha-fetoprotein (AFP) >7 ng/mL (HR = 3.059; 95% CI = 1.215-7.669; p = 0.018), EOT₁₂ albumin-bilirubin (ALBI) grade ≥ 2 (HR = 2.664; 95% CI = 1.158-6.128; p = 0.021), and body mass index (BMI) ≥ 25 kg/m² (HR = 2.214; 95% CI = 1.011-4.852; p = 0.047).

Conclusion:

Despite achieving viral clearance with DAAs, CHC patients still face a residual risk of de novo HCC. Establishing a risk stratification model based on independent variables could facilitate the prediction of future HCC development and enhance screening strategies.

1. INTRODUCTION

Chronic hepatitis C virus (HCV) infection causes persistent hepatic inflammation and results in progressive liver disease. Studies have estimated that approximately 10%-20% of patients develop liver cirrhosis over a period of 20-30 years following HCV infection, with an annual risk of hepatic decompensation ranging from 3% to 6% and an annual risk of hepatocellular carcinoma (HCC) ranging from 1% to 5%.¹ In addition to inducing hepatic injury and fibrosis, ultimately leading to liver cirrhosis, HCV possesses direct carcinogenic potential, exerting pro-oncogenic effects on infected cells through oxidative stress, dysregulation of host cell checkpoints, and DNA damage.²

The therapeutic landscape for chronic hepatitis C (CHC) has undergone a remarkable transformation with the advent of interferon (IFN)-free oral direct-acting anti-viral agents (DAAs). These all-oral DAAs, which target viral proteins such as NS3/4A protease, NS5B polymerase, and the NS5A replication complex, have achieved sustained virological response (SVR) rates of over 95% in patients. They offer shorter treatment duration, a good safety profile, and excellent tolerability.^3–6 Successful anti-viral therapy has had substantial documented benefits, including improved liver function and a significant reduction in overall and liver-related mortality.^7–9

Kanwal et al¹⁰ conducted a large cohort study involving 22 500 DAA-treated patients, with a mean follow-up period of 1.02 years, to investigate the risk of HCC. The study demonstrated a significantly reduced risk of HCC in patients achieving SVR compared with non-SVR patients (0.90 vs 3.45 HCC/100 person-years; adjusted hazard ratio [HR] = 0.28, 95% CI = 0.22-0.36). Another multicenter prospective cohort study by Carrat et al,⁷ with a mean follow-up period of 33.4 months, reported that DAAs therapy was associated with reduced mortality and risk of HCC. The study included 7344 patients who received DAAs treatment and 2551 patients who did not. After adjusting for variables, DAAs treatment was linked to a decrease in HCC (adjusted HR = 0.66, 95% CI = 0.46-0.91) and all-cause mortality (adjusted HR = 0.48, 95% CI = 0.33-0.70).⁷ However, despite achieving viral clearance through DAAs, previous studies have identified a residual risk of HCC development and progression of liver disease in a small proportion of patients.^11,12 Risk factors for de novo HCC post-SVR include advanced liver fibrosis, hypoalbuminemia, thrombocytopenia, elevated alpha-fetoprotein (AFP) levels, and co-existing fatty liver disease.^13,14

In Taiwan, the National Health Insurance Administration of the Ministry of Health and Welfare (NHIA) has provided conditional reimbursement for all-oral DAAs since January 24, 2017. However, there is a scarcity of real-world data in Taiwan that assesses the risk factors associated with de novo HCC in CHC patients following DAAs treatment with SVR. Consequently, this study was conducted to address this significant issue. Furthermore, the study successfully established a risk stratification model based on identifying independent predictors.

2. METHODS

2.1. Study population

This was a retrospective, single-center study. Between January 2017 and December 2019, 1097 consecutive adults (≥20 years) CHC patients who received all-oral DAAs at our hospital were considered potential candidates. CHC was defined as detectable HCV antibody (anti-HCV) (Abbott HCV EIA 2.0, Abbott Laboratories, Abbott Park, IL) and quantifiable serum HCV RNA (Cobas TaqMan HCV Test version 2.0; Roche Diagnostics GmbH, Mannheim, Germany; lower limit of quantification: 15 IU/mL) for ≥6 months. The exclusion criteria included a history of any of the following: (1) HCC (n = 200); (2) liver transplantation (n = 3); (3) a short period of follow-up time (≤1 year) (n = 204); and (4) DAAs treatment failure (n = 19). Ultimately, a total of 671 patients were enrolled for statistical analyses. This study was approved by the Institutional Review Board of Taipei Veterans General Hospital and was conducted in accordance with the principles of the Declaration of Helsinki and the International Conference on Harmonization for Good Clinical Practice. Written informed consent was obtained from all the patients before recruitment. The flow chart of the case collection is depicted in Fig. 1.

Fig. 1.

Flow chart of case collection. CHC = chronic hepatitis C; DAAs = direct-acting antivirals; HCC = hepatocellular carcinoma; SVR = sustained virological response.

