Abstract
Background
Recurrence rates following thermal ablation for hepatocellular carcinoma (HCC) remain high, but the patterns of recurrence and post-recurrence outcomes are not well characterized. This study aimed to investigate the recurrence patterns and long-term post-recurrence survival (PRS) in patients with HCC after thermal ablation to inform post-recurrence treatment strategy.
Methods
A retrospective analysis was conducted on 824 patients who underwent thermal ablation for HCC between 2007 and 2023. Recurrence patterns and factors influencing PRS in patients with recurrence within and beyond the Milan criteria were evaluated. An independent cohort of 198 patients served as an external validation cohort for the prognostic models.
Results
During a median follow-up of 54.5 months, 536 patients experienced HCC recurrence, with 83.8% within and 16.2% beyond Milan criteria. For patients with recurrence within Milan criteria, early recurrence, recurrent tumor size, AFP, ALBI grade, and FIB-4 score were independent predictors of PRS. In patients with recurrence beyond Milan criteria, diabetes mellitus, macrovascular invasion, AFP, ALBI grade, and FIB-4 score independently predicted PRS. Based on PRS predictors, a risk model stratified patients with recurrence within Milan criteria into four risk groups, with median PRS of 103.7, 65, 48.7, and 28.6 months, respectively (p < 0.001). For recurrence beyond Milan criteria, a separate risk model classified patients into three risk groups, showing median PRS of 70.1, 24.7, and 8.1 months, along with probabilities of successful downstaging of 66.7%, 39.3%, and 0%, respectively (p < 0.001). The external validation results showed that both the Milan-in and Milan-out models demonstrated significant discriminative performance in the validation cohort.
Conclusions
PRS in patients with recurrent HCC after thermal ablation is significantly influenced by recurrence patterns, tumor characteristics, and host factors. These findings may guide post-recurrