Development of a nomogram for predicting the long-term risk of hepatocellular carcinoma after antiviral treatment for hepatitis C Authors: Jing Li, Baima Yangjin, Chengbin Zhu, Weicheng Xu, Yingli He Published: 2025 ABSTRACT BackgroundThe impact of direct acting antivirals (DAAs) on hepatitis C virus related hepatocellular carcinoma (HCV-HCC) incidence after sustained virologic response (SVR) remains controversial. We aimed to compare long-term HCC incidence after SVR induced by interferon (IFN) and DAA therapy, also to construct a nomogram model for predicting long-term HCC incidence after HCV cure.MethodsA total of 405 patients treated with IFN or DAA were analyzed. 200 patients who achieved SVR were assigned to IFN and DAA group by propensity score matching. The COX regression, LASSO logistic regression, and best subset regression (BSR) were adopted to select variables independently associated with HCC incidence and construct nomogram.ResultsAfter a median follow-up period of 72.0 (interquartile range 57.5–90.0) months, HCC developed in 7 (7%) and 15 (15%) patients in IFN and DAA groups, respectively. The cumulative HCC rates were comparable between the two groups (HR: 0.917, 95% CI: 0.262–1.968, P = 0.520). Out of all three models constructed by COX, LASSO and BSR, model 2 and 3 both contained gender, age, platelet count and alkaline phosphatase, exhibited the minimum Akaike Information Criterion (174.54) and the best predictive performance for 8-year HCC incidence (AUROC = 0.905). Using these four factors, a novel nomogram was established to predict 8-year HCC occurrence following HCV eradication. It showed significant clinical benefit according to decision curve analysis.ConclusionsDAAs exert the same long-term protect role against HCC as IFN treatment, though the development of HCC is still inevitable. Herein, we developed a 8-year HCC-predicting nomogram that can be used to guide clinical screening strategies.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12876-025-04412-x. Background Hepatocellular carcinoma (HCC) remains a major worldwide health problem, with an annual incidence rate of 1–6% [1]. Chronic hepatitis C virus (HCV) infection is one of the major risk factors for HCC, and it is estimated that 2.5% of the world population is infected with HCV. The reported annual risk of developing HCC in chronic hepatitis C (CHC) patients is 1–4%, and rises to 3–8% in CHC patients with liver cirrhosis [2]. The combination of interferon (IFN) and ribavirin has been a standard treatment of chronic hepatitis C. However, the cure rate is unsatisfactory [3, 4]. The recent introduction of direct acting antivirals (DAAs) has revolutionized HCV treatment. It improved the rate of viral eradication, defined by sustained virologic response (SVR) to 95–100% [5, 6], even in elder patients and patients with advanced liver disease that were not candidates for IFN-based therapies [7, 8]. It is well accepted that HCV cure following IFN-based therapy could reduce the development of HCC [9–12]. However, whether the HCC occurrence or recurrence becomes higher or lower by DAA treatment is still controversial. Conti et al. reported a 28.9% (17/59) HCC recurrence in patients with previous HCC [13]. Another study by Elkassas et al. reported an even higher recurrence rate (37.7%) after 16 months of follow up [14]. However, some later studies suggest that DAAs may even suppress HCC development [15]. In a retrospective cohort analysis of 793 patients with HCV-related HCC, the recurrence rate between treated and untreated patients was 42.1% and 58.9%, respectively [16]. Imai et al. also revealed that HCV eradication by DAA was an independent factor for the prevention of HCC recurrence [17]. Considering that immune response and metabolic status are different between CHC patients with or without previous HCC, we assumed that DAAs may affect the occurrence and recurrence rate of HCC in different ways. DAAs also seem to increase early HCC occurrence in CHC patients with cirrhosis [13, 18]. However, several reports found that DAA therapy may suppress HCC development, which is similar to IFN therapy. For instance, Li et al. found that those treated with either IFN or DAA therapy had a reduced HCC incidence when compared with no treatment [19]. A retrospective study revealed that the cumulative incidence rate of 3-year HCC was similar in the two cohorts of HCV patients treated with DAAs and IFN [20]. There remain some concerns regarding the real benefit obtained from DAA therapy. Firstly, the long-term benefit of DAA therapy have not been fully established since the follow-up time of DAAs treated patients is not long enough. Until now, the follow-up period time ranged from 1 to 3 years in most studies. Secondly, given that patients with decompensated cirrhosis or elder patients are not suitable for IFN therapy, the populations treated with IFN and DAAs are different in the background characteristics. Overall, it is unclear if HCC occurrence will decline and whether there are risk factors related to HCC over the long term after HCV cure. Previous studies have demonstrated that male sex, older age, high liver stiffness and diagnosis of cirrhosis are the risk factors of early occurrence of HCC [21–23]. But there is still a lack of easy-to-apply method for accurately estimating HCC incidence in patients who achieve SVR. For this purpose, we evaluated and compared the long-term incidence of HCC after successful IFN or DAA treatment in chronic hepatitis C patients and patients with compensated cirrhosis. We also evaluated the factors associated with HCC occurrence using pre-treatment data and developed a simple HCC-predicting nomogram of patients at specific time points during follow-up after SVR. Methods Patients We undertook an analysis of a retrospective cohort study at the First Affiliated Hospital of Xi’an Jiaotong University from January 2002 to May 2018, enrolling a total of 405 consecutive patients with HCV infection who received IFN or DAA treatment. Patients with the following conditions were excluded: (1) baseline age less than 18 years old; (2) patients with a previous history of HCC or other malignant tumors; (3) patients who have underwent liver transplantation; (4) patients co-infected with hepatitis B virus or human immunodeficiency virus, or patients with other liver diseases, including non-alcoholic fatty liver disease, autoimmune hepatitis and so on. Finally, a total of 368 patients were included in the present study. Among them, 288 cases achieved SVR after antiviral treatment. To mitigate the impact of baseline characteristic differences, 100 cases were selected from each group (IFN group and DAA group) after 1:1 Propensity Score Matching (PSM) based on gender, age, diabetes and baseline liver disease status. Liver disease severity was assessed using Child-Turcotte-Pugh score. The selection of the study population is illustrated in Supplementary Fig. 1. Liver cirrhosis was diagnosed by the liver imaging examination (ultrasonography, computed tomography or magnetic resonance imaging), as well as the patient’s clinical state. SVR was defined as undetectable HCV RNA performed at least 12 weeks after the end of the antiviral treatment. Diabetes mellitus (DM) was defined as DM pattern in 75 g oral glucose tolerance test, HbA1c >6.5% or taking anti-diabetes drugs and/or injecting insulin before starting DAA therapy [24]. The study was carried out in accordance with the guidelines of the 1975 Declaration of Helsinki and was approved by the Institutional Ethics Committee for Human Studies at the First Affiliated Teaching Hospital of Xi’an Jiaotong University. Written informed consent was obtained from all subjects. Hepatocellular carcinoma occurrence HCC surveillance was performed in all patients every three to six months by blood test including α-fetoprotein (AFP) and imaging examination. Imaging examination was performed every 6 months for patients with chronic hepatitis and every 3 months for patients with liver cirrhosis. HCC was diagnosed histologically and/or based on imaging findings according to the international guidelines [25]. Clinical and laboratory assessments Clinical and laboratory parameters, including liver function tests (alanine aminotransferase (ALT), total bilirubin (TB), serum albumin (ALB), alkaline phosphatase (ALP), gamma-glutamyl transpeptidase (GGT)), renal function (serum urea(UA)), complete blood count (red blood cell count (RBC), white blood cell count (WBC), platelet counts (PLT), neutrophil count (NEUT)) and coagulation (international normalized ratio) were performed at the central lab of the hospital. Serum HCV RNA levels were estimated by quantitative real time PCR assay using the COBAS Ampliprep/Cobas Taqman HCV test v.2.0, with a lower limit of quantification of 15 IU/mL. Clinical and laboratory parameters on admission and follow-ups (every three to six months) were retrieved from the hospital information system. Statistical analysis The data were analyzed using SPSS version 19.0 (SPSS, Chicago, IL, USA) and R software (version 4.3.2, Regression Modelling Strategies package. The R Foundation for Statistical Computing, Vienna, Austria). The data are presented as the mean ± standard derivation or median (interquartile ranges) for quantitative data and as proportions for categorical data. Differences between subgroups were analyzed using the Chi-squared test for categorical parameters and the Wilcoxon matched-pairs signed-rank test or Mann-Whitney U-test for continuous parameters, as appropriate. The HCC incidence rate was evaluated using Kaplan-Meier survival analysis. The log-rank test was used to identify differences in survival among groups. Multivariate Cox hazard regression analysis was performed to predict significant factors related to HCC incidence. Factors that were statistically significant in the univariate analysis were entered into the multivariate Cox hazard regression analysis to identify significant independent factors. Hazard ratios (HRs) and 95% confidence intervals (CIs) were also calculated. All statistical tests were two-tailed, and a probability level of P < 0.05 was considered statistically significant. In order to avoid over-fitting or under-fitting of the model, three statistical methods, namely, COX regression screening, the selection operator (LASSO) logistic regression, and the best subset regression (BSR) were adopted to select variables independently associated with HCC incidence using a backward stepwise selection, respectively. The final model was determined by the Akaike Information Criterion (AIC) and the receiver operating characteristics (ROC) curves with best performance. In the model development phase, a backward stepwise selection process was performed using a threshold of P < 0.05 to establish a model, and a nomogram was constructed to predict the HCC incidence for each patient at 8-year of follow up, according to calculated points associated with each risk factor evaluated. Each variable had its own score scale. To evaluate the discriminative ability of our model, we used the area under the receiver operating characteristic curve (AUROC). Calibration was assessed through calibration curves, and the clinical use was evaluated by a decision curve analysis (DCA). Results Patients’ characteristics A total of 288 patients, who underwent IFN or DAA treatment and obtained SVR were enrolled in the present study. Patients’ characteristics at baseline are shown in Supplementary Table 1. 200 patients from the IFN and DAA groups were matched for gender, age, diabetes and baseline liver disease status. There were 42 patients with compensated cirrhosis and 58 patients with CHC in the IFN group, and the number of patients with compensated cirrhosis and CHC in the DAA group was 39 and 61, respectively. Baseline characteristics of the study population after PSM are shown in Table 1. The median observation period was 88.5 and 63.0 months for the IFN and DAA group, respectively. As expected, baseline characteristics were well balanced between the two groups, except that the observation period was longer in the IFN group. It’s noted that the mean PLT levels of both IFN and DAA group were below the normal lower limit (125 × 109/L), possibly because of the including of patients with compensated liver cirrhosis in both groups. Table 1Baseline characteristics of the study population after propensity score matching based on gender, age, diabetes and baseline liver disease statusCharacteristicTotal (n = 200)IFN (n = 100)DAA (n = 100)P valueHCC (%)22 (11.0)7 (7.0)15 (15.0)0.114Male sex (%)81 (40.5)42 (42.0)39 (39.0)0.773Age (years)54 ± 1055 ± 1154 ± 90.720DM (%)30 (15.0)15 (15.0)15 (15.0)1.000Cirrhosis (%)81 (40.5)39 (39.0)42 (42.0)0.773HCV RNA (lg IU/mL)5.62 (4.79–6.42)5.71 (5.05–6.45)5.56 (4.65–6.23)0.429ALT (U/L)59.1 (32.3–107.1.3.1)52.0 (30.0–107.1.0.1)62.1 (34.0–103.5.0.5)0.266ALP (U/L)90.4 (71.5–113.5.5.5)88.0 (70.7–114.3.7.3)90.9 (73.2–112.6.2.6)0.584GGT (U/L)44.6 (25.2–83.5)43.2 (25.2–81.7)47.0 (25.1–85.5)0.708TBIL (µmol/L)15.1 (10.5–20.7)14.5 (10.1–20.3)16.0 (11.6–20.8)0.214ALB (g/L)38.3 ± 4.637.9 ± 4.638.7 ± 4.60.270 UA (µmol/L) 295 (239–354) 301 (248–360) 279 (231–346) 0.488 TC (mmol/L) 3.7 (3.2–4.2) 3.6 (3.1–4.1) 3.7 (3.2–4.2) 0.485 INR 1.10 (1.04–1.16) 1.10 (1.05–1.15) 1.09 (1.03–1.16) 0.601RBC (1012/L)4.1 ± 0.64.1 ± 0.64.2 ± 0.50.210WBC (109/L)4.2 (3.1–5.1)3.9 (3.1–4.8)4.4 (3.2–5.3)0.126PLT (109/L)110 (71–150)111 (70–145)109 (72–156)0.740NEUT (109/L)2.0 (1.5–2.8)2.0 (1.5–2.9)2.0 (1.4–2.7)0.583Observation period (months)72.0 (57.5–90.0)88.5 (73.0–96.0)63.0 (47.3–72.5)< 0.001IFN Interferon, DAA Direct-active-antiviral, HCC Hepatocellular carcinoma, DM Diabetes mellitus, ALT Alanine aminotransferase, ALP Alkaline phosphatase, GGT Gammaglutamyl transpeptidase, TBIL Total bilirubin, ALB Serum albumin, UA Serum urea, TC Cholesterol, INR International normalized ratio, RBC Red blood cell count, WBC White blood cell count, PLT Platelet count, NEUT Neutrophil count Table 1Baseline characteristics of the study population after propensity score matching based on gender, age, diabetes and baseline liver disease statusCharacteristicTotal (n = 200)IFN (n = 100)DAA (n = 100)P valueHCC (%)22 (11.0)7 (7.0)15 (15.0)0.114Male sex (%)81 (40.5)42 (42.0)39 (39.0)0.773Age (years)54 ± 1055 ± 1154 ± 90.720DM (%)30 (15.0)15 (15.0)15 (15.0)1.000Cirrhosis (%)81 (40.5)39 (39.0)42 (42.0)0.773HCV RNA (lg IU/mL)5.62 (4.79–6.42)5.71 (5.05–6.45)5.56 (4.65–6.23)0.429ALT (U/L)59.1 (32.3–107.1.3.1)52.0 (30.0–107.1.0.1)62.1 (34.0–103.5.0.5)0.266ALP (U/L)90.4 (71.5–113.5.5.5)88.0 (70.7–114.3.7.3)90.9 (73.2–112.6.2.6)0.584GGT (U/L)44.6 (25.2–83.5)43.2 (25.2–81.7)47.0 (25.1–85.5)0.708TBIL (µmol/L)15.1 (10.5–20.7)14.5 (10.1–20.3)16.0 (11.6–20.8)0.214ALB (g/L)38.3 ± 4.637.9 ± 4.638.7 ± 4.60.270 UA (µmol/L) 295 (239–354) 301 (248–360) 279 (231–346) 0.488 TC (mmol/L) 3.7 (3.2–4.2) 3.6 (3.1–4.1) 3.7 (3.2–4.2) 0.485 INR 1.10 (1.04–1.16) 1.10 (1.05–1.15) 1.09 (1.03–1.16) 0.601RBC (1012/L)4.1 ± 0.64.1 ± 0.64.2 ± 0.50.210WBC (109/L)4.2 (3.1–5.1)3.9 (3.1–4.8)4.4 (3.2–5.3)0.126PLT (109/L)110 (71–150)111 (70–145)109 (72–156)0.740NEUT (109/L)2.0 (1.5–2.8)2.0 (1.5–2.9)2.0 (1.4–2.7)0.583Observation period (months)72.0 (57.5–90.0)88.5 (73.0–96.0)63.0 (47.3–72.5)< 0.001IFN Interferon, DAA Direct-active-antiviral, HCC Hepatocellular carcinoma, DM Diabetes mellitus, ALT Alanine aminotransferase, ALP Alkaline phosphatase, GGT Gammaglutamyl transpeptidase, TBIL Total bilirubin, ALB Serum albumin, UA Serum urea, TC Cholesterol, INR International normalized ratio, RBC Red blood cell count, WBC White blood cell count, PLT Platelet count, NEUT Neutrophil count Baseline characteristics of the study population after propensity score matching based on gender, age, diabetes and baseline liver disease status IFN Interferon, DAA Direct-active-antiviral, HCC Hepatocellular carcinoma, DM Diabetes mellitus, ALT Alanine aminotransferase, ALP Alkaline phosphatase, GGT Gammaglutamyl transpeptidase, TBIL Total bilirubin, ALB Serum albumin, UA Serum urea, TC Cholesterol, INR International normalized ratio, RBC Red blood cell count, WBC White blood cell count, PLT Platelet count, NEUT Neutrophil count Cumulative incidence rate of HCC During a median follow-up period of 72.0 (interquartile range 57.5–90.0) months, 7 and 15 out of 100 patients in the IFN and DAA group developed HCC, respectively (Table 2). Therefore, the overall HCC incidence rate was 11% in the present study. The 36-, 60-, and 96-month cumulative incidence rates of HCC post-SVR were 4%, 6%, and 7% for patients in the IFN group, compared to 0%, 2%, and 11% for patients in the DAA group, respectively. However, the cumulative HCC rates were not significantly different between these two groups after HCV eradication (HR: 0.917, 95% CI: 0.262–1.968, P = 0.520) (Fig. 1). Table 2Univariate and multivariate analysis of HCC incidence in HCV-infected patients achieving SVR after antiviral therapySingle-factor Cox analysisMulti-factor Cox analysisHR95%CIP valueHR95%CIP valueAge (years)1.0571.008–1.1080.0221.0490.995–1.1050.075Sex (Male:1)2.2240.948–5.2170.0662.5040.923–6.7950.072ALB (g/L)0.9020.825–0.9850.0220.8980.804–1.0020.055ALP(U/L)1.0090.998–1.0210.117ALT(U/L)1.0010.996–1.0060.718AST(U/L)1.0030.998–1.0080.264CRE(µmol/L)1.0010.986–1.0170.855GGT(U/L)1.0020.998–1.0060.357LYMPH (109/L)0.6430.344–1.200.344.2000.165NEUT (109/L)0.7020.405–1.2170.208PLT (109/L)0.9800.969–0.990< 0.0010.9750.961–0.9890.001RBC (1012/L)0.5570.252–1.2300.148TBIL(µmol/L)1.0260.987–1.0660.192TC (mmol/L)0.5260.310–0.8910.0171.3470.597–3.0390.472UA(µmol/L)1.0051.000–1.009.000.0090.0311.0010.996–1.0070.642WBC (109/L)0.7660.564–1.0410.0881.0510.718–1.5390.798HCV RNA (lg IU/mL)1.0050.765–1.3210.971DM(DM:1)1.1320.382–3.3570.823Baseline liver disease (Cirrhosis :1)2.7481.148–6.5760.0231.2200.323–4.5990.769Antiviral therapy modes(DAA:1)0.7080.257–1.9500.504HCC Hepatocellular carcinoma, SVR Sustained virological response, ALB Serum albumin, ALP Alkaline phosphatase, ALT Alanine aminotransferase, AST Aspartate aminotransferase, CRE Creatinine, DBIL Direct bilirubin, GGT Gammaglutamyl transpeptidase, LYMPH Lymphocyte count, NEUT Neutrophil count, PLT Platelet count, RBC Red blood cell count, TBIL Total bilirubin, TC Cholesterol, UA Serum urea, WBC White blood cell count, DM Diabetes mellitus, DAA Direct-active-antiviral Table 2Univariate and multivariate analysis of HCC incidence in HCV-infected patients achieving SVR after antiviral therapySingle-factor Cox analysisMulti-factor Cox analysisHR95%CIP valueHR95%CIP valueAge (years)1.0571.008–1.1080.0221.0490.995–1.1050.075Sex (Male:1)2.2240.948–5.2170.0662.5040.923–6.7950.072ALB (g/L)0.9020.825–0.9850.0220.8980.804–1.0020.055ALP(U/L)1.0090.998–1.0210.117ALT(U/L)1.0010.996–1.0060.718AST(U/L)1.0030.998–1.0080.264CRE(µmol/L)1.0010.986–1.0170.855GGT(U/L)1.0020.998–1.0060.357LYMPH (109/L)0.6430.344–1.200.344.2000.165NEUT (109/L)0.7020.405–1.2170.208PLT (109/L)0.9800.969–0.990< 0.0010.9750.961–0.9890.001RBC (1012/L)0.5570.252–1.2300.148TBIL(µmol/L)1.0260.987–1.0660.192TC (mmol/L)0.5260.310–0.8910.0171.3470.597–3.0390.472UA(µmol/L)1.0051.000–1.009.000.0090.0311.0010.996–1.0070.642WBC (109/L)0.7660.564–1.0410.0881.0510.718–1.5390.798HCV RNA (lg IU/mL)1.0050.765–1.3210.971DM(DM:1)1.1320.382–3.3570.823Baseline liver disease (Cirrhosis :1)2.7481.148–6.5760.0231.2200.323–4.5990.769Antiviral therapy modes(DAA:1)0.7080.257–1.9500.504HCC Hepatocellular carcinoma, SVR Sustained virological response, ALB Serum albumin, ALP Alkaline phosphatase, ALT Alanine aminotransferase, AST Aspartate aminotransferase, CRE Creatinine, DBIL Direct bilirubin, GGT Gammaglutamyl transpeptidase, LYMPH Lymphocyte count, NEUT Neutrophil count, PLT Platelet count, RBC Red blood cell count, TBIL Total bilirubin, TC Cholesterol, UA Serum urea, WBC White blood cell count, DM Diabetes mellitus, DAA Direct-active-antiviral Univariate and multivariate analysis of HCC incidence in HCV-infected patients achieving SVR after antiviral therapy Antiviral therapy modes (DAA:1) HCC Hepatocellular carcinoma, SVR Sustained virological response, ALB Serum albumin, ALP Alkaline phosphatase, ALT Alanine aminotransferase, AST Aspartate aminotransferase, CRE Creatinine, DBIL Direct bilirubin, GGT Gammaglutamyl transpeptidase, LYMPH Lymphocyte count, NEUT Neutrophil count, PLT Platelet count, RBC Red blood cell count, TBIL Total bilirubin, TC Cholesterol, UA Serum urea, WBC White blood cell count, DM Diabetes mellitus, DAA Direct-active-antiviral Fig. 1Cumulative incidence of HCC development in patients with SVR after IFN or DAA therapy. IFN, interferon; DAA, direct-active-antiviral; HCC, hepatocellular carcinoma; SVR, sustained virological response Cumulative incidence of HCC development in patients with SVR after IFN or DAA therapy. IFN, interferon; DAA, direct-active-antiviral; HCC, hepatocellular carcinoma; SVR, sustained virological response Risk factors of HCC development Because no significant difference of the cumulative incidence rate of HCC post-SVR was found between the IFN and DAA group, and the sample size were relative small for each group after PSM, we therefore analyzed predictive factors using the total 200 patients (IFN group + DAA group). Potential prognostic factors for HCC occurrence included baseline liver disease status, age, gender and diabetes status were analysed by Kaplan–Meier