2.2. DDAs treatment protocol

In the current study, the DAA regimen was selected by clinical physicians based on factors such as the HCV genotype/subtype, viral load, patient characteristics, drug-drug interaction (DDIs) profiles, and evolving NHIA reimbursement criteria.^15,16 The HCV genotype was determined using a commercially available assay (Cobas HCV GT; Roche Diagnostics GmbH). In summary, the patients received the following treatment regimens: paritaprevir/ritonavir, ombitasvir, and dasabuvir (PrOD) for 12 weeks; daclatasvir (DCV) and asunaprevir (ASV) for 24 weeks; sofosbuvir (SOF) in combination with ribavirin (RBV), daclatasvir (DCV), ledipasvir (LDV), or velpatasvir (VEL) for 12 weeks; elbasvir (EBR) and grazoprevir (GZR) for 12 weeks; and glecaprevir (GLE) plus pibrentasvir (PIB) for 8 or 12 weeks. Before initiating DAA therapy, a comprehensive survey of the regular medications taken by the enrolled patients was conducted to evaluate potential DDIs. Medications that could potentially interact were either discontinued, switched to alternative drugs, or initiated at the lowest dose with frequent monitoring by physicians.^17,18 After assessment by clinical physicians, ribavirin (RBV) (200 mg capsule of Robatrol; Genovate Biotechnology Co., Ltd., Taiwan, China) was added to the regimen of selected patients to enhance their virological response. Oral RBV (200-1200 mg total daily dose) was administered based on the patient’s body weight and estimated glomerular filtration rate. The need for dose adjustments, temporary interruptions, or discontinuation of RBV was determined based on the manufacturer’s prescribing information.

During the treatment period, patients underwent assessments by physicians at weeks 1 and 2, followed by subsequent evaluations every 2 weeks (or more frequently in cases of adverse events) until the completion of therapy. Serum quantitative HCV RNA levels were measured prior to initiating DAA treatment, at week 4 (optional), at the end of treatment, and at post-treatment week 12 to determine the virological response.^19,20 SVR₁₂ was defined as an undetectable HCV RNA level on real-time polymerase chain reaction (PCR) (≤15 IU/mL) after DAA treatment and 12 weeks after completing therapy. Patients without SVR₁₂ data were considered to have not achieved SVR₁₂.^19,20

2.3. Surveillance of HCC before and after DAAs

In the current study, all patients underwent periodic serum AFP tests and imaging follow-ups to detect the occurrence of HCC during and after anti-viral treatment. Abdominal ultrasound (US) was performed every 3-6 months, while dynamic studies such as computed tomography (CT) or magnetic resonance imaging (MRI) were conducted every 4-6 months. In cases where an uncharacterized nodule was identified during imaging, or there was an unexplained elevation in AFP levels, further investigation through contrast-enhanced studies and/or biopsy was arranged to avoid any selection bias, particularly before initiating DAAs. HCC was diagnosed based on histology or typical imaging findings observed during contrast-enhanced CT or MRI, which exhibited hyper-enhancement in the arterial phase and a washout pattern in the portal or delayed phases.

2.4. Primary end-point

The primary end-point analyzed in the current study was the cumulative incidence of de novo HCC in CHC patients who were successfully treated with all-oral DAAs. The date of HCC diagnosis, whether determined by histology or typical dynamic CT or MRI findings, was used to construct the cumulative incidence curves.

2.5. Definitions

The fibrosis-4 (FIB-4) score is a noninvasive scoring system that utilizes laboratory tests (platelets, alanine aminotransferase [ALT], aspartate aminotransferase [AST]) and age to estimate the degree of hepatic fibrosis. The FIB-4 index is calculated using the following formula: FIB-4 = Age (years) × AST (U/L)/(platelet [10⁹/L] × √ALT [U/L]). A FIB-4 index greater than 3.25 has a positive predictive value of 82.1% and a specificity of 98.2% for confirming the presence of advanced fibrosis (F3-4).²¹ In the current study, advanced fibrosis (F3-4) was defined by either the presence of typical clinical or radiological manifestations of cirrhosis, with evidence of portal hypertension such as esophageal or gastric varices, or a FIB-4 score higher than 3.25.

The albumin-bilirubin (ALBI) score/grade is a liver function measure that was initially developed as a prognostic factor for patients with HCC and has now become well-established. The formula for this scoring system is based on the following equation: ALBI score = (log₁₀ bilirubin [µmol/L] × 0.66) + (albumin [g/L] × –0.0852). Consequently, three ALBI grades (1, 2, and 3) were defined as follows: ALBI score ≤ –2.60 (ALBI grade 1), > –2.60 to ≤ –1.39 (ALBI grade 2), and > –1.39 (ALBI grade 3). The ALBI score is an objective measure that can detect even slight changes in liver dysfunction, surpassing the sensitivity of the Child-Pugh or Model for End-stage Liver Disease (MELD) scores. The ALBI score/grade has been effectively utilized to predict survival outcomes in patients with nonmalignant liver diseases of diverse etiologies.²²

Fatty liver was diagnosed by evaluating the following ultrasound parameters, routinely assessed before initiating DAAs treatment: parenchymal brightness, liver-to-kidney contrast, deep beam attenuation, bright vessel walls, and gallbladder wall definition. Patients were categorized as overweight if their body mass index (BMI) exceeded 25 kg/m², as defined by the World Health Organization.

