Abstract
Background/Aims
Surgical resection and percutaneous ablation are considered curative treatments for hepatocellular carcinoma (HCC). However, both approaches are associated with a high risk of recurrence. Real-world data on outcomes and post-curative management remain scarce. This study aimed to assess survival and healthcare pathways of patients with HCC treated with resection or ablation in routine clinical practice.
Methods
We conducted a retrospective analysis using the French National Health Data System. Patients who underwent resection or percutaneous ablation as first-line treatment for HCC between January 2014 and December 2021 were included. The primary endpoint was overall survival (OS); secondary endpoints focused on subsequent treatment sequences and healthcare pathways
Results
Among 10,810 patients included, 5,488 underwent ablation and 5,322 resection as first-line treatment. One- and 2-year OS rates were 90% and 78% after ablation (median OS [mOS], 56 months; 95% confidence interval [CI], 54–58) and 88% and 80% after resection (mOS, 75 months; 95% CI, 72–79). Median time to next treatment or death was 20 months (95% CI, 18–21) after ablation and 29 months (95% CI, 27–30) after resection. Following initial treatment, 27% of patients in the ablation group and 39% in the resection group received no further therapy.
Conclusions
This nationwide real-world study provides comprehensive data on survival and treatment sequences after curative-intent therapy for HCC in France. Despite frequent recurrence, survival outcomes remained favourable, reflecting access to repeated and effective subsequent treatments, underscoring the importance of long-term follow-up in routine care.
Keywords: Carcinoma, hepatocellular; Fibrosis; Resection; Ablation; Prognosis
GRAPHICAL ABSTRACT
INTRODUCTION
Hepatocellular carcinoma (HCC) is a significant global public health concern, with an annual incidence reaching one million new cases worldwide.1 The rising prevalence of alcohol consumption and obesity has contributed to a 9% yearly increase in HCC cases, with metabolic dysfunction-associated fatty liver disease (MAFLD) now being the leading cause of disease.2,3 Approximately 80% of HCC cases develop in a cirrhotic liver, while a smaller percentage arise in chronic liver disease without cirrhosis.2
Screening programs enable the diagnosis of around 40% of HCC cases at a curative stage, allowing for potentially curative treatments such as resection, ablation, and liver transplantation.4 In patients with localized HCC (Barcelona Clinic Liver Cancer [BCLC] 0 or A), resection is preferred for patients with a single tumor, preserved liver function, and no significant portal hypertension, while ablation serves as the only curative alternative for non-surgical candidates, particularly prevalent in Europe due to the frequent coexistence of cirrhosis or comorbidities.5,6
However, recurrence rates remain high, with over 70% of patients experiencing recurrence within 5 years after resection or ablation, even among those deemed optimal candidates.7 The risk is even greater in patients with more advanced tumor burden at the time of curative-intent treatment and HCC is now the second leading cause of cancer-related deaths worldwide.8
HCC recurrence follows a bimodal distribution, with the highest risk of disease recurrence occurring within 1 to 2 years post curative-intent treatments.9-11 Early recurrence is often linked to micrometastases from the primary tumor, while late recurrence, peaking between 4 and 5 years, is associated with new tumor development due to underlying liver disease.12,13 Factors contributing to early disease recurrence include tumor size, multiplicity, vascular invasion, high histological grade, and elevated serum α-fetoprotein (AFP) levels. Early recurrence is generally more aggressive and leads to poorer long-term outcomes compared to late recurrence.
Adjuvant therapy has been explored as a means to prevent or delay disease recurrence. However, phase III clinical trials have yet to demonstrate a significant benefit in recurrence-free or overall survival, and current data are not sufficient to recommend adding adjuvant or neo-adjuvant treatment to a resection or ablation of HCC.6,14-16 Additionally, clinical trial data may not fully represent real-world patient populations, as trials often exclude specific patient groups and may not capture the heterogeneity of HCC in routine clinical practice. In France, the National Health Data System (Système National de Données de Santé, SNDS) offers a comprehensive database covering approximately 99% of the French population (over 66 million people). This database provides detailed records of outpatient and inpatient healthcare services, including physician visits, reimbursed medications, laboratory tests (without test results), medical procedures, and hospital stays. By integrating the hospital database (National Programme for Hospital Information System, PMSI), it captures critical diagnostic and procedural information, enabling a thorough analysis of healthcare utilization and disease management. In cases of severe chronic conditions, such as HCC, general practitioners can request full coverage for patients under the long-term disease (LTD) scheme.
This study aimed to provide information on the epidemiology, management and outcomes of patients with localized HCC who underwent percutaneous ablation or surgical resection.
MATERIALS AND METHODS
Study design
The LIVERPOOL study was a retrospective study of historical data constituted from the SNDS, including patients who underwent a liver resection or radiofrequency ablation as a first-line curative treatment for HCC. The identification of hospitalizations for HCC, HCC-related procedures and specific treatments were searched on the overall study period (2012–2021) in order to identify the disease onset.
