A novel scoring system to identify patients at low risk for hepatocellular carcinoma five years after DAA-Induced sustained virological response: a retrospective cohort study Authors: Hideki Fujii, Takeshi Nishimura, Hiroyuki Kimura, Yasuhide Mitsumoto, Toshihide Shima, Yuya Seko, Michihisa Moriguchi, Seita Kataoka, Saiyu Tanaka, Hiroki Ishikawa, Shiro Takami, Yasuyuki Nagao, Atsushi Umemura, Yoshito Itoh, Kanji Yamaguchi Published: 2025 ABSTRACT AbstractBackgroundTreating chronic hepatitis C virus infection with direct-acting antivirals (DAAs) results in a sustained virological response (SVR) in most cases, but hepatocellular carcinoma (HCC) can occur in some patients after achieving SVR. Thus, this study comprehensively assessed the risk factors associated with the development of HCC during long-term follow-up in patients who achieved SVR. Furthermore, we aimed to identify a group of patients without prior HCC treatment who are at extremely low risk for HCC long after achieving SVR.MethodsThis retrospective cohort study included patients who achieved SVR after DAA therapy, and predictors contributing to the incidence of HCC were determined via Cox regression analysis. Specifically, we evaluated the factors contributing to the incidence of HCC among patients without previous HCC treatment later than 5 years after achieving SVR, and these were used to construct a predictive scoring system for HCC.ResultsOverall, 267 cases of HCC were observed during follow-up. Male sex, older age, previous HCC treatment before starting DAA therapy, excessive alcohol use, diabetes mellitus, elevated γ-GTP, low albumin, and low platelet count were independent predictors of HCC. Since previous HCC treatment strongly influenced the incidence of HCC after SVR, we further analyzed the risk factors for HCC among patients without previous HCC treatment. In that cohort, parameters at 5 years after achieving SVR that contributed to the incidence of HCC later than 5 years after SVR were an age of ≥ 70 years, diabetes mellitus (DM), and fibrosis-4 index(Fib-4) of ≥ 2.67. These three factors were used to create a scoring system, with scores of 0 to 3 predicting a 0.8%, 2.7%, 11.9%, and 28.8%, respectively, cumulative incidence rate of HCC later than 5 years after SVR.ConclusionsUsing long-term follow-up data, we investigated the risk factors for HCC among patients who achieved SVR after DAA therapy. Previous HCC treatment strongly predicted HCC after SVR. Although this was a retrospective cohort study, among patients without previous HCC treatment, those younger than 70 years, without DM and with a Fib-4 index of < 2.67 had a very low risk of developing HCC later than 5 years after SVR.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12876-025-04451-4.Keywords: Chronic hepatitis c virus infection, Direct-acting antiviral, Hepatocellular carcinoma, Sustained virological response, Age, Diabetes mellitus, Fibrosis-4 index, Long-term follow-up Background Hepatitis C virus (HCV) infection is a major public health concern and it is a leading cause of liver diseases that lead to cirrhosis or hepatocellular carcinoma (HCC). The World Health Organization reports that the worldwide prevalence of HCV infection is 50 million, with a yearly incidence of 1.0 million new cases, and 242,000 deaths annually are attributed to HCV infection [1]. In Japan, approximately 1.5 to 2 million are infected with HCV [2, 3]. Fortunately, with the advent of direct-acting antiviral (DAA) drugs, almost all patients (>97%) are able to achieve a sustained virological response (SVR) regardless of HCV genotype or disease stage [4]. The Japan Society of Hepatology recommends that patients with HCV infection are treated with DAA therapies unless contraindications exist [5]. DAA therapies have led to the eradication of HCV in many patients [6, 7], but the risk of HCC has not been eliminated, thereby significantly impacting their survival [8, 9]. This raises various clinical questions about the management and follow-up of HCV-cured patients, particularly regarding the need for HCC surveillance and follow-up with a specialist. In Japan, most patients who achieved SVR have no history of HCC and tend to be older, and these patients are regularly followed up at hospitals or clinics for extended periods of time after achieving SVR. However, implementing continuous regular HCC surveillance in all patients would be a psychological burden and not cost-effective [10]. Therefore, there is great clinical value in identifying populations with a high risk for developing HCC and developing surveillance parameters for their follow-up. This study comprehensively assessed the factors associated with the incidence of HCC among patients who achieved SVR after DAA therapy for HCV infection. We also suggest guidelines for long-term follow-up in these patients to identify those with a high risk of developing HCC, especially among those without previous HCC treatment. Methods Patients This retrospective study collected data from patients with DAA-treated chronic HCV infection between September 2014 and December 2023 at the Kyoto Prefectural University of Medicine and six affiliated hospitals in the Kansai region of Japan (Kyoto, Osaka, Nara, and Shiga Prefecture). The baseline characteristics of the patients were collected before initiating DAA therapy (Table 1). Patients who achieved SVR were defined as those with an undetectable HCV