2.6. Statistical analysis

All statistical analyses were performed using SPSS Statistics version 23.0 (SPSS Inc., Chicago, IL). Baseline patient characteristics were presented as means with SDs or as percentages, as appropriate. Quantitative HCV RNA levels (IU/mL) were logarithmically transformed for analysis. Categorical variables were analyzed using the chi-square test or Fisher exact test, while continuous variables were analyzed using the Mann-Whitney U test. Survival and cumulative incidence curves were compared using the log-rank test for Kaplan-Meier analysis. The risk of developing HCC at any point during follow-up, considering baseline and post-SVR₁₂ variables, was estimated using multivariable Cox regression analysis with HRs and 95% CIs. In the multivariate analysis, if a surrogate marker such as advanced fibrosis (F3-4) or ALBI grade was included as an independent variable for residual HCC, its main components, such as platelet count, AST, or albumin, were not incorporated to avoid dilutional effects. The length of follow-up was measured from the date of DAA initiation until death, the last visit, or the last abdominal image follow-up. All statistical tests were two-sided, and a significance level of p < 0.05 was used.

3. RESULTS

3.1. Clinical characteristics of the enrolled patients

A total of 671 patients were included in the analysis. The mean age of the enrolled patients was 65.1 ± 12.8 years. Of the total, 266 (39.6%) patients were male, and 143 (21.3%) patients had previously failed pegylated IFN therapy. The median baseline HCV RNA level was 6.07 log₁₀ IU/mL, with an interquartile range of 5.12-6.73 log₁₀ IU/mL. The distribution of HCV genotypes was as follows: 7.2% had genotype 1a, 60.8% had genotype 1b, 28.1% had genotype 2, 0.3% had genotype 3, 3.6% had genotype 6, and 0.7% had unclassified genotype.

According to the predefined criteria, 205 (30.6%) patients in the current study were classified as having advanced (F3-4) fibrosis, and 2.2% of the patients were diagnosed with decompensated (CTP class B-C) cirrhosis. Among the enrolled patients, 53.7% received treatment with SOF-based regimens. Detailed information on the DAA regimens used is provided in Table 1. The median follow-up period after DAA initiation was 2.90 years, with an interquartile range of 2.03-3.70 years.

Table 1.

Clinical characteristics of the study population

Characteristics	Patients (n = 671)
Age, y, means ± SD	65.1 ± 12.8
Male sex, n (%)	266 (39.6)
Body mass index, means ± SD	24.98 ± 4.11
HCV RNA, n (%)
≥1 000 000 (IU/mL)	355 (52.9)
<1 000 000 (IU/mL)	316 (47.1)
HCV genotype, n (%)
1a	48 (7.2)
1b	405 (60.8)
2	187 (28.1)
3	2 (0.3)
6	24 (3.6)
Unclassified	5 (0.7)
Previous IFN failure, n (%)	143 (21.3)
Diabetes mellitus, n (%)	118 (17.6)
Coinfected with HBV, n (%)	37 (5.5)
Coinfected with HIV, n (%)	30 (4.5)
Advanced fibrosis (F3-4), n (%)	205 (30.6)
Decompensated (CTP class B-C) cirrhosis, n (%)	15 (2.2)
DAAs regimen, n (%)
PrOD	136 (20.3)
DCV + ASV	69 (10.3)
SOF + RBV	65 (9.7)
SOF + DCV	29 (4.3)
SOF + LDV	260 (38.7)
SOF + VEL	6 (0.9)
EBR + GZR	18 (2.7)
GLE + PIB	88 (13.1)
Median observation period after DAAs, y	2.90

ASV = asunaprevir; CTP = Child-Turcotte-Pugh; DAAs = direct-acting antivirals; DCV = daclatasvir; EBR = elbasvir; GLE = glecaprevir; GZR = grazoprevir; HBV = hepatitis B virus; HCV = hepatitis C virus; HIV = human immunodeficiency virus; IFN = interferon; LDV = ledipasvir; PIB = pibrentasvir; PrOD = paritaprevir/ritonavir, ombitasvir, and dasabuvir; RBV = ribavirin; SOF = sofosbuvir; VEL = velpatasvir.

3.2. Laboratory data at 12 weeks post-DAAs of enrolled patients

After the eradication of HCV, the median ALT and AST levels at 12 weeks post-DAAs (EOT₁₂) were 18 IU/L and 22 IU/L, respectively. At the same time point, the median albumin level was 4.3 g/dL, the total bilirubin level was 0.70 mg/dL, and the platelet count was 175 × 1000/μL. Utilizing the laboratory data obtained at 12 weeks post-DAAs, the distribution of ALBI grade was as follows: grade 1: 83.6%, grade 2: 15.5%, and grade 3: 0.9% (Table 2). Since the laboratory values affected by hepatitis can be altered after DAAs treatment, we used the EOT₁₂ laboratory data to examine the development of de novo HCC in subsequent statistical analyses.

Table 2.

Laboratory data at 12 weeks post-DAAs of enrolled patients

Characteristics	Patients (n = 671)
Albumin (g/dL)	4.3 (4.1-4.4)
Aspartate aminotransferase (U/L)	22 (18-28)
Alanine aminotransferase (U/L)	18 (13-25)
Total bilirubin (mg/dL)	0.70 (0.53-1.00)
Creatinine (mg/dL)	0.81 (0.71-0.98)
Alpha-fetoprotein (ng/mL)	3.02 (2.05-4.51)
White cell count (×1000/μL)	5.5 (4.5-6.7)
Hemoglobin (g/dL)	13.6 (12.5-14.7)
Platelet count (×1000/μL)	175 (130-223)
Prothrombin time (s)	11.2 (10.7-11.9)
FIB-4 score	2.00 (1.37-3.23)
ALBI grade 1/2/3, n (%)	561/104/6 (83.6%/15.5%/0.9%)

Continuous variables are expressed as the median with 25th and 75th percentiles.