The study was performed in accordance with the authorization DR-2024-010 of French Data Protection Authority (CNIL) allowing an exception to the obligation to provide information to the subjects of the study.17 Nevertheless, it was registered on the transparency portal of the Health Data Hub. To exercise their rights the persons concerned shall send their request to the director of the Health Data Platform or through the director of the management organisation of the mandatory health insurance scheme to which the data subject is affiliated, in accordance with the provisions of article R.1461-9 of the French Public Health Code.
Study population
Eligible patients were included at the first date of admission to hospital for a curative liver resection or radiofrequency ablation for HCC over the period from January 1st 2014 to December 31st 2021 and followed up until death or December 31st 2021. A minimum 2-year lookback period was considered (i.e., from January 1st 2012 to January 1st 2014 for a patient included at January 1st 2014) to ensure the absence of any previous HCC treatment and any previous marker of HCC evidence or management.
Patients were 18 years or older, with a primary liver cancer as confirmed by an LTD status or a diagnosis of malignant neoplasm of liver (ICD-10 code C22.0) between 2012 and 2021, or a diagnosis of malignant neoplasm of liver during a hospital stay between 2012 and 2021 (main diagnosis or secondary diagnosis), and treated by curative-intent liver resection or radiofrequency ablation over the period 2014 to 2021.
Patients were excluded if they had received a transarterial chemoembolization (TACE) or transarterial radioembolization (TARE) or systemic treatment over the lookback period before the first curative resection or ablation treatment (index date), if they had another primitive cancer diagnosed after HCC and without further hospital stay with HCC diagnosis, or if they had another record of a malignant tumor of the liver and biliary tract or malignant tumor of the liver without precision (C22.9), carcinoma of the intrahepatic bile duct (C22.1), hepatoblastoma (C22.2), angiosarcoma of the liver (C22.3), other sarcomas of the liver (C22.4) or other specified carcinomas of the liver (C22.7) before the index date. Patients with another associated primitive cancer or markers of management for advanced or metastatic cancer prior to the first identification of HCC and patients who underwent ablation and resection on the same day were also excluded.
Variables and data collection
For all patients, we registered characteristics such as the age at inclusion, gender, residence area, HCC aetiologies for chronic liver disease (alcohol-related, metabolic-related or virus-related), cirrhosis (yes/no), healthcare pathway (hospitalizations including hospitalization for best supportive care [BSC] associated with a HCC), all HCC related procedures, systemic treatments for HCC by line of treatment (drugs, duration of treatment), and the medical staff organization (MSO) hospitalizations during the follow-up period.
Cirrhosis was identified using ICD-10 code K74 (fibrosis and cirrhosis of liver) recorded during hospitalizations in the SNDS database.
Several liver disease aetiologies could be identified for a single patient based on ICD-10 codes. For outcome analyses, patients were assigned to mutually exclusive aetiological categories according to a predefined hierarchy: viral aetiology>alcohol-related aetiology>metabolic-related aetiology.
However, since the database does not contain clinical information on HCC prognostic factors or disease staging, such as BCLC stage, liver function assessments (Child-Pugh score), tumor characteristics (number, size, vascular invasion, extrahepatic spread, AFP level), or patient performance status, these factors could not be included in the analyses.
The codes used for the analysis are available in Supplementary Tables 2–5.
Outcomes
The primary endpoint was overall survival (OS) of patients treated by curative liver resection or ablation for HCC, defined as the time from the first curative-intent resection or ablation identified in the study period until death from any cause, loss to follow-up or the end date of the study (31 December 2021), whichever occurred first. We also studied OS with respect to cirrhosis status before ablation or resection (yes/no) and different groups of disease etiologies.
Secondary endpoints were the description of patient characteristics at baseline, their healthcare pathway and treatment sequences after curative resection or ablation for HCC. For the purpose of treatment sequence description, only patients with a minimum of 4 years of follow-up were considered (i.e., patients included between 2014 and 2017).
We also estimated an exploratory endpoint, time to next treatment or death (TTNT-D), defined as the duration between the date of the first curative liver resection or ablation identified in the study period and the first TTNT-D, whichever occurred first. The first next treatment following ablation or resection was identified by any of the markers of the initiation of first-line systemic treatment, radiotherapy, TACE, TARE, hospitalisation for BSC associated with an HCC ICD-10 code, or another curative treatment including liver transplantation.
Since a second curative treatment shortly after the initial resection or ablation could indicate technical failure, early disease recurrence, or suboptimal initial indication, we performed a sensitivity analysis evaluating whether a second curative procedure occurring within 90 days of the first should be considered part of the initial treatment or the first next treatment. We also assessed the correlation between TTNT-D and OS.
Statistical analysis
The Kaplan-Meier method was used to estimate median OS (mOS), median TTNT-D (mTTNT-D), and survival rates with corresponding 95% confidence intervals (CI). Data were censored at the time of the last reimbursed care (ambulatory or hospital) recorded in the study period. Landmark survival with 95% CI was estimated at 6, 12 and 18 months after the date of first ablation or resection. All analyses were performed separately within the ablation and resection groups, without direct comparative statistical testing between treatment modalities.
The log-rank test was used to evaluate differences in OS between groups of disease aetiology. The reverse Kaplan-Meier method was used to estimate the median follow-up with its 95% CI.