survival analysis. As assessed by the log-rank test, the cumulative HCC incidence was significantly higher in liver cirrhosis (LC) patients than in CHC patients (HR = 2.748; 95% CI:1.148–6.576, P = 0.023). Similarly, the cumulative HCC incidence was significantly higher in younger patients (≤ 60 years old) than in patients over 60 years old (HR = 2.450; 95% CI:1.059–5.669, P = 0.036). Furthermore, the cumulative HCC incidence was higher in male than in female, though the difference was not statistically significant (HR = 2.224; 95% CI:0.948–5.217, P = 0.066). As for diabetes status, there was no difference in patients with or without diabetes (HR = 1.132; 95% CI:0.382–3.357, P = 0.823) (Fig. 2). Fig. 2The cumulative HCC incidence stratified according to the factors that might contribute to the development of HCC. A Cumulative incidence of HCC in patients with CHC (red line) and patients with compensated liver cirrhosis (blue line); (B) Cumulative incidence of HCC in patients who were ≤ 60 years old (red line) or > 60 years old; (C) Cumulative incidence of HCC in male (red line) and that in female (blue line); (D) Cumulative incidence of HCC in patients without diabetes (red line) and patients with diabetes (blue line). HCC, hepatocellular carcinoma; CHC, chronic hepatitis C; DM, diabetes mellitus The cumulative HCC incidence stratified according to the factors that might contribute to the development of HCC. A Cumulative incidence of HCC in patients with CHC (red line) and patients with compensated liver cirrhosis (blue line); (B) Cumulative incidence of HCC in patients who were ≤ 60 years old (red line) or > 60 years old; (C) Cumulative incidence of HCC in male (red line) and that in female (blue line); (D) Cumulative incidence of HCC in patients without diabetes (red line) and patients with diabetes (blue line). HCC, hepatocellular carcinoma; CHC, chronic hepatitis C; DM, diabetes mellitus The results of multivariate Cox hazard regression analysis with the factors significantly associated with HCC occurrence after HCV elimination were shown in Table 2. Univariate analysis demonstrated that age (P = 0.022), ALB (P = 0.022), PLT (P < 0.001), TC (P = 0.017), UA (P = 0.031), and baseline liver disease status (P = 0.023) were significant risk factors that may contribute to HCC occurrence. Significant factors (P < 0.1) were entered into the multivariate analysis, which identified that PLT (HR = 0.975; 95% CI:0.961–0.989, P = 0.001) were independent risk factors for incidence of HCC post-SVR (Table 2). Development of a HCC-predicting nomogram We then adopted three statistical methods (COX, LASSO and BSR) to select variables independently associated with HCC incidence using a backward stepwise selection. Accordingly, we developed three models: model 1 consisted of three variables based on the COX regression, including gender, age and PLT; model 2 and model 3 both consisted of four variables, including gender, age, PLT and ALP, based on the LASSON regression and BSR, respectively. We further used AIC and the ROC curve to examine efficiency of these models. As a result, model 2 and 3 showed the smallest AIC among the three regression models (model 1:175.52; model 2:174.54; model 3:174.54). We then plotted ROC curves for HCC incidence at 36, 60 and 96 months and quantified the area under the ROC curve(AUROC). As shown in Fig. 3, model 2 and 3 had an AUROC of 0.812, 0.901 and 0.905, respectively. In contrast, model 1 gave an AUROC of 0.745, 0.869 and 0.866, confirming the better performance of model 2 and 3. Fig. 3ROC analysis of HCC incidence at (A) 36-, (B) 60-, and (C) 96- month for the three models. HCC, hepatocellular carcinoma ROC analysis of HCC incidence at (A) 36-, (B) 60-, and (C) 96- month for the three models. HCC, hepatocellular carcinoma Therefore, the nomogram predicting HCC incidence post-SVR should be developed based on four factors of model 2 and 3: gender, age, PLT and ALP. It was assigned a score to each variable, and thus by summing up each point to obtain a total score and, drawing a vertical line from the total score (Total Point) up to the probability scale, at 8-year of HCC incidence at the bottom of the nomogram (Fig. 4). For example, a 65-year-old (41 points) male patient (13 points), with a PLT