RNA level 24 weeks after finishing treatment. Patients with other active liver diseases, including HIV co-infection, active hepatitis B infection, Wilson’s disease, or a history of liver transplantation, were excluded. Cases that did not achieve SVR or cases with major protocol deviations were excluded from the study (Fig. 1). The Fib-4 index was calculated using the following formula: age (years) × AST (U/L)/[platelet count (109/L) × ALT [(U/L)](1/2)] [11, 12]. Patients taking diabetic medications, a fasting plasma glucose ≥ 126 mg/dL and HbA1c ≥ 6.5%, were considered to have diabetes mellitus (DM). Excessive alcohol consumption was defined as >60 g/day for men and >50 g/day for women with reference on the Metabolic dysfunction-associated steatotic liver disease to intake alcohol levels at the initiation of therapies [13]. Table 1Overall baseline characteristics of 2,129 DAA-treated patients who achieved SVRParametersNumber or medianIQR or %Sex, male/female, n (%)913/1216(42.9%/57.1%)Age, years, median67.0(57.5–75.0)Previous interferon-based therapy, n (%)571(26.8%)Previous HCC treatment, n (%)223(10.5%)Genotypes (1a, 1b, type1, 2a, 2b, type2, 3a, mix, other), n (%)14/1396/120/358/179/40/5/9/80.7/65.6/5.6/16.8/8.4/1.9/0.2/0.4/0.4Regimen of DAAs (DA, DAB, EG, GP, OP, SL, SR, SV, other), n (%)370/1/84/424/135/739/357/16/317.4/0.0/3.9/19.9/6.3/34.7/16.8/0.8/0.2Excess alcohol use, n (%)140(6.6%)Diabetes mellitus, n (%)274(12.9%)AST (IU/L)39.0(27.0–59.0)ALT (IU/L)36.0(23.0–62.0)γ-GTP (IU/L)30.0(19.0–53.0)Albumin (g/dL)4.0(3.8–4.3)Bilirubin (mg/dL)0.72(0.60–1.00)Platelet count (×104/µL)16.0(11.7–20.6)Fib-4 index2.83(1.74–4.57)Fib-4 index > 3.25900(42.3%)AFP (ng/mL)4.70(3.00–9.00)Data are presented as numbers with percentages or median with IQR in parenthesesDAAs Direct-acting antivirals, SVR Sustained virological response, IQR Interquartile range, HCC Hepatocellular carcinoma, DA daclatasvir + asunaprevir, DAB Daclatasvir + asunaprevir + beclabuvir, EG Elbasvir + grazoprevir, GP Glecaprevir + pibrentasvir, OP Ombitasvir + paritaprevir + ritonavir, SL Sofosbuvir + ledipasvir, SR Sofosbuvir + ribavirin, SV Sofosbuvir + velpatasvir, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 Fibrosis-4, AFP Alpha-fetoprotein Overall baseline characteristics of 2,129 DAA-treated patients who achieved SVR Data are presented as numbers with percentages or median with IQR in parentheses DAAs Direct-acting antivirals, SVR Sustained virological response, IQR Interquartile range, HCC Hepatocellular carcinoma, DA daclatasvir + asunaprevir, DAB Daclatasvir + asunaprevir + beclabuvir, EG Elbasvir + grazoprevir, GP Glecaprevir + pibrentasvir, OP Ombitasvir + paritaprevir + ritonavir, SL Sofosbuvir + ledipasvir, SR Sofosbuvir + ribavirin, SV Sofosbuvir + velpatasvir, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 Fibrosis-4, AFP Alpha-fetoprotein Flowchart of patient selection. SVR, sustained virological response; HCC, hepatocellular carcinoma DAA therapy was initiated following the guidelines developed by board-certified hepatologists from the Japan Society of Hepatology for the management of HCV infection [5], specifically in patients without active or untreated HCC or life-threatening comorbidities. The study protocol was approved by the ethics committee of each institution and adhered to the provisions of the Declaration of Helsinki. This cohort of DAA-treated patients who achieved SVR was followed for 8 years, until the end of December 2023, or until the occurrence of HCC, whichever came first. Cases of HCC diagnosed within the initial 5-year period were not included in the cohort analyzed beyond the 5-year landmark study. If patients were lost to follow-up, the final visit was defined as the endpoint. Patients who died during the observation period were censored at the time of death. The patient characteristics before DAA therapy were collected as baseline data. In patients who received multiple DAA therapies, the starting point was defined by the end of the final treatment. HCC surveillance Abdominal imaging examinations and blood tests, including tumor marker tests, were performed at the start of DAA therapy, upon achieving SVR, and at least every 6–12 months thereafter for HCC surveillance. A hepatologist diagnosed the presence of HCC via ultrasonography, contrast-enhanced computed tomography, magnetic resonance imaging, contrast-enhanced ultrasonography with perflubutane, and angiography. If imaging modalities revealed nontypical imaging findings, the final diagnosis of HCC was made via biopsied specimens interpreted by an expert pathologist. Statistical analysis Cumulative incidence of HCC was calculated according to the Kaplan–Meier method and compared between groups by the log-rank test. After assessing multicollinearity, the clinical variables that could potentially influence the incidence of HCC in patients with SVR were analyzed via univariate and multivariate Cox’s proportional hazards regression analysis. Variables with a p-value of < 0.05 were considered for inclusion in the multivariate analysis. Receiver operating characteristic (ROC) curves and area under the curve (AUC) were used to test the predictive power of each variable, while the Youden index was used to determine the optimal cut-off values. Using simplified score transformation, a predictive scoring system was developed from the significant factors identified on multivariate Cox’s proportional