Cox proportional hazards models were used to estimate hazard ratios (HR) for death and next treatment or death associated with cirrhosis, with corresponding 95% CI. Both unadjusted and adjusted analyses (baseline covariates or propensity score adjustment for cirrhosis status) were performed.
Restricted mean survival time (RMST) was estimated as the average survival over a predefined time horizon, and the difference in RMST between cirrhotic and non-cirrhotic patients was calculated to complement Cox model results in case of non-proportional hazards. The RMST truncation time was selected according to the maximum available follow-up with stable survival estimation (approximately 92 months for OS and 36 months for TTNT-D).
The correlation between OS and TTNT-D was assessed using a method proposed by Schemper et al.,17 equivalent to a Spearman-type correlation coefficient ranging from 0 (no association) to 1 (perfect association).
Patient characteristics were summarised as medians (interquartile range [IQR]) for continuous variables and as counts (percentages) for categorical data. The Pearson test was used to compare percentages; alternatively, Fisher’s exact test was used when chisquare assumptions were not satisfied. Kruskal-Wallis and Wilcoxon tests were also applied. Statistical significance was defined as P≤0.05.
Treatment sequences were described using Sankey diagrams with up to three lines of treatment following ablation or resection. All statistical analyses were performed using SAS® V9.4 software (SAS Institute, Cary, NC, USA).
RESULTS
Patients’ characteristics
A total of 21,764 patients were screened and among them 19,469 patients were selected for the study. After applying the exclusion criteria previously described (Fig. 1), 10,810 patients were included in the study: 5,488 patients who underwent ablation, and 5,322 who underwent resection. The majority of patients were male (81.9% for ablation, 80.1% for resection) with a median age of 68.0 (IQR, 61–74) years for both groups (Table 1). The most common underlying liver disease aetiology was excessive alcohol consumption (57.7% vs. 30.8% of patients had at least an excessive alcohol consumption), followed by metabolic disease (diabetes or obesity), hepatitis C infection, and non-alcoholic steatohepatitis (NASH). 83.4% of patients in the ablation group had cirrhosis, vs 49.8% in the resection group. The median follow-up was 2.7 years (IQR, 1.1–4.0) in the ablation group and 2.9 years in the resection group (IQR, 1.0–4.4).
Figure 1.
Flowchart of patient enrollment and exclusion criteria. HCC, hepatocellular carcinoma.
Table 1.
Patient characteristics and etiologies
| Ablated group (n=5,488) | Resected group (n=5,322) | Global (n=10,810) | |
|---|---|---|---|
| Year of inclusion | |||
| 2014 | 492 (9.0) | 528 (9.9) | 1,020 (9.4) |
| 2015 | 588 (10.7) | 625 (11.7) | 1,213 (11.2) |
| 2016 | 631 (11.5) | 690 (13.0) | 1,321 (12.2) |
| 2017 | 713 (13.0) | 666 (12.5) | 1,393 (12.8) |
| 2018 | 751 (13.7) | 651 (12.2) | 1,402 (12.3) |
| 2019 | 761 (13.9) | 714 (13.4) | 1,475 (13.6) |
| 2020 | 696 (12.7) | 674 (12.7) | 1,370 (12.7) |
| 2021 | 856 (15.6) | 774 (14.5) | 1,630 (15.1) |
| Gender | |||
| Male | 4,497 (81.9) | 4,263 (80.1) | 8,760 (81.0) |
| Female | 991 (18.1) | 1,059 (19.9) | 2,050 (19.0) |
| Age (years) at index date | |||
| Number (response rate) | 5,488 (100.0) | 5,322 (100.0) | 10,810 (100.0) |
| Mean [standard deviation] | 67.3 [−9.7] | 66.7 [10.9] | 67.0 [10.3] |
| Median/min/max | 68.0/18.0/96.0 | 68.0/18.0/94.0 | 68.0/18.0/96.0 |
| Quantile 25/quantile 75 | 61.0/74.0 | 61.0/74.0 | 61.0/74.0 |
| Etiologies identified (at least one) | |||
| Alcohol disease | 3,168 (57.7) | 1,637 (30.8) | 4,805 (44.4) |
| Hepatitis B | 373 (6.8) | 447 (8.4) | 820 (7.6) |
| Hepatitis C | 1,256 (22.9) | 914 (17.2) | 2,170 (20.1) |
| Other hepatitis | 135 (2.5) | 96 (1.8) | 231 (2.1) |
| Diabetes | 1,715 (31.3) | 1,368 (25.7) | 3,083 (28.5) |
| Obesity | 1,405 (25.6) | 1,336 (25.1) | 2,741 (25.4) |
| NASH | 1,008 (18.4) | 818 (15.4) | 1,826 (16.9) |
| Hemochromatosis | 250 (4.6) | 276 (5.2) | 526 (4.9) |
| None of these etiologies | 516 (9.4) | 1,268 (23.8) | 1,784 (16.5) |
Values are presented as number (%), if the value is not specially defined.
NASH, non-alcoholic steatohepatitis.
Overall survival
One-year and 2-year survival rates were 90% (95% CI, 90–91) and 78% (95% CI, 76–79), respectively, for the ablation group, with a mOS of 56 months (95% CI, 54–58) (Fig. 2A). For the resection group of patients, the respective survival rates were 88% (95% CI, 87–89) and 80% (95% CI, 79–81), with an mOS of 75 months (95% CI, 72–79) (Fig. 2B).