ALBI = albumin-bilirubin; DAAs = direct-acting antivirals; FIB-4 = fibrosis-4.

3.3. Cumulative incidence of de novo HCC development

Following the initiation of all-oral DAAs, a total of 32 patients were diagnosed with de novo HCC. In our current study, the cumulative incidence of developing HCC after DAA initiation for CHC patients post SVR12 was 1.6% at 1 year, 4.4% at 2 years, 4.8% at 3 years, 5.3% at 4 years, and 6.1% up to 4.8 years (Fig. 2). The median time to HCC development from DAA initiation was 1.18 years, with an interquartile range of 0.78-1.68 years.

Fig. 2.

The overall cumulative incidence to develop HCC was at 1.6% (95% CI = 1.1%-2.1%) at 1 y, 4.4% (95% CI = 3.6%-5.2%) at 2 y, 4.8% (95% CI = 3.9%-5.7%) at 3 y, 5.3% (95% CI = 4.3%-6.3%) at 4 y, and 6.1% (95% CI = 4.8%-7.4%) till 4.8 y, respectively. DAAs = direct-acting antivirals; HCC = hepatocellular carcinoma.

Among the patients who developed HCC during the follow-up period, 18 patients (56.3%) were classified as Barcelona-Clinic Liver Cancer (BCLC) stage 0, while 11 patients (34.4%) were classified as BCLC stage A. Only three patients (9.4%) had more advanced stage HCC (BCLC B or C). The median tumor size was 1.65 cm (IQR: 1.40-2.95 cm). Of the patients, 26 (81.3%) developed a single nodule, while 6 (18.7%) developed two or more nodules. Only two patients showed evidence of vascular invasion based on imaging studies. There was no identified correlation between the specific type of DAAs used and the characteristics of de novo HCC.

3.4. Comparison of the characteristics between the patients who developed HCC and non-HCC patients

The comparison between patients who developed HCC and those who did not reveal several distinct characteristics. The patients who developed HCC were found to be older and had a higher incidence of previous interferon treatment failure, diabetes mellitus, BMI > 25 kg/m², and advanced (F3-4) fibrosis compared with the non-HCC patients (Table 3). Regarding the blood test results of the HCC patients, the EOT₁₂ levels of aspartate aminotransferase (AST), alanine aminotransferase (ALT), total bilirubin, prothrombin time, and alpha-fetoprotein (AFP) were significantly higher, while the albumin, hemoglobin (Hb) level, and platelet count were significantly lower (Table 3). Furthermore, patients who developed HCC had a significantly higher likelihood of belonging to EOT₁₂ ALBI grade 2 + 3 compared with patients without HCC (59.4% vs 14.2%; p < 0.001) (Table 3).

Table 3.

Comparison of the characteristics between the patients who developed HCC and non-HCC patients

Characteristics	No HCC (n = 639)	HCC (n = 32)	p
Mean age (y)	64.8 ± 12.8	72.0 ± 10.0	0.002
Male sex (%)	253 (39.6)	13 (40.6)	0.907
BMI ≥25 kg/m2 (%)	249 (44.9)	21 (67.7)	0.013
HCV RNA ≥1 000 000 (IU/mL) (%)	337 (52.7)	18 (56.3)	0.698
HCV genotype 1 (%)	431 (67.9)	22 (71)	0.718
Previous IFN failure (%)	131 (20.5)	12 (37.5)	0.022
Diabetes mellitus (%)	106 (16.6)	12 (37.5)	0.002
SOF-based DAAs (%)	343 (53.7)	17 (53.1)	0.951
Advanced fibrosis (F3-4) (%)	178 (27.9)	27 (84.4)	<0.001
Decompensated (CTP class B-C) cirrhosis (%)	13 (2.0)	2 (6.3)	0.115
Mean baseline AFP (ng/mL)	8.75 ± 15.91	29.77 ± 75.64	<0.001
Mean EOT12 albumin (g/dL)	4.31 ± 1.81	3.86 ± 0.47	<0.001
Mean EOT12 AST (U/L)	24.7 ± 12.3	33.9 ± 16.9	<0.001
Mean EOT12 ALT (U/L)	21.6 ± 18.9	27.1 ± 15.0	0.001
Mean EOT12 total bilirubin (mg/dL)	0.82 ± 0.50	1.71 ± 2.99	0.001
Mean EOT12 AFP (ng/mL)	3.72 ± 2.97	30.83 ± 93.18	<0.001
Mean EOT12 WBC (×1000/μL)	5.8 ± 2.0	5.0 ± 1.9	0.06
Mean EOT12 Hgb (g/dL)	13.6 ± 1.8	12.7 ± 2.0	0.004
Mean EOT12 platelet count (×1000/μL)	180 ± 71	118 ± 58	<0.001
Mean EOT12 prothrombin time (s)	11.4 ± 1.3	12.2 ± 1.4	0.001
Mean EOT12 FIB-4 score	2.60 ± 2.08	5.10 ± 3.55	<0.001
EOT12 decompensated cirrhosis (%)	9 (1.4)	1 (3.1)	0.389
EOT12 ALBI grade 2 + 3 (%)	91 (14.2)	19 (59.4)	<0.001

AFP = alpha-fetoprotein; ALBI = albumin-bilirubin; ALT = alanine aminotransferase; AST = aspartate aminotransferase; BMI = body mass index; CTP = Child-Turcotte-Pugh; DAAs = direct-acting antivirals; EOT₁₂ = end-of-treatment 12 wk; FIB-4 = fibrosis-4; HCC = hepatocellular carcinoma; HCV = hepatitis C virus; Hgb = hemoglobin; IFN = interferon; SOF = sofosbuvir; WBC = white blood cell.