hazard regression model. Each item was assigned a specific score derived from the logistic regression coefficient of the multivariable model. The regression coefficient of each item was divided by the lowest coefficient then rounded up to the nearest integer. This scoring system predicted the incidence of HCC later than 5 years after SVR among those without a previous HCC treatment. Statistical analysis was performed using IBM SPSS Statistics 26.0 and R. Results Patient characteristics, HCC incidence rate, and risk factors in the total group After applying the exclusion criteria, the study population included 2,129 patients who achieved SVR (Fig. 1). This cohort included a total of 10,379.0 person-years (PY) and had a median follow-up duration of 5.25 (interquartile range [IQR] 2.67–7.33) years after SVR. The patient characteristics are summarized in Table 1. During the follow-up period, 267 (12.5%) patients developed HCC, while 900 patients (42.3%) were expected to have advanced fibrosis (Fib-4 index >3.25) [11, 12]. In this cohort, 223 patients (10.5%) had previous HCC treatment before starting DAA therapy. Table 2 presents the results of the Cox regression analyses regarding the incidence of HCC in the overall cohort. Multivariable analyses revealed the following independent predictors of HCC: male sex, older age, previous HCC treatment before starting DAA therapy, excess alcohol use, DM, elevated γ-GTP, low albumin, and low platelet count (Table 2). Based on hazard ratio, having a history of HCC treatment was most significantly associated with the cumulative incidence rate of HCC. Table 2Risk factors for HCC occurrence by Cox multivariate model in all patientsUnivariate analysisMultivariate analysisParametersNo HCCHCCHR (95% CI)P valueAdjusted HR (95% CI)P valueTotal number, n1862267Sex, male/female768/1094145/1221.78 (1.40–2.27)< 0.0011.62 (1.26–2.09)< 0.001Age, years66.0 (56.0–74.0)73.0 (67.0–78.0)1.06 (1.04–1.07)< 0.0011.03 (1.02–1.05)< 0.001Previous IFN-based therapy472 (25.3%)99 (37.1%)1.42 (1.11–1.83)0.005Previous HCC treatment93 (5.0%)130 (48.7%)11.97 (9.39–15.2)< 0.0016.03 (3.74–7.88)< 0.001Excess alcohol use117 (6.3%)23 (8.6%)1.61 (1.05–2.46)0.0301.76 (1.09–2.85)0.021Diabetes mellitus201 (10.8%)73 (27.3%)2.718 (2.076–3.558)< 0.0011.52 (1.15–2.00)0.003AST (IU/L)37.0 (26.0–57.0)51.0 (38.0–69.0)1.005 (1.003–1.007)< 0.001ALT (IU/L)35.0 (23.0–61.0)44.0 (29.0–65.0)1.002 (1.001–1.003)0.370γ-GTP (IU/L)29.0 (19.0–51.0)39.0 (25.0–69.0)1.002 (1.001–1.003)0.0011.001 (1.000–1.003)< 0.001Albumin (g/dL)4.10 (3.80–4.30)3.70 (3.40–4.00)0.22 (0.17–0.27)< 0.0010.46 (0.35–0.61)< 0.001Bilirubin (mg/dL)0.70 (0.60–0.94)0.80 (0.61–1.13)2.28 (1.80–2.88)< 0.001Platelet count (×104/µL)16.7 (12.5–21.3)11.2 (8.60–14.6)0.87 (0.85–0.89)< 0.0010.93 (0.91–0.96)< 0.001Fib4 index2.58 (1.64–4.11)5.04 (3.69–7.46)1.12 (1.10–1.14)< 0.001AFP (ng/mL)4.30 (2.90–7.83)10.0 (5.20–19.3)1.008 (1.006–1.010)< 0.001HCC Hepatocellular carcinoma, HR Hazard ratio, CI confidence interval, IFN Interferon, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 fibrosis-4, AFP Alpha-fetoprotein Risk factors for HCC occurrence by Cox multivariate model in all patients HCC Hepatocellular carcinoma, HR Hazard ratio, CI confidence interval, IFN Interferon, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 fibrosis-4, AFP Alpha-fetoprotein Annual incidence of HCC with and without previous HCC treatment before DAA therapy In the Cox regression analyses of the overall cohort, the presence of previous HCC treatment before starting DAA therapy strongly influenced the incidence of HCC. The incidence rate of HCC was markedly higher among patients with a history of HCC treatment (17.3 per 100 PY [95% CI 14.5–20.5]) versus those without a history of HCC treatment (1.4 per 100 PY [95% CI 1.2–1.7]), with a 12.2-fold incidence rate ratio (95% CI. 9.5–15.6, p < 0.01). Furthermore, approximately 70% of patients with a history of HCC treatment developed HCC recurrence within 8 years after achieving SVR (Fig. 2). Cumulative incidence of hepatocellular carcinoma. Cumulative rate of hepatocellular carcinoma (HCC) in patients with and without previous HCC treatment. Dotted line, patients with previous HCC treatment before direct-acting antiviral (DAA) therapy. Continuous line, patients without a history of HCC treatment. The log-rank test revealed a significant difference between those with versus those without a history of HCC treatment (p < 0.001). (each 95%CI, with a history of HCC;3.842–4.684, without a history or HCC;7.503–7.646) The annual incidence of HCC was then plotted among patients with and without previous HCC treatment. Among patients with previous HCC treatment, the incidence of HCC was generally decreased. Meanwhile, among patients without previous HCC treatment, the incidence of HCC was lowest in the fifth year of follow-up, and this was generally maintained throughout the observation period (Fig. 3). Table 3 and 4 show the results of the Cox regression analysis regarding the occurrence of HCC among patients with and without a history of HCC treatment. High aspartate aminotransferase and low albumin levels remained independent predictors of HCC among patients with a history of HCC treatment (Table 3). Meanwhile, male sex, older age, excess alcohol use, DM, low albumin, and low platelet count remained independent predictors of HCC among patients without a history of HCC treatment (Table 4). In