Figure 2.
Overall survival (OS) in the ablation (A) and resection group (B), according to the presence of cirrhosis for ablation group (C) and resection group (D), according to the presence of a cirrhosis for ablated patients (E) or resected (F) with the RMST procedure and according to the aetiology of the liver disease for ablation group (H) and resection group (G). CI, confidence interval; RMST, restricted mean survival time.
In the group of patients who underwent ablation, survival rates were significantly lower among those with cirrhosis compared to those without 89% (95% CI, 88–90) vs. 95% (95% CI, 94–97) at 1 year, and 76% (95% CI, 75–78) vs. 85% (95% CI, 82–87) at 2 years. The respective mOS was 53 months (95% CI, 51–55) and 64 months (95% CI, 59–71) in patients with and without cirrhosis (P<0.0001) (cirrhosis associated unadjusted HR for death, 1.418 [95% CI, 1.249–1.610]) (Fig. 2C). The unadjusted 92-month RMST was 55.54 months for patients with cirrhosis and 63.35 months for patients without cirrhosis, showing a difference in RMST of −7.81 months (95% CI, −10.76 to −4.84, P<0.001) (Fig. 2E).
Similarly, in the group of patients who underwent resection, survival rates were significantly lower among those with cirrhosis compared to those without: 87% (95% CI, 86–89) vs. 89% (95% CI, 88–90) at 1 year, and 80% (95% CI, 78–81) vs. 81% (95% CI, 79–83) at 2 years. The respective mOS was 70 (95% CI, 67–72) and 87 months (95% CI, 82–80) in patients with and without cirrhosis (P<0.0007) (unadjusted HR for death in the group of patients with cirrhosis, 1.178 [95% CI, 1.071–1.295]) (Fig. 2D). Unadjusted 92-month RMST was 61.1 months vs. 65.4 months in patients with and without cirrhosis, with a difference in RMST of −4.3 months (95% CI, −6.6 to −2.0, P<0.001) (Fig. 2F).
Of note, the effect of cirrhosis on OS in terms of HR and the difference in RMST was numerically lower but still significant after adjustment for baseline covariates or propensity scores in both ablation and resection group of patients (Supplementary Fig. 1).
Regarding etiologies (analysis based on patients with a single etiology), the ablation group with alcohol-related HCC had a lower mOS when compared to groups with virus-related and metabolic-related etiologies for HCC (mOS of 49 months [95% CI, 46–53]) vs. 81 months (95% CI, 72-NR) vs. 58 months (95% CI, 50–63) for alcohol-related, virus-related and metabolic-related HCC, respectively; overall P<0.0001) (Fig. 2G). Pairwise comparison of etiologies also confirmed these results (P-values from log-rank tests for pairwise comparisons of alcohol-related HCC vs. virus-related HCC and alcohol-related HCC vs. metabolic-related HCC both <0.05) (Supplementary Fig. 1A–C). In the resection group of patients, those with alcohol-related HCC etiology had a significantly lower mOS compared to the group of patients with virus-related etiology, but there was no difference in mOS between alcohol-related and metabolic-related etiology groups (mOS of 72 months [95% CI, 64–78] vs. not reached vs. 73 months [95% CI, 66–77] for alcohol-related, virus-related and metabolic-related HCC, respectively; P<0.00031) (Fig. 2H). Pairwise comparisons for etiologies in the resection group of patients are reported in Supplementary Fig. 1D–F.
Therapeutic sequence
Ablation cohort
Among the 2,424 patients who underwent percutaneous ablation for HCC between 2014 and 2017, 650 (27%) did not receive any further treatment: 307 were alive at last follow-up and 343 had died without receiving any subsequent treatment (median time to death, 20 months [Q1Q3, 7.28–37.77]) (Fig. 3A). A second-line curative treatment was identified in 717 patients (30%), including ablation in 653 (median time, 15 months [Q1Q3, 7.34–27.51]) and resection in 64 patients (median time to 2nd line ablation, 22 months [Q1Q3, 10.57–33.44]). In addition, 124 patients (5%) underwent liver transplantation as second-line therapy (median time, 13 months [Q1Q3, 7.25–22.28]). Other second-line treatments included locoregional therapy in 559 patients (23%; median time, 12 months [Q1Q3, 4.43–24.82]), systemic therapy in 171 (7%; median time, 18 months [Q1Q3, 7.93–30.66]), and best supportive care in 203 (8%; median time, 22 months [Q1Q3, 11.74–27.31]).
Figure 3.
Therapeutic sequences for patients with (A) ablation and (B) resection (patients with at least 4 years of follow-up).