3.5. Factors associated with the development of HCC in DAA-treated patients who achieved SVR₁₂

According to the characteristic differences between patients with or without the development of HCC, we categorized continuous variables for Cox regression analysis. In the univariate analysis, the variables associated with de novo HCC development were as follows: age ≥ 65 years, diabetes mellitus, BMI ≥ 25 kg/m², advanced (F3-4) liver fibrosis, high ALBI grade (2 + 3), EOT₁₂ platelet count less than 100 × 10⁹/L, prothrombin time > 12 seconds, albumin level ≤ 3.5 g/dL, AST ≥ 40 U/L, total bilirubin ≥ 1 mg/dL, baseline AFP > 7 ng/mL, and EOT¹² AFP > 7 ng/mL (Table 4).

Table 4.

Cox regression analysis for factors associated with HCC development

Variables	Univariate		Multivariate
	HR (95% CI)	p	HR (95% CI	p
Age, y	2.616 (1.175-5.823)	0.019	1.974 (0.794-4.907)	0.143
≥65 vs <65
Sex	0.928 (0.458-1.880)	0.836	0.624 (0.279-1.398)	0.252
Female vs male
RNA (IU/mL)
≥1 000 000 vs <1 000 000	1.149 (0.571-2.312)	0.697
Genotype
Type non-1 vs type 1a + 1b	0.960 (0.439-2.096)	0.917
Previous IFN regimens
Yes vs no	1.961 (0.950-4.051)	0.069
Diabetes mellitus
Yes vs no	2.844 (1.390-5.818)	0.004	1.286 (0.611-2.711)	0.508
BMI (kg/m2)
≥25 vs <25	2.473 (1.164-5.252)	0.018	2.214 (1.011-4.852)	0.047
Coinfected with HBV
Yes vs no	0.561 (0.077-4.110)	0.569
Coinfected with HIV
Yes vs no	0.803 (0.109-5.885)	0.829
Advanced fibrosis (F3-4)
Yes vs no	13.246 (5.100-34.401)	<0.001	6.745 (1.960-23.218)	0.002
EOT12 ALBI grade
2 + 3 vs 1	8.200 (4.049-16.606)	<0.001	2.664 (1.158-6.128)	0.021
EOT12 Hb (g/dL)
<11 vs ≥11	1.826 (0.640-5.206)	0.260
EOT12 PLT (×109/L)
<100 vs ≥100	4.450 (2.197-9.014)	<0.001
EOT12 PT (s)
>12 vs ≤12	3.427 (1.711-6.868)	0.001	1.055 (0.478-2.329)	0.894
EOT12 albumin (g/dL)
≤3.5 vs >3.5	9.905 (4.569-21.474)	<0.001
EOT12 AST (U/L)
≥40 vs <40	3.523 (1.582-7.846)	0.002
EOT12 ALT (U/L)
≥40 vs <40	1.463 (0.446-4.805)	0.530
EOT12 AFP (ng/mL)
>7 vs ≤7	5.165 (2.489-10.718)	<0.001	3.059 (1.215-7.669)	0.018
Baseline AFP (ng/mL)
>7 vs ≤7	3.019 (1.508-6.046)	0.002	0.844 (0.352-2.025)	0.704

AFP = alpha-fetoprotein; ALBI = albumin-bilirubin; ALT = alanine aminotransferase; AST = aspartate aminotransferase; BMI = body mass index; EOT₁₂ = end-of-treatment 12 wk; Hb = hemoglobin; HBV = hepatitis B virus; HCC = hepatocellular carcinoma; HIV = human immunodeficiency virus; HR = hazard ratio; IFN = interferon; PLT = platelet; PT = prothrombin time.

We then performed a multivariable Cox model to identify the variables independently associated with the composite end-point. The variables found to be independently associated with the composite end-point were advanced liver fibrosis (HR = 6.745; 95% CI = 1.960-23.218; p = 0.002), EOT12 AFP > 7 ng/mL (HR = 3.059; 95% CI = 1.215-7.669; p = 0.018), EOT¹² ALBI grade ≥ 2 (HR = 2.664; 95% CI = 1.158-6.128; p = 0.021), and BMI ≥ 25 kg/m² (HR = 2.214; 95% CI = 1.011-4.852; p = 0.047) (Table 4).

3.6. Comparison of the cumulative rates of HCC development by independent factors

Fig. 3A illustrates the impact of advanced (F3-4) fibrosis on future HCC risk. The cumulative incidence of HCC development was significantly higher in individuals with advanced fibrosis for up to 4.8 years compared with patients without advanced fibrosis (17.2% vs 1.2%; p < 0.001).