multivariable analyses, neither the era of DAA regimen nor HCV genotype was independently associated with the risk of HCC after adjustment for other covariates. These factors did not significantly influence the outcomes and were therefore not retained in the final predictive model. Annual incidence of hepatocellular carcinoma after achieving a sustained virological response. The graph depicts the incidence per 100 person-years for each year after achieving a sustained virological response across patients with and without previous hepatocellular carcinoma treatment before direct-acting antiviral therapy Table 3HCC risk factors by Cox multivariate model with previous HCC treatment before DAA therapyUnivariate analysisMultivariate analysisParametersNo HCCHCCHR (95% CI)P valueAdjusted HR (95% CI)P valueTotal number, n93130Sex, male/female47/4669/611.13 (0.80–1.60)0.49Age, years75.0 (69.0–80.0)74.0 (68.0–78.3)1.00 (0.98–1.02)0.71Previous IFN-based therapy34 (36.6%)47 (36.2%)0.86 (0.60–1.24)0.42Excess alcohol use6 (6.5%)3 (2.3%)0.42 (0.13–1.31)0.13Diabetes mellitus25 (26.9%)35 (26.9%)1.07 (0.72–1.57)0.75AST (IU/L)42.0 (26.5–57.0)49.0 (36.8–71.3)1.01 (1.00–1.01)0.0071.01 (1.00–1.01)0.04ALT (IU/L)35.0 (20.0–51.0)46.5 (28.0–63.0)1.00 (1.00–1.01)0.13γ-GTP (IU/L)31.0 (19.0–48.0)36.5 (23.8–62.8)1.00 (1.00–1.01)0.09Albumin (g/dL)3.90 (3.70–4.10)3.70 (3.30–4.00)0.49 (0.34–0.70)< 0.0010.51 (0.36–0.74)< 0.001Bilirubin (mg/dL)0.78 (0.60–1.00)0.80 (0.61–1.13)1.33 (0.93–1.90)0.13Platelet count (×104/µL)11.9 (10.3–15.0)10.9 (8.6–14.9)0.98 (0.94–1.01)0.17Fib-4 index4.31 (3.04–6.50)5.10 (3.69–7.60)1.06 (1.02–1.10)0.007AFP (ng/mL)5.8 (3.6–12.0)10.1 (5.38–20.3)1.00 (1.00–1.01)0.226HCC Hepatocellular carcinoma, DAA Direct-acting antiviral, HR Hazard ratio, CI Confidence interval, IFN Interferon, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 Fibrosis-4, AFP Alpha-fetoprotein HCC risk factors by Cox multivariate model with previous HCC treatment before DAA therapy HCC Hepatocellular carcinoma, DAA Direct-acting antiviral, HR Hazard ratio, CI Confidence interval, IFN Interferon, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 Fibrosis-4, AFP Alpha-fetoprotein Table 4HCC risk factors by Cox multivariate model without previous HCC treatment before DAA therapyUnivariate analysisMultivariate analysisParametersNo HCCHCCHR (95% CI)P valueAdjusted HR (95% CI)P valueTotal number, n1769137Sex, male/female721/104876/611.96 (1.40–2.78)< 0.0012.02 (1.41–2.90)< 0.001Age, years66.0 (55.0–74.0)71.0 (66.0–77.0)1.048 (1.03–1.07)< 0.0011.05 (1.03–1.08)< 0.001Previous IFN-based therapy438 (24.8%)52 (38.0%)1.49 (1.06–2.11)0.024Excess alcohol use111 (6.3%)20 (14.6%)3.05 (1.90–4.91)< 0.0013.27 (1.90–5.63)< 0.001Diabetes mellitus176 (9.9%)34 (27.7%)3.12 (2.15–4.54)< 0.0012.21 (1.50–3.24)< 0.001AST (IU/L)37.0 (26.0–57.0)53.0 (38.5–68.0)1.01 (1.00–1.01)< 0.001ALT (IU/L)34.0 (23.0–61.5)45.0 (30.0–65.0)1.00 (1.00–1.01)0.14γ-GTP (IU/L)29.0 (19.0–52.0)42.0 (26.0–73.50)1.00 (1.00–1.00)< 0.001Albumin (g/dL)4.10 (3.80–4.30)3.80 (3.40–4.10)0.22 (0.16–0.31)< 0.0010.48 (0.33–0.70)< 0.001Bilirubin (mg/dL)0.70 (0.60–0.94)0.80 (0.610–1.140)2.66 (1.93–3.67)< 0.001Platelet count (×104/µL)16.9 (13.0–21.45)11.6 (8.40–14.20)0.86 (0.83–0.88)< 0.0010.89 (0.86–0.92)< 0.001Fib-4 index2.52 (1.60–3.92)4.99 (3.64–7.35)1.12 (1.10–1.15)< 0.001AFP (ng/mL)4.2 (2.9–7.6)9.5 (5.1–19.0)1.01 (1.01–1.01)< 0.001HCC Hepatocellular carcinoma, DAA Direct-acting antiviral, HR Hazard ratio, CI Confidence interval, IFN Interferon, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 fibrosis-4, AFP Alpha-fetoprotein HCC risk factors by Cox multivariate model without previous HCC treatment before DAA therapy HCC Hepatocellular carcinoma, DAA Direct-acting antiviral, HR Hazard ratio, CI Confidence interval, IFN Interferon, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 fibrosis-4, AFP Alpha-fetoprotein Examination of factors influencing the incidence of HCC later than 5 years after achieving SVR Many patients in the overall cohort did not have a history of previous HCC treatment (n = 223, with previous HCC treatment; n = 1906, without previous HCC treatment). The frequency and factors contributing to the incidence of HCC after a long period of SVR are of great clinical interest. Since a 5-year follow-up is customarily used in clinical practice, we set landmark the 5th year after achieving SVR as the point of interest in this study. We then examined the factors influencing the incidence of HCC (de-nove HCC) later than 5 years after achieving SVR using laboratory data taken at the beginning of treatment and at the 5th year of SVR. Univariate Cox regression analysis using data at the start of treatment revealed that many factors influenced the incidence of HCC later than 5 years after achieving SVR (Table 5). ROC curves demonstrated the high predictive value of the Fib-4 index and α-fetoprotein as continuous variables (Fig. 4a). Multivariate Cox regression analysis identified age, DM, γ-GTP, and platelet count as independent predictors of HCC. Using these independent factors, we initially attempted to predict the incidence of HCC later than 5 years after SVR based on data at the start of treatment. However, the predictive score of these factors was not beyond that of the continuous Fib-4 index, which was not an independent factor (Fig. 4a). Meanwhile, univariate Cox regression analysis using data taken at the 5th year of SVR