Resection cohort
Among the 2,509 patients who underwent surgical resection, 986 (39%) did not receive any further treatment: 649 were alive at last follow-up and 337 had died without receiving any subsequent treatment (median time to death, 11 months [Q1Q3, 0.59–44.10]) (Fig. 3B). A second-line curative treatment was administered in 446 patients (18%), including ablation in 290 (median time, 24 months [Q1Q3, 13.74–38.69]) and resection in 156 (median time, 25 months [Q1Q3, 13.67–41.05]). In addition, 55 patients (2%) underwent liver transplantation as second-line therapy (median time, 17 months [Q1Q3, 7.74–26.52]). Other second-line treatments included locoregional therapy in 556 patients (22%; median time, 16 months [Q1Q3, 7.74–30.56]), systemic therapy in 291 patients (12%; median time, 11 months [Q1Q3, 5.11–24.72]), and best supportive care in 175 patients (7%; median time, 20 months [Q1Q3, 7.77–33.25]).
Second-line, third-line, and subsequent treatments administered during follow-up are detailed in Fig. 3.
TTNT-D
Finally, mTTNT-D in the ablation group was 20 months (95% CI, 18–21), with 12-month and 24-month landmark TTNT-D of 65% (95% CI, 64–67) and 43% (95% CI, 41–44) (Fig. 4A), respectively. In the resection group, the corresponding TTNT-Ds were 29 months (95% CI, 27–30), 70% (95% CI, 69–72) and 55% (95% CI, 53–56), respectively (Fig. 4B).
Figure 4.
Time to next treatment or death (TTNT-D) for the ablation group (A) and the resection group (B), according to the presence of a cirrhosis for ablated patients (C) or resected (D) with the RMST procedure, and according to the aetiology of the liver disease for ablation group (E) or resection group (F). RMST, restricted mean survival time; CI, confidence interval.
Patients with cirrhosis who initially underwent ablation for HCC had a significantly lower mTTNT-D compared to those without cirrhosis (19 months [95% CI, 17–20] vs. 26 months [95% CI, 21–28] respectively; HR for TTNT-D, 1.267 [95% CI, 1.157–1.339]; P<0.0001) (Fig. 4C). The 3-year RMST for TTNT-D was 29.1 months in cirrhotic patients vs. 34.5 months in the non-cirrhotic patients, with a difference in RMST of −5.4 months (95% CI, −7.8 to −2.9, P<0.001).
Similarly, patients with cirrhosis who initially underwent resection for HCC had a significantly lower mTTNT-D compared to those without cirrhosis (26 months [95% CI, 24–28] vs. 32 months [95% CI, 30–34], respectively; HR for TTNT-D, 1.215 [95% CI, 1.130–1.306]; P<0.0001) (Fig. 4D). The 3-year RMST for TTNT-D was 37.1 months in cirrhotic patients vs. 44.3 months in non-cirrhotic patients, with a difference in RMST of −7.2 months (95% CI, −9.3 to −5.0, P<0.001).
Regarding aetiologies, the ablation group with an alcohol-related HCC had a significantly lower mTTNT-D compared to the others (20 months [95% CI, 18–26] for alcohol-related vs. 24 months [95% CI, 21–26] for virus-related HCC and 22 months [95% CI, 20–25] for metabolic-related HCC; P<0.0112) (Fig. 4E). Resection group experienced similar TTNT-D whatever the etiologies of their liver disease (mTTNT-D of 28 months [95% CI, 25–35] for alcohol-related HCC vs. 31 months [95% CI, 27–39] for virus-related HCC and 30 months [95% CI, 27–33] for metabolic-related HCC; P<0.2439) (Fig. 4F).
Correlation between OS and TTNT-D in the whole population (Supplementary Table 6) was 0.59 (95% CI, 0.46–0.72), with 0.51 (95% CI, 0.48–0.54) in the ablation group and 0.65 (95% CI, 0.64–0.68) in the resection group.
The first next treatment following ablation or resection identified within 90 days has been described to assess post curative act complications. It concerned 10% of the ablation group (n=529, mean of 1.4 treatments) and 6.1% of the resection group (n=315, mean of 1.7 treatments), with 6.2% and 1.7% respectively of curative treatment, 3.1% and 1.4% respectively of loco-regional treatment, 0.6% and 2.7% respectively of systemic treatment, and 0.7% and 0.8% respectively of palliative care.
Ten percent of patients of the ablation group (n=529) and 6.1% in the resection group (n=315) underwent an additional treatment within 90 days of their initial curative intervention. Among these, we identified that 16.1% (n= 85) of patients in the ablation group and 8.3% (n=26) in the resection group underwent a second curative act. In a sensitivity analysis, the exclusion of patients who received two curative-intent procedures within 90 days did not affect TTNT-D outcomes.
Practice evolution
A gradual increase was observed in the use of curative treatments between 2014 and 2021. A crossover point was reached in 2017 when ablation surpassed 50% of total curative procedures, even though the difference between the two procedures remains small. A marked drop-in activity was also observed in 2020, linked to the COVID-19 pandemic (Supplementary Fig. 2).
DISCUSSION
Hepatocellular carcinoma remains a leading cause of liver cancer-related mortality, with high recurrence rates following curative-intent treatments such as resection or ablation. Despite early-stage diagnosis in a substantial proportion of cases, no adjuvant therapy has yet demonstrated a clear survival benefit. The LIVERPOOL study aimed to evaluate the management and prognosis of patients undergoing curative-intent treatment for HCC in routine clinical practice using French nationwide reimbursement data.