count of 100 × 10^9/L (71 points), and ALP level of 80U/L (10 points) has a total scoring of 135 points; corresponding to a probability of developing HCC at 8-year of about 30%. Finally, we evaluated the clinical usage of the nomogram by ROC curves, calibration curves and DCA, respectively. The AUROCs for 36-mouth, 60-mouth, and 96-mouth HCC-incidence predictions were 0.812, 0.901, and 0.889, respectively (Supplementary Fig. 2 A). The calibration curve showed strong agreement between predicted and observed HCC incidence, closely aligning with the ideal 45º line (Supplementary Fig. 2B). DCA further demonstrated good consistency between estimated values and clinical results (Supplementary Fig. 2 C). Fig. 4Nomogram of the HCC-predicting model based on gender, age, PLT and ALP. HCC, hepatocellular carcinoma; PLT, platelet count; ALP, alkaline phosphatase Nomogram of the HCC-predicting model based on gender, age, PLT and ALP. HCC, hepatocellular carcinoma; PLT, platelet count; ALP, alkaline phosphatase In the clinical practice, we recommend stratifying patients into high-risk and low-risk groups based on the nomogram. High-risk individuals (8-year predicted incidence ≥ 20%) are suggested to maintain intensive surveillance every 6 months. The screening methods should at least include abdominal ultrasound combined with AFP testing. For low-risk individuals (8-year predicted incidence < 20%), the screening interval can be changed to once a year. Additionally, we recommend reassessing the HCC-predicting nomogram at least once a year to provide a precise and dynamic clinical management for patients. Discussion In the management of chronic hepatitis C, the achievement of SVR by IFN treatment was demonstrated to reduced the incidence of HCC, with a residual annual incidence lower than 2% [26, 27]. Recently, the development of DAAs has brought excellent rates of viral eradication, even in older and cirrhotic patients. As a result, the vast majority of treated patients will achieve HCV eradication in the DAA era. Although this remarkable rate of DAA-induced SVR was assumed to be associated with a reduction in HCC risk, it should be noted that HCC may inevitably occur or recur in some patients after achievement of SVR, especially in patients with advanced fibrosis and cirrhosis. However, the occurrence rate of HCC after DAA-induced HCV cure remains inconclusive, probably because of the differences in the background characteristics, such as different populations, genetic variabilities, different follow-up periods, different sample size and so on. Of note, most studies failed to show the long-term association between DAAs therapy and the development of HCC. Also, the population characteristics are not comparable between IFN and DAA treatment. Based on this background, we evaluated the cumulative HCC rate of HCV cure after a median follow-up period of 72 months, which is the longest among the present studies as we known. We further used a propensity score analysis to overcome bias due to the different distributions of covariates among patients treated with IFN or DAAs. According to our study, the 36-, 96-, and 96-month incidence rates of HCC for patients in IFN and DAA group were 4%, 6%, 7% and 0%, 2%, 11%, respectively. We found that the cumulative HCC rates were not significantly different between patients in IFN and DAA group (HR: 0.917, 95% CI: 0.262–1.968, P = 0.520)(Fig. 1). Overall, our results indicated that HCV cure achieved by DAAs exerted the same long-term protect role against HCC as IFN treatment, though the development of HCC is still inevitable in some particular populations in the DAA era. Current clinical guidelines recommended that clinicians should perform HCC surveillance using ultrasound with or without AFP every 6–12 months in all F3 and F4 patients after HCV cure [25, 28]. However, there are still some concerns. Firstly, the performance of non-invasive assessments remain imperfect. Secondly, our data are in general agreement that DAA-induced HCV elimination does not eradicate the risk of HCC, even in patients without cirrhosis and patients with compensated cirrhosis. Finally, patients who are destined to develop HCC will benefit from the surveillance, whereas others may incur more expenses and potential harms, including