Fig. 3.

Comparison of the cumulative rates of HCC development by independent factors. A, The cumulative incidence of HCC development according to liver fibrosis stage. F3 to F4: 17.2% (95% CI = 13.4%-21.0%) vs F0 to F2: 1.2% (95% CI = 0.6%-1.8%) till 4.8 y (p < 0.001). B, The cumulative incidence of HCC development according to ALBI grade: ALBI 2-3: 24.2% (95% CI = 17.8%-30.6%) vs ALBI 1: 2.6% (95% CI = 1.9%-3.3%) till 4.8 y (p < 0.001). ALBI = albumin-bilirubin; DAAs = direct-acting antivirals; HCC = hepatocellular carcinoma.

In CHC patients who achieved viral eradication through DAAs but belonged to EOT₁₂ ALBI grade 2 or 3, the cumulative incidence of developing HCC was significantly higher compared with patients who belonged to EOT12 ALBI grade 1, for up to 4.8 years (24.2% vs 2.6%; p < 0.001) (Fig. 3B).

3.7. Dynamic changes of AFP value before and after DAAs therapy and its predictive value of the occurrence of de novo HCC

Using a cutoff value of 7 ng/mL, we analyzed the dynamic changes in AFP values before and after DAAs treatment and categorized our study subjects into four groups. Interestingly, patients whose AFP values increased from normal (≤7 ng/mL) to abnormal (> 7 ng/mL) after DAAs treatment were found to have the highest risk of future HCC development. The cumulative incidence rates of HCC in this group were 20% at 1 year, 60% at 2 years, and 60% at 3 years.

Compared with the other groups, patients who maintained a persistently normal AFP value (≤7 ng/mL) throughout the course of DAAs treatment had the lowest incidence of developing HCC. The cumulative incidence rates of HCC in this group were 0.8% at 1 year, 2.2% at 2 years, 2.5% at 3 years, and 2.5% at 4 years (Fig. 4).

Fig. 4.

The dynamic changes of AFP value before and after DAAs therapy predicted the occurrence of HCC. Patients with AFP value increased from normal (≤7 ng/mL) to abnormal (>7 ng/mL) after DAAs treatment were associated with the highest risk of future HCC development. AFP = alpha-fetoprotein; DAAs = direct-acting antivirals; HCC = hepatocellular carcinoma; n.s. = nonsignificant.

3.8. Risk of HCC occurrence stratified by combining advanced fibrosis, AFP, ALBI grade, and BMI

Based on the results of the multivariate analysis, two risk stratification models were proposed. The first model (model A) was based on the presence of independent factors, including advanced (F3-4) fibrosis, EOT₁₂ AFP > 7 ng/mL, EOT₁₂ ALBI grade ≥ 2, and BMI ≥ 25 kg/m². Among the 671 study patients, 508 (75.7%) had 0 or 1 independent factor, placing them in the low-risk group. A total of 110 patients (16.4%) had two independent factors, constituting the medium-risk group, while 53 patients (7.9%) had 3 or 4 independent factors, forming the high-risk group.

Fig. 5A illustrates the cumulative incidence of HCC for these three groups. The log-rank test demonstrated a significant difference in cumulative HCC incidence among the groups (p < 0.001). The incidence rates of HCC at 1, 2, 3, and 4 years in the low-risk group were 0.6%, 1.5%, 1.8%, and 1.8%, respectively, compared with 2.7%, 6.8%, 6.8%, and 6.8% in the medium-risk group, and 9.4%, 27.0%, 29.2%, and 33.6% in the high-risk group.

Fig. 5.

Risk of HCC occurrence stratified by combining advanced fibrosis, AFP, ALBI grade, and BMI. A, The cumulative incidence of HCC development stratified by the presence of independent factor including advanced (F3-4) fibrosis, EOT₁₂ AFP >7 ng/mL, EOT₁₂ ALBI grade ≥2, and BMI ≥ 25 kg/m² (model A). Low-risk: 0-1 factor, medium-risk: 2 factors, high-risk: 3-4 factors. As determined by the log-rank test, the cumulative HCC incidence was significantly different between these groups (p < 0.001). B, The cumulative incidence of HCC development stratified after consideration the impact of HR (model B). We assigned advanced fibrosis as three points due to its high HR, and one point each to EOT₁₂ AFP >7 ng/mL, EOT₁₂ ALBI grade ≥2, and BMI ≥ 25 kg/m², respectively. Low-risk: 0-1 points, medium-risk: 2-4 points, high-risk: 5-6 points. As determined by the log-rank test, the cumulative HCC incidence was significantly different between these groups (p < 0.001). AFP = alpha-fetoprotein; ALBI = albumin-bilirubin; BMI = body mass index; DAAs = direct-acting antivirals; EOT₁₂ = end-of-treatment 12 wk; HCC = hepatocellular carcinoma; HR = hazard ratio.

Due to the wide variation in HRs for each risk factor (HR, 6.745 for fibrosis; HR, 2.214 for BMI), we developed another risk stratification model (model B) based on the impact of HR. We assigned three points to advanced fibrosis, considering its high HR and one point each to EOT₁₂ AFP > 7 ng/mL, EOT₁₂ ALBI grade ≥ 2, and BMI ≥ 25 kg/m², respectively. The total score could range from 0 to 6. The residual risk of HCC development after achieving SVR was categorized as low if the sum of the score was 0-1, medium if the sum of the score was 2-4, and high if the sum of the score was 5-6.