Despite the high recurrence rates traditionally associated with resection and ablation in early-stage HCC, this large real-world cohort demonstrates encouraging survival outcomes.10 Our study therefore reflects outcomes in patients undergoing curative-intent procedures rather than strictly BCLC-defined early-stage disease, given the absence of detailed clinical staging variables in the SNDS database.
The relatively short mTTNT-D following resection and ablation reflects the frequent need for additional therapies. However, the majority of patients benefited from subsequent curative, locoregional, or systemic treatments, which likely contributed to sustained survival. Nearly 30% of patients in the ablation group received subsequent curative treatment, highlighting ablation as part of a stepwise multimodal treatment strategy rather than a definitive approach in patients with underlying liver disease. These findings underscore the importance of structured long-term surveillance and access to sequential therapeutic strategies, which together help mitigate the impact of disease recurrence and maintain favourable outcomes in clinical practice. Notably, in France, curative strategies such as resection or local ablation are commonly performed as first-line treatments, while liver transplantation is often considered at recurrence.11 These results are broadly consistent with those reported in nationwide cohorts from Asia, including recent Korean population-based studies, although differences in underlying liver disease etiology, surveillance practices, and healthcare system organization may influence survival outcomes.18
In the current study, although no formal comparison between resection and ablation was performed, median overall survival was numerically higher in the resection group. This observation is consistent with previous observational studies and meta-analyses. 19 However, treatment allocation between resection and ablation largely reflects patient characteristics in routine clinical practice. According to current recommendations, resection is typically proposed to patients with preserved liver function and limited portal hypertension, whereas ablation is preferentially offered to patients who are not surgical candidates, often due to cirrhosis or comorbidities. These differences introduce substantial confounding by indication and should be considered when interpreting survival differences between the two groups. In our cohort, most patients in the ablation group had cirrhosis, a well-known determinant of mortality in liver disease.20 While resection remains the preferred curative option for selected patients, ablation remains a key therapeutic option for cirrhotic or frail individuals, emphasizing the importance of careful patient selection. In addition, in routine clinical practice, both treatment options are rarely considered equally feasible for a given patient, as reflected by the limited number of randomized trials directly comparing resection and ablation. Therefore, these approaches should be viewed as complementary strategies applied to distinct patient populations rather than competing interventions.
Alcohol-related liver disease, the most frequent etiology of HCC in our study, was associated with poorer outcomes compared with viral or metabolic-related HCC. Our findings support previous reports suggesting a worse prognosis in alcohol-related HCC and represent one of the largest contemporary cohorts to confirm this association regardless of the curative procedure.21 Although etiologies may be incompletely captured in administrative databases, particularly for alcohol-related disease or NASH/NAFLD, these results underline the importance of etiology-driven risk stratification. They also highlight the need for tailored surveillance strategies and potentially targeted therapeutic approaches in high-risk populations such as patients with alcohol-related cirrhosis.
This study provides a large nationwide real-world analysis of patients undergoing curative-intent treatment for HCC based on the French SNDS database, which covers nearly the entire French population. By analysing more than 10,000 patients over an 8-year period, it offers valuable insight into how patients are managed outside the strict inclusion criteria of clinical trials. Unlike randomized controlled trials, real-world studies include patients with comorbidities, heterogeneous liver function, and diverse sociodemographic characteristics, thereby providing a broader picture of routine clinical practice.
However, several limitations must be acknowledged. First, the SNDS database was primarily developed for administrative and reimbursement purposes and therefore lacks detailed clinical information. Important prognostic factors such as BCLC stage, liver function scores, tumor burden, AFP levels, and performance status are not available. Consequently, treatment allocation and survival outcomes cannot be fully adjusted for these clinical determinants. Second, diagnoses and comorbidities are identified through administrative coding, which may lead to misclassification or underreporting of some conditions. In addition, treatments administered in the context of clinical trials or outside reimbursed care may not be captured in the database. Finally, follow-up for patients included in the last years of the study period was necessarily shorter due to administrative censoring at the end of the observation period.
TTNT-D was explored as a pragmatic endpoint reflecting clinically meaningful disease progression requiring therapeutic intervention. Such treatment-based endpoints have been increasingly investigated in oncology as complementary measures to OS, particularly in real-world studies where direct measures of disease progression or recurrence may not be available.22-24 However, TTNT-D should not be considered a direct surrogate for recurrence-free survival, as some recurrences may not immediately lead to treatment. In our study, the correlation between TTNT-D and OS was moderate. This finding is consistent with the broader oncology literature, where treatment-based intermediate endpoints capture only part of the survival trajectory, while OS is strongly influenced by subsequent therapies and competing clinical events. In HCC specifically, underlying liver disease, cirrhosis-related complications, and non-tumor mortality represent important competing risks, which may partly explain the moderate association observed, particularly in the ablation group. TTNT-D should therefore be interpreted as a pragmatic marker of clinically meaningful progression requiring therapeutic intervention rather than a direct surrogate for recurrence-free survival.