revealed age, DM, platelet count, and Fib-4 index as significant factors (Table 6). On multivariate Cox regression analysis, age, DM, and Fib-4 index measured at the 5th year of SVR remained independent predictors of the incidence of HCC later than 5 years after SVR. In constructing the multivariable predictive model, we considered the potential multicollinearity between platelet count and the FIB-4 index, as platelet count is one of the components of FIB-4 index. Comparative analysis revealed that the model including age, DM), and FIB-4 index (excluding platelet count) showed improved stability and predictive accuracy. Therefore, platelet count was excluded from the final model to avoid redundancy and enhance model robustness. On ROC analysis the optimal cut-offs were age ≥ 70 years at the 5th year((sensitivity:89.2%, and specificity:46.6%) and Fib-4 index ≥ 2.67 at the 5th year (sensitivity:70.3%, and specificity:70.2%). Therefore, age ≥ 70 years, the presence of DM and Fib-4 index ≥ 2.67 were selected as potential predictors. Each potential predictor on the multivariable model was assigned a specific score derived from the multivariate Cox’s proportional hazard regression coefficient. The coefficients derived via Cox’s proportional hazards models were 3.715 (95% CI: 1.413–9.767, p = 0.008) for age ≥ 70, 3.041 (95% CI: 1.471–6.289, p = 0.003) for DM, and 3.638 (95% CI: 1.767–7.490, p < 0.001) for Fib-4 index ≥ 2.67. The regression coefficient of each item was divided by the lowest coefficient, then rounded up to the nearest integer to develop the HCC incidence prediction scoring. All items were assigned to have a score of 1, since all variables were deemed to have an equal impact on the dependent variable. Age ≥ 70 years, DM, and Fib-4 index ≥ 2.67 provided the best discrimination value based on AUC compared to the other single continuous variables (Fig. 4b). We further evaluated a 4-item score including platelet count along with age, diabetes mellitus, and FIB-4 index. However, the addition of platelet count did not improve the model’s discrimination (AUC 0.751 vs. 0.765 for the original 3-item score), nor did it enhance sensitivity. Therefore, we opted to retain the simpler 3-item model for its comparable predictive performance and clinical practicality. Finally, the Kaplan–Meier curve was used to describe the incidence of HCC later than 5 years after SVR using the above score (Fig. 5). The cumulative incidence rates of HCC after 5 years were 0.8%, 2.7%, 11.9%, and 28.8% for patients with scores of 0, 1, 2, and 3, respectively. A score of 0 indicates a very low incidence rate of HCC among patients treated with DAA therapy later than 5 years after SVR. Table 5HCC risk factors at 5 years or later after achieving SVR by Cox multivariate model. Examination with factors at the beginning of DAA therapyUnivariate analysisMultivariate analysisParametersNo HCCHCCHR (95% CI)P valueAdjusted HR (95% CI)P valueTotal number, n101037Sex, male/female370/64017/201.446 (0.757–2.761)0.264Age, years66.0 (57.0–73.0)70.0 (67.0–77.0)1.071 (1.031–1.112)< 0.0011.07 (1.03–1.11)0.001Previous IFN-based therapy302 (29.9%)13 (35.1%)1.071 (0.545–2.105)0.842Excess alcohol use41 (4.1%)4 (10.8%)2.740 (0.970–7.739)0.057Diabetes mellitus100 (9.9%)10 (27.0%)3.284 (1.589–6.786)0.0013.00 (1.45–6.21)0.003AST (IU/L)37.0 (27.0–56.0)51.0 (38.5–72.5)1.007 (1.002–1.012)0.010ALT (IU/L)34.0 (23.00–56.00)46.0 (29.5–62.0)1.002 (0.996–1.009)0.471γ-GTP (IU/L)27.0 (18.0–46.3)42.0 (23.0–100.0)1.004 (1.001–1.008)0.0071.01 (1.00–1.01)0.002Albumin (g/dL)4.1 (3.9–4.3)3.9 (3.6–4.1)0.289 (0.142–0.587)0.001Bilirubin (mg/dL)0.70 (0.60–0.92)0.70 (0.6–1.2)2.896 (1.345–6.234)0.007Platelet count (×104/uL)16.8 (12.9–21.0)12.0 (7.9–15.3)0.859 (0.806–0.915)< 0.0010.87 (0.82–0.93)< 0.001Fib-4 index2.49 (1.68–3.81)4.99 (3.97–7.29)1.128 (1.086–1.171)< 0.001AFP (ng/mL)4.4 (3.0–7.9)9.0 (6.1–15.3)1.010 (1.003–1.016)0.004HCC Hepatocellular carcinoma, SVR Sustained virological response, DAA Direct-acting antiviral, HR Hazard ratio, CI Confidence interval, IFN Interferon, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 Fibrosis-4, AFP Alpha-fetoprotein HCC risk factors at 5 years or later after achieving SVR by Cox multivariate model. Examination with factors at the beginning of DAA therapy HCC Hepatocellular carcinoma, SVR Sustained virological response, DAA Direct-acting antiviral, HR Hazard ratio, CI Confidence interval, IFN Interferon, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 Fibrosis-4, AFP Alpha-fetoprotein Receiver operating characteristics (ROC) curves of continuous variables for the incidence of hepatocellular carcinoma later than 5 years after achieving SVR. a ROC curves using factors at the beginning of treatment of age, AST, γ-GTP, Albumin, Bilirubin, platelet count, Fib-4 index, and AFP. These factors are significant differences in univariate analysis of COX regression analysis. In addition, dotted line of ROC curve of score made from age, diabetes mellitus, γ-GTP and platelet count, which are independent factors at the beginning of direct-acting antiviral therapy, is shown.; (b) ROC curves using factors at the 5th year after achieving a sustained virological response of age, Fib-4 index and platelet count. These factors are significant differences in univariate analysis of COX regression analysis. In addition, dotted line of ROC curve of Score made from age, diabetes mellitus, and Fib-4 