According to this model, there was a significant difference in cumulative HCC incidence among the three groups (p < 0.001). The incidence rates of HCC at 1, 2, 3, and 4 years in the low-risk group were 0.2%, 1.0%, 1.3%, and 1.3%, respectively, compared with 3.0%, 6.3%, 6.3%, and 6.3% in the medium-risk group, and 9.4%, 27.0%, 29.2%, and 33.6% in the high-risk group (Fig. 5B).

3.9. Impact of overweight on the prevalence of fatty liver disease and post-treatment laboratory value

In the current study, 46.2% of patients were classified as overweight, with a BMI ≥ 25 kg/m². Among individuals in the overweight group, the ratio of having fatty liver disease was significantly higher compared with those in the non-overweight group (56.3% vs 39.0%; p < 0.001). Additionally, the proportion of individuals who achieved sustained ALT normalization post-DAAs therapy was significantly lower in the overweight group compared with the nonoverweight group (78.1% vs 91.4%; p < 0.001).

4. DISCUSSION

With superior efficacy, shorter treatment duration, broader eligibility, and excellent safety profiles, the therapeutic landscape of chronic hepatitis C (CHC) has evolved dramatically following the introduction of all-oral DAAs. Successful anti-viral therapy has been shown to have beneficial effects, leading to improved clinical outcomes and a reduced risk of HCC.^10,23–26 However, a small proportion of patients with advanced fibrosis still face a residual risk of liver disease even after viral clearance. In this study, we provide insight into the development of de novo HCC in patients with CHC who have been successfully treated with all-oral DAAs. The results of this retrospective cohort study show that the cumulative incidence of de novo HCC was 6.1% over a period of 4.8 years in CHC patients who achieved sustained virologic response at 12 weeks (SVR₁₂). Our study identified four independent predictors: advanced fibrosis (F3-4), alpha-fetoprotein (AFP) levels greater than 7 ng/mL at the end of treatment (EOT12), ALBI grade of 2 or higher at EOT₁₂, and a BMI of 25 kg/m² or higher.

When considering the risk factors for developing HCC after successful treatment with DAAs, it is widely accepted that the presence of cirrhosis prior to therapy is the most significant risk factor.^27–29 This risk is particularly pronounced in patients with decompensated liver disease or severe portal hypertension, which can be indicated by a low platelet count or the presence of esophageal varices.^30–33 Thrombocytopenia, a common hematological abnormality in patients with chronic liver disease, has been identified as an independent risk factor for complications of cirrhosis and death in individuals with cirrhosis.^34,35 Recent studies have also shown that thrombocytopenia increases the likelihood of HCC after DAA treatment.^33,36 In our study, we aimed to utilize an integrated index that could more accurately reflect the severity of liver fibrosis. Following previous reports,²¹ we selected the validated FIB-4 scoring system, which incorporates age, platelet count, and aminotransferase levels to estimate the extent of hepatic fibrosis. We combined the score with typical clinical and imaging findings to define advanced liver fibrosis. Consistent with earlier research, our study demonstrated that the presence of advanced fibrosis (F3-4) was associated with a higher incidence of de novo HCC.^37–41 Despite achieving HCV clearance with DAAs, patients with advanced fibrosis (F3-4) had a cumulative incidence of 17.2% for developing HCC over 4.8 years, highlighting the importance of this finding for clinicians.

As mentioned earlier, the ALBI score/grade is a measure of liver function originally developed as a prognostic factor for patients with HCC. The ALBI score provides an objective assessment and can detect smaller changes in liver dysfunction compared with the Child-Pugh or MELD scores. Numerous publications have demonstrated the high prognostic value of ALBI in patients with various types and stages of chronic liver disease.²² In our study, we found that in CHC patients who achieved viral eradication with DAAs, an EOT₁₂ ALBI grade of 2 or 3 was independently associated with the occurrence of HCC in multivariate analysis. The cumulative incidence of HCC development was significantly higher in the EOT₁₂ ALBI grade 2 or 3 group than the EOT₁₂ ALBI grade 1 group over a period of 4.8 years (24.2% vs 2.6%; p < 0.001). These results emphasize the importance of functional liver reserve in relation to the residual risk of HCC in CHC patients successfully treated with DAAs, which is consistent with recently published studies.^42,43

Alpha-fetoprotein (AFP) is a glycoprotein produced by the liver and yolk sac. In normal physiological conditions, AFP levels rapidly decline after birth and remain low throughout life. AFP serves as a tumor marker for HCC and also reflects the activation of hepatic progenitor cells.⁴⁴ The baseline AFP levels can indicate liver regeneration, inflammation, fibrosis, as well as microscopic HCC. Treatment with DAAs may reduce inflammation and unmask tumor-derived AFP. Additionally, the proliferative activity of hepatocytes may contribute to hepatocarcinogenesis.⁴⁵ Therefore, we chose to evaluate the impact of the AFP level at the end of treatment (EOT₁₂) on the development of de novo HCC. Our study discovered that an EOT₁₂ AFP level > 7 ng/mL, but not a baseline AFP > 7 ng/mL, was an independent factor associated with de novo HCC. Patients with abnormal AFP levels after DAA therapy face a higher risk of future HCC development and should be monitored more closely. Similar findings have been reported in previous studies.^37,38,46 Importantly, our current study demonstrated that dynamic changes in AFP values before and after DAA therapy can predict the occurrence of HCC. Patients whose AFP values increased from normal (≤7 ng/mL) to abnormal (>7 ng/mL) after DAA treatment had the highest risk of future HCC development. Conversely, patients who consistently maintained normal AFP values (≤7 ng/mL) throughout the course of DAA treatment had the lowest risk of HCC development.