The role of adjuvant therapy after curative treatment for HCC remains an area of active investigation. In our cohort, most patients had cirrhosis and a large proportion had alcohol-related or metabolic liver disease, conditions associated with a high risk of de novo carcinogenesis. Despite this elevated risk, recent randomized phase III trials evaluating adjuvant therapies have not consistently demonstrated a survival benefit in unselected populations. 7,14-16 These results highlight the need for improved risk stratification to identify patients most likely to benefit from adjuvant strategies. Future approaches integrating molecular profiling, tumor microenvironment characterization, and real-world clinical data may help refine patient selection and guide personalized adjuvant treatment strategies.
In conclusion, this nationwide real-world study provides comprehensive data on survival and treatment pathways following curative-intent therapy for HCC in a Western population. Our findings highlight the complexity of HCC management, the importance of sequential therapeutic strategies after recurrence, and the need for improved risk stratification to optimize long-term outcomes.
Funding Statement
Roche SAS financed the study in its entirety.
Footnotes
Acknowledgements
The authors thank Stephane Bouée from CEMKA for his contribution in the analyses.
Conflicts of Interest
Marie Decraecker: Roche, Servier. Boris Guiu: Roche, Boston Scientific, Canon Medical, Terumo, Sirtex, Guerbet, Astra Zeneca. Emmanuel Boleslawski: Roche. Pierre Nahon: Astra Zeneca, BMS, EISAI, Roche. Manel Dhaoui: Roche. Danko Stamenic: Roche. Moreau Manoel: Roche. Alicia Fravalo: Roche. Jean Frédéric Blanc: Roche, Astra-Zeneca, BMS, MSD, Bayer, Sirtex.
Ethics Statement
As stated by the Health Data Hub and as the study was solely conducted on data from SNDS, the information of the data subjects was not necessary and ethics approval was not required. For the exercise of rights on the main basis of the SNDS, a request has been made to the director of the body that manages the compulsory health insurance scheme to which the person is related, in accordance with Article R1461-9 of the Public Health Code. Finally, data were pseudonymized.
Funding Statement
Roche SAS financed the study in its entirety.
Author Contributions
Conceptualization: MM
Data curation: AF
Formal analysis: AF
Funding acquisition: MM
Investigation: not applicable
Methodology: AF, DS, MM
Project administration: AF, DS, MDh, MM
Resources: not applicable
Software: AF
Supervision: MM, DS, MDh
Validation: AF, MM, DS, MD, BG, EB, JFB, MDh, PN
Visualization: AF
Writing - original draft: DS, MD, MM
Writing - review & editing: AF, BG, EB, JFB, MDh, PN
Data Availability
The aggregated data that support the findings of this study are available from the corresponding author upon reasonable request. SNDS individual data are not available.
Supplementary Materials
Supplementary data can be found with this article online https://doi.org/10.17998/jlc.2026.04.24.
REFERENCES
- 1.Kulik L, El-Serag HB. Epidemiology and management of hepatocellular carcinoma. Gastroenterology. 2019;156:477–491e1. doi: 10.1053/j.gastro.2018.08.065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Younossi ZM. Non-alcoholic fatty liver disease - a global public health perspective. J Hepatol. 2019;70:531–544. doi: 10.1016/j.jhep.2018.10.033. [DOI] [PubMed] [Google Scholar]
- 3.Nabi O, Lacombe K, Boursier J, Mathurin P, Zins M, Serfaty L. Prevalence and risk factors of nonalcoholic fatty liver disease and advanced fibrosis in general population: the French nationwide NASH-CO study. Gastroenterology. 2020;159:791–793e2. doi: 10.1053/j.gastro.2020.04.048. [DOI] [PubMed] [Google Scholar]
- 4.Desai A, Sandhu S, Lai JP, Sandhu DS. Hepatocellular carcinoma in non-cirrhotic liver: a comprehensive review. World J Hepatol. 2019;11:1–18. doi: 10.4254/wjh.v11.i1.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Huang Y, Wallace MC, Adams LA, MacQuillan G, Garas G, Ferguson J, et al. Rate of nonsurveillance and advanced hepatocellular carcinoma at diagnosis in chronic liver disease. J Clin Gastroenterol. 2018;52:551–556. doi: 10.1097/MCG.0000000000000916. [DOI] [PubMed] [Google Scholar]
- 6.Reig M, Forner A, Rimola J, Ferrer-Fàbrega J, Burrel M, Garcia-Criado Á, et al. BCLC strategy for prognosis prediction and treatment recommendation: the 2022 update. J Hepatol. 2022;76:681–693. doi: 10.1016/j.jhep.2021.11.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Association Française pour l’Étude du Foie. Recommandations AFEF sur la prise en charge du carcinome hépatocellulaire & outils pour la pratique [Internet] Paris (FR): Association Française pour l’Étude du Foie; [cited 2025 Apr 1]. Available from: https://afef.asso.fr/recommandation/recommandations-afef-sur-la-prise-en-charge-du-carcinome-hepatocellulaire-outils-pour-la-pratique/ [Google Scholar]