index, which are independent factors at the 5th year after achieving a sustained virological response, is shown Table 6HCC risk factors at 5 years or later after achieving SVR by Cox multivariate model. Examination with factors at the 5th year after achieving SVRUnivariate analysisMultivariate analysisParametersNo HCCHCCHR (95% CI)P valueAdjusted HR (95% CI)P valueTotal number, n101037Sex, male/female374/64817/201.45 (0.76–2.76)0.26Age, years66.0 (57.0–73.0)70.0 (67.0–77.0)1.07 (1.03–1.11)< 0.0011.06 (1.02–1.10)0.006Previous IFN-based therapy307 (30.0%)13 (35.1%)1.07 (0.55–2.11)0.84Excess alcohol use41 (4.1%)4 (10.8%)2.74 (0.97–7.74)0.06Diabetes mellitus101 (9.9%)10 (27.0%)3.28 (1.59–6.79)0.0013.11 (1.50–6.45)0.002AST at 5th year (IU/L)21.0 (18.0–26.0)24.0 (20.5–27.5)1.02 (0.99–1.05)0.24ALT at 5th year (IU/L)15.0 (11.0–20.0)18.0 (12.0–23.0)1.00 (0.97–1.03)0.92γ-GTP at 5th year (IU/L)18.0 (14.0–27.0)27.0 (14.0–54.0)1.00 (1.00–1.01)0.30Albumin at 5th year (g/dL)4.3 (4.1–4.5)4.2 (4.0–4.4)0.58 (0.24–1.36)0.21Bilirubin at 5th year (mg/dL)0.70 (0.70–1.00)0.85 (0.60–1.1)1.96 (0.97–1.52)0.35Platelet count at 5th year (×104/uL)19.1 (15.3–22.7)15.1 (11.8–19.4)0.89 (0.84–0.94)< 0.001Fib-4 index at 5th year2.06 (1.51–2.82)2.83 (2.21–4.72)1.24 (1.14–1.35)< 0.0011.21 (1.10–1.33)< 0.001AFP (ng/mL) at 5th year2.9 (2.0–4.0)3.6 (2.8–4.7)1.11 (0.98–1.21)0.06SVR Sustained virological response, HCC Hepatocellular carcinoma, HR Hazard ratio, CI Confidence interval, IFN Interferon, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 Fibrosis-4, AFP Alpha-fetoprotein HCC risk factors at 5 years or later after achieving SVR by Cox multivariate model. Examination with factors at the 5th year after achieving SVR SVR Sustained virological response, HCC Hepatocellular carcinoma, HR Hazard ratio, CI Confidence interval, IFN Interferon, AST Alanine aminotransferase, ALT Alanine aminotransferase, γ-GTP Gamma-glutamyltransferase, Fib-4 Fibrosis-4, AFP Alpha-fetoprotein Cumulative incidence of hepatocellular carcinoma after achieving a sustained virological response This analysis was done in the cohort of patients without previous hepatocellular carcinoma treatment. The cumulative incidence rate of each score is illustrated on the right side of the graph. The log-rank test showed a significant difference (p < 0.001) Discussion Although DAA therapy has dramatically changed the treatment of chronic hepatitis C, data regarding the long-term incidence of HCC after achieving SVR remains limited. This study explored the factors contributing to the incidence of HCC after achieving SVR. Male sex, older age, previous HCC treatment, excess alcohol use, DM, elevated γ-GTP, low albumin, and low platelet count were identified independent risk factors throughout the entire observation period. These risk factors closely resemble those linked to the incidence of HCC after previous interferon (IFN)-based therapies [14, 15]. In this study, previous HCC treatment before DAA therapy was the strongest predictor of HCC after achieving SVR. Tumor recurrence after treatment with HCC is common even if patients are clinically considered to be cured of HCC [16]. Since the cumulative incidence of HCC recurrence until 8 years after SVR was about 70% in patients with previous HCC treatment, the predictors shown in Table 3 may not be very accurate in predicting the incidence of HCC. Although several references [17, 18] have reported that DAA treatment could reduce the risk of recurrence after curative treatment, HCC surveillance remains crucial for patients with a history of HCC treatment. Since most patients who achieved SVR do not have a history of HCC treatment before DAA therapy, we additionally examined the incidence and risk factors of HCC in this group. In this cohort, the annual incidence of HCC was high (1.4 per 100 PY) comparable to those with previous IFN treatments [19]. Factors such as male sex, older age, excessive alcohol consumption, DM, low albumin levels, and low platelet counts were confirmed as independent risk factors across the entire observation period. The information regarding these risk factors can help both patients and practitioners predict the actual risk of HCC following virologic cure. The frequency and factors contributing to the incidence of HCC after a long period of SVR are of clinical interest. Many patients did not have a history of HCC treatment before DAA therapy. Thus, we focused on creating a guide which involves a longer follow-up period, especially since only a 5-year follow-up is customarily used in clinical practice. In our study, the annual incidence of HCC decreased toward the 5th year then rebounded slightly afterward among those without previous HCC treatment, demonstrating an inflection point in the 5th year. Additionally, clinical factors, including Fib-4 index [20], also changed over time during the follow-up period. Thus, it would be clinically valuable to reevaluate whether follow-up should be continued even 5 years after SVR. Thus, we examined the factors influencing the incidence of HCC later than 5 years after SVR, using data at start of DAA therapy and data at the5th year of follow-up. We created a scoring model to predict the incidence of HCC later than 5 years after SVR. Although age, DM, γ-GTP and platelet count at the start of DAA therapy were independent risk factors, we were unable to create a model that had better predictive ability than the Fib-4 index