anxiety and labour productive losses. Therefore, the development of methods that can predict HCC risk will help to balance medical expenses and the benefits of screening, and then help to assist clinical decision-making. One aim of this study was to assess potential predictive factors associated with HCC occurrence after achieving SVR by antiviral therapy. Previous studies have shown that sex, age, PLT, GGT, advanced liver fibrosis, DM and obesity were risk factors related to HCC occurrence [21–23, 29–32]. According to our results, cirrhosis and advanced age were associated with higher HCC incidence (Fig. 2A-B), which is consistent with other studies [22, 23, 29, 30]. Male also exhibited a higher cumulative HCC incidence, though the difference was not statistically significant(P = 0.066). Moreover, DM, different from other results [22, 29, 31], was not associated with an increase risk of HCC, potentially explained by different study population and observation time. Indeed, Alonso et al. found that the risk of HCC did not differ in patients with DM after a median follow up of 45 months [33]. In the present study, multivariate analysis further identified PLT as independent factor that contributed to the development of HCC after HCV cure (Table 2). However, cirrhosis and gender were not extracted by multivariate analysis. In order to establish a reliable HCC-predicting nomogram, we adopted three statistical methods to select variables independently associated with HCC incidence. The LASSON regression and BSR performed better according to the AIC and ROC curves. Of note, the model of LASSON and BSR exhibited excellent predictive value, with an AUROC of 0.812, 0.901 and 0.905 for the prediction of HCC incidence at 36, 60 and 96 months of post-SVR, respectively (Fig. 3). Based on the LASSON and BSR methods, we screened out four candidate indicators, including gender, age, PLT and ALP. Among them, age and PLT were also extracted by multivariate analysis as mentioned above. In line with our finding, several studied reported that advanced age and a lower level of PLT were high risk factors for the development of HCC [22, 23, 29, 32]. Furthermore, male gender are known to be linked with HCC development after HCV eradication [22, 34]. As for ALP, it was rarely combined in models or scoring systems to predict the risk of HCC. However, ALP have been recognized as independent risk factors for the early recurrence of combined hepatocellular-cholangiocarcinoma [35]. We then developed a nomogram to predict the 8-year HCC occurrence after HCV cure (Fig. 4). More importantly, We employed multiple approaches to validate the performance of our model(Supplementary Fig. 2). Therefore, using our predictive model may further differentiate populations who are at high risk of developing HCC in the long-term of 8 years, even after HCV eradication. This will facilitate prompt clinical decision-making and play an important role in the long-term management of patients after HCV cure. It should be noted that there were some limitations to our study. First, since the study is retrospective, the results need to be verified in additional prospective studies. Second, because of the lack of internal or external validation, the generalizability of our HCC-predicting nomogram need to be validated in another multi-center study using larger sample sizes. Third, only pre-treatment clinical characteristics were used. Finally, the incomplete information of BMI and liver-stiffness measurement may induce confounding and bias in the estimation of predictive factors that cannot be ruled out. However, we evaluated PLT and liver disease status, which were related to the liver fibrosis progression. Conclusions Our study revealed that the long-term HCC occurrence was comparable in patients treated with DAAs and IFN after HCV cure, signified that DAA exerted the same protect role against HCC as IFN therapy in the long-term. Furthermore, the novel HCC-predicting nomogram offers a reliable and easy-to-use method to assess the 8-year risk of HCC occurrence after HCV eradication. Clinicians may take it into account in their identification of patients who would benefit from HCC surveillance. Supplementary Information Supplementary Material 1. Supplementary Material 1.