- 8.Singal AG, Kanwal F, Llovet JM. Global trends in hepatocellular carcinoma epidemiology: implications for screening, prevention and therapy. Nat Rev Clin Oncol. 2023;20:864–884. doi: 10.1038/s41571-023-00825-3. [DOI] [PubMed] [Google Scholar]
- 9.Filgueira NA. Hepatocellular carcinoma recurrence after liver transplantation: risk factors, screening and clinical presentation. World J Hepatol. 2019;11:261–272. doi: 10.4254/wjh.v11.i3.261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kim J, Kang W, Sinn DH, Gwak GY, Paik YH, Choi MS, et al. Substantial risk of recurrence even after 5 recurrence-free years in early-stage hepatocellular carcinoma patients. Clin Mol Hepatol. 2020;26:516–528. doi: 10.3350/cmh.2020.0016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Herrero A, Toubert C, Bedoya JU, Assenat E, Guiu B, Panaro F, et al. Management of hepatocellular carcinoma recurrence after liver surgery and thermal ablations: state of the art and future perspectives. Hepatobiliary Surg Nutr. 2024;13:71–88. doi: 10.21037/hbsn-22-579. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Nevola R, Ruocco R, Criscuolo L, Villani A, Alfano M, Beccia D, et al. Predictors of early and late hepatocellular carcinoma recurrence. World J Gastroenterol. 2023;29:1243–1260. doi: 10.3748/wjg.v29.i8.1243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Hsiao CY, Ho CM, Ho MC, Cheng HY, Wu YM, Lee PH, et al. Risk factors, patterns, and outcome predictors of late recurrence in patients with hepatocellular carcinoma after curative resection: a large cohort study with long-term follow-up results. Surgery. 2024;176:2–10. doi: 10.1016/j.surg.2024.02.006. [DOI] [PubMed] [Google Scholar]
- 14.Bruix J, Takayama T, Mazzaferro V, Chau GY, Yang J, Kudo M, et al. Adjuvant sorafenib for hepatocellular carcinoma after resection or ablation (STORM): a phase 3, randomised, double-blind, placebo-controlled trial. Lancet Oncol. 2015;16:1344–1354. doi: 10.1016/S1470-2045(15)00198-9. [DOI] [PubMed] [Google Scholar]
- 15.Yopp A, Kudo M, Chen M, Cheng AL, Kaseb AO, Lee HC, et al. LBA39 updated efficacy and safety data from IMbrave050: phase III study of adjuvant atezolizumab (atezo) + bevacizumab (bev) vs active surveillance in patients (pts) with resected or ablated high-risk hepatocellular carcinoma (HCC) Ann Oncol. 2024;35(Suppl 2):S1230. [Google Scholar]
- 16.Kuang M, Chen Z, Shen S, Peng Z, Li B. 71MO Randomized phase II study of adjuvant radiotherapy after curative resection of hepatocellular carcinoma with narrow margin (≤1 cm) (RAISE) Ann Oncol. 2022;33(Suppl 9):S1458–S1459. [Google Scholar]
- 17.Schemper M, Kaider A, Wakounig S, Heinze G. Estimating the correlation of bivariate failure times under censoring. Stat Med. 2013;32:4781–4790. doi: 10.1002/sim.5874. [DOI] [PubMed] [Google Scholar]
- 18.An J, Chang Y, Choi GH, Sohn W, Song JE, Shin H, et al. Hepatocellular carcinoma in Korea: an analysis of the 2016–2018 Korean Nationwide Cancer Registry. J Liver Cancer. 2025;25:109–122. doi: 10.17998/jlc.2025.02.20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Singal AG, Pillai A, Tiro J. Early detection, curative treatment, and survival rates for hepatocellular carcinoma surveillance in patients with cirrhosis: a meta-analysis. PLoS Med. 2014;11:e1001624. doi: 10.1371/journal.pmed.1001624. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Pinter M, Trauner M, Peck-Radosavljevic M, Sieghart W. Cancer and liver cirrhosis: implications on prognosis and management. ESMO Open. 2016;1:e000042. doi: 10.1136/esmoopen-2016-000042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kraglund F, Villadsen GE, Jepsen P. Effects of curative-intent treatments on hepatocellular carcinoma survival in alcohol-related cirrhosis: a nationwide study. Clin Epidemiol. 2023;15:39–48. doi: 10.2147/CLEP.S393118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Branchoux S, Sofeu CL, Gaudin AF, Kurt M, Moshyk A, Italiano A, et al. Time to next treatment or death as a candidate surrogate endpoint for overall survival in advanced melanoma patients treated with immune checkpoint inhibitors: an insight from the phase III CheckMate 067 trial. ESMO Open. 2022;7:100340. doi: 10.1016/j.esmoop.2021.100340. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Agapow P, Mulla R, Markuzon N, Ottesen LH, Meulendijks D. Systematic review of time to subsequent therapy as a candidate surrogate endpoint in advanced solid tumors. Future Oncol. 2023;19:1627–1639. doi: 10.2217/fon-2022-0616. [DOI] [PubMed] [Google Scholar]
- 24.Terashima T, Yamashita T, Toyama T, Arai K, Kawaguchi K, Kitamura K, et al. Surrogacy of time to progression for overall survival in advanced hepatocellular carcinoma treated with systemic therapy: a systematic review and meta-analysis of randomized controlled trials. Liver Cancer. 2019;8:130–139. doi: 10.1159/000489505. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The aggregated data that support the findings of this study are available from the corresponding author upon reasonable request. SNDS individual data are not available.