alone. Meanwhile, age, DM and high Fib-4 index at the 5th year of follow-up were independent risk factors. After determining optimal cut-off values for age and Fib-4 index, we were able to create a model that could predict and stratify the incidence of HCC at least 5 years after SVR. Patients under 70 years of age, without DM, and with a Fib-4 index less than 2.67 had a sufficiently low incidence of HCC later than 5 years after SVR, and in our cohort, these included about 1/3 of the patients. To assess the appropriateness of considering these three variables as having equivalent clinical impact, we compared the predictive performance between a weighted score based on the regression coefficients and that of the simplified equal-weight model using the concordance index and AUC. No significant difference was observed between the two models (concordance-index: 0.777 vs. 0.773, p = 0.672). We consider that this simplified model suggests sufficient predictive accuracy. Moreover, we compared the predictive performance of our proposed score with other established HCC risk models, including the Watanabe et al. score [21] and ADRES scores [22], among the various scoring systems that have been developed to assess the risk of hepatocarcinogenesis following SVR, using data obtained 5 years after SVR. The AUCs for later than 5years after SVR of HCC incidence prediction were 0.766 for our score, 0.730 for Watanabe et al. score, 0.608 for ADRES score. DeLong test showed the differences between our score and the Watanabe et al. score was p = 0.543, and the differences between our score and ADRES score were p = 0.038. Our scoring system showed predictive ability comparable to, or possibly exceeding, that of previously these two established scores. Since both the Watanabe et al. score and ADRES scores were originally developed using parameters obtained before DAA therapy or at SVR12, their predictive ability may be limited for long-term (>5 years) HCC risk estimation. In contrast, our model, which incorporates clinical data obtained at 5 years after SVR, may provide better discrimination for late HCC development. Nevertheless, the potential overfitting should be considered when interpreting these results. Recent studies suggest initiating surveillance among patients with cirrhosis if they cross the HCC incidence threshold of approximately 1.0% [6, 23]. Thus, although patients identified as very low-risk group by this scoring system should not immediately discontinue surveillance, certain populations may not require prolonged, intensive follow-up depending on their findings after the 5th year of SVR. Meanwhile, approximately 1/3 of the patients in this study had scores of 2 or 3, which indicate a non-negligible frequency of carcinogenesis even after 5 years of SVR, and thus, continuous surveillance may be necessary. The identified risk factors, including age, Fib-4 index, and DM, likely represent the residual oncogenic potential of the liver even after viral eradication. Although achievement of SVR leads to improvement in inflammation and partial regression of fibrosis, these processes are highly heterogeneous, and fibrosis regression may be incomplete in some patients. Moreover, metabolic alterations resulting from DM or alcohol consumption can persist or even worsen after viral clearance, contributing to the incidence of HCC. Model based on 5-year post-SVR data may therefore capture these dynamic changes in hepatic and metabolic status more accurately than a model constructed using baseline data. However, time-dependent changes in variables such as DM status and alcohol consumption use could not be fully captured in this retrospective design. This remains an important area for future investigation. The Fib-4 index, which measures the degree of fibrosis, had relatively good accuracy for predicting the incidence of HCC when measured at the start of DAA therapy and also after 5 years of SVR. Since the Fib-4 index is a continuous variable, appropriate cut-off values need to be determined for clinical use. Interestingly, the threshold for Fib-4 index after 5 years of SVR closely corresponds to the cut-off value used to diagnose advanced fibrosis or other liver diseases such as metabolic dysfunction-associated steatohepatitis [24]. Although FIB-4 index values may fluctuate following SVR, patients presenting with elevated baseline FIB-4 index suggestive of cirrhosis, as well as those with persistently high FIB-4 index values during long-term follow-up, should be considered appropriate candidates for continued HCC surveillance. This approach aligns with current recommendations and reflects the importance of fibrosis severity in guiding post-SVR management strategies [25]. Moreover, it should be noted that the FIB-4 index includes age as one of its calculation components, which may lead to an overestimation of fibrosis in elderly patients. However, in the present analysis, both diabetes mellitus and age remained independent risk factors for hepatocellular carcinoma development beyond 5 years after SVR. Moreover, being aged 70 years or older was identified as a significant variable during the process of constructing the risk score. Therefore, the influence of age is considered to be, at least in part, inherently incorporated into the scoring system. As assessment of liver fibrosis, elastography was not routinely performed in this multicenter study because the majority of participating institutions did not have access to the equipment. Furthermore, serum fibrosis markers were not consistently measured at baseline. In addition, the APRI score at 5 year did not show superior predictive performance for the incidence of HCC later than 5 years after SVR compared with FIB-4 at (AUC; Fib-4 index 0.704 vs. APRI 0.474) These factors may have limited the comprehensive assessment of liver fibrosis in the present study. This study has several limitations. The study cohort was exclusively Japanese. Given the known ethnic and regional differences in HCV genotypes, host genetics, and health insurance system, the findings and the proposed scoring system may not be generalizable to other populations (e.g., North American or European). Further studies are required to validate these findings in other Japanese cohorts and in different populations abroad. As the data were extracted from medical records, it was difficult to incorporate variables such as total alcohol consumption, treatment regimens for DM, met the diagnostic criteria for MASLD, and other lifestyle-related factors into the analysis. Although alcohol is a well-established carcinogenic factor, its contribution to HCC risk may have been underestimated in our analysis. Moreover, the degree of hepatic steatosis was not uniformly evaluated and could not be quantitatively assessed in this multicenter setting. In addition, variability in the surveillance intervals in clinical practice may have introduced bias in the detection of HCC. Moreover, the retrospective nature of the study introduces potential biases. The presence of selection bias and dropout cases on follow-up, which are inevitable in clinical research, may have influenced our results. To account for the potential competing risk of death, we reanalyzed the data after excluding patients who had died. Given that there were only 16 death cases later than 5 years after achieving SVR without previous HCC treatment group, the reanalysis of cumulative HCC incidence were 0.8%, 2.3%, 10.0%, and 31.1% for patients with scores of 0, 1, 2, and 3, respectively, and these results were generally consistent with those presented in the results section. Clinicians may have rushed DAA therapy after HCC treatment without a sufficient interval to prevent the decline of hepatic functional reserve, or they may have imposed long-term follow-up of patients that were suspected to have a higher risk of HCC. Additionally, due to the presence of censored cases, the graph for examining the frequency of HCC would have larger error bars with longer follow-up. To ensure clinical relevance and consistency, our interpretation aligns with surveillance thresholds recommended in current clinical guidelines. However, we emphasize that further external validation is necessary before any modifications to existing surveillance strategies or clinical practice can be considered. Nevertheless, it remains clinically meaningful to create guidelines for long-term follow-up. Conclusions This study investigated the factors associated with the development of HCC after achieving SVR due to DAA therapy for patients with HCV. Previous HCC treatment was the strongest risk factor for the development of HCC after achieving SVR. Moreover, for patients without previous HCC treatment, we developed a scoring system to predict the incidence of HCC later than 5 years after SVR, which included age of ≥ 70 years, DM, and Fib-4 index of ≥ 2.67 at the 5th year. Among patients without previous HCC treatment, age of < 70 years, without DM, and Fib-4 index of < 2.67 were associated with a very low risk of HCC development possibly long after SVR. Supplementary Information Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Acknowledgements The authors would like to thank Enago (www.enago.jp) for the English language review. Abbreviations Area under the curve Direct-acting antiviral Diabetes mellitus Hepatocellular carcinoma Hepatitis C virus Person-years Receiver operating characteristic Sustained virological response Interferon Authors’ contributions Hideki Fujii, Kanji Yamaguchi, Takeshi Nishimura, Hiroyuki Kimura, Yasuhide Mitsumoto, Toshihide Shima, Yuya Seko, Michihisa Moriguchi, Seita Kataoka, Saiyu Tanaka, Hiroki Ishikawa, Shiro Takami, Yasuyuki Nagao, Atsushi Umemura, and Yoshito Itoh collected the data of the research subjects. Hideki Fujii analyzed and interpreted these data. Hideki Fujii and Yoshito Itoh were major contributors in writing the manuscript. All authors read and approved of the final manuscript. Funding No funding support was obtained. Data availability The approval of ethical committee of Kyoto Prefectural University of Medicine and each of the participating institutions in the study are needed when the data are requested. Declarations Ethics approval and consent to participate This study was conducted in accordance with the tenets of the Declaration of Helsinki. The ethical committee of Kyoto Prefectural University of Medicine and each of the participating institutions in the study approved the study. Informed consent was obtained at each facility via an opt-out method, as approved by the institutional review boards. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Data Availability Statement The approval of ethical committee of Kyoto Prefectural University of Medicine and each of the participating institutions in the study are needed when the data are requested.