Interestingly, our present study revealed the clinical significance of metabolic derangement in the development of de novo HCC. In multivariate analysis, a BMI ≥ 25 kg/m² emerged as an independent risk factor for de novo HCC. Moreover, overweight patients were more likely to present with fatty liver and had lower rates of sustained ALT normalization after DAA treatment. Overweight individuals are prone to developing nonalcoholic fatty liver disease, which leads to hepatocyte inflammation, accelerated liver fibrosis progression, and an increased risk of HCC.⁴⁷ Although HCV clearance can result in partial or complete regression of liver steatosis, a significant number of patients with a SVR may still have steatosis unrelated to HCV. In a retrospective study of 515 patients with advanced fibrosis who achieved SVR with DAAs, multivariate analysis demonstrated that pretreatment steatosis was associated with a 7.5-fold increased risk of the combined end-point of all-cause mortality and HCC.⁴⁸ Additionally, another study showed that obesity (BMI ≥ 25 kg/m²) and heavy alcohol consumption (≥60 g/d) were independent factors for HCC development following HCV eradication with anti-virals.⁴⁹ Therefore, the presence of metabolic derangement in CHC patients, particularly in those without sustained ALT normalization after DAA treatment, may contribute to the development of de novo HCC.

Since the pathogenesis of de novo HCC may have multiple factors involved, the most effective approach for risk stratification is to combine independent risk factors into algorithms. To our knowledge, our study is the first large-scale, real-world investigation from Taiwan that examines the residual risk of HCC in DAAs-treated patients with SVR using a risk stratification model. Taking model A as an example, which comprises four independent factors, including advanced (F3-4) fibrosis, alpha-fetoprotein (AFP) level greater than 7 ng/mL at 12 weeks post-treatment (EOT₁₂), ALBI grade of 2 or higher at EOT₁₂, and BMI of 25 kg/m² or higher, the incidence rates of HCC at 1, 2, 3, and 4 years were 9.4%, 27.0%, 29.2%, and 33.6%, respectively, for SVR patients with at least three risk factors. In contrast, SVR patients with zero or only one risk factor had very low rates of developing HCC within 4 years (approximately 1.8%). With the widespread use of DAAs and the growing population of chronic hepatitis C (CHC) patients achieving SVR, our study introduces the concept that post-SVR monitoring frequency and intensity for liver disease progression and HCC development can be stratified using this informative model.

Our study results also offer valuable insights into public health policy. During the interferon (IFN) era, the initiation of anti-viral therapy was hindered by low success rates, frequent side effects, and ineligibility. In this article, we found that the incidence of de novo HCC following treatment with DAAs is low for patients with mild liver disease. However, patients with more advanced liver fibrosis stages still face a higher risk of developing HCC. Therefore, our study supports the early initiation of DAAs in the course of liver disease, aiming to prevent disease progression and reduce the costs associated with cirrhosis and HCC complications. Additionally, previous research has shown that CHC patients with decompensated cirrhosis have lower SVR rates with DAAs, which also impacts patient outcomes.^50,51 Given this, it may be worthwhile and cost-effective to screen and treat our CHC patients with DAAs at an earlier stage.

We must acknowledge several limitations in our study. First, the study design was retrospective. Second, the follow-up period was relatively short, and the data was collected from only one tertiary center in Taiwan, which limits the generalizability of the findings. Third, despite collecting numerous controlled factors, there may be variables that were not measured or are unmeasurable, which could potentially impact our results. Lastly, the sample size was not sufficiently large to allow for the division of study subjects into model derivation and validation cohorts for statistical analysis. Additionally, we would welcome external validation of the risk models we proposed to assess their predictive accuracy in the future.

In conclusion, our study demonstrates that the risk of de novo HCC persists in chronic hepatitis C (CHC) patients despite viral clearance through the use of DAAs. Independent variables associated with HCC occurrence include advanced (F3-4) fibrosis, alpha-fetoprotein (AFP) level greater than 7 ng/ml at 12 weeks post-treatment (EOT₁₂), ALBI grade of 2 or higher at EOT₁₂, and a BMI of 25 kg/m² or higher. Treating all CHC individuals, regardless of disease stage, offers the advantage of preventing disease progression and reducing the costs associated with cirrhosis-related complications, including developing hepatocellular carcinoma. By creating a risk stratification model using the aforementioned independent variables, we can accurately predict future HCC development post-SVR and refine screening strategies for this expanding population.

ACKNOWLEDGMENTS

This study was supported in part by Taipei Veterans General Hospital (V105C-068) and Szu-Yuan Research Foundation of Internal Medicine (108-033, 109-020, 110-016 and 111-008).

We thank Miss Yu-Ling Chou and Miss Chiao-Yu Chang for patient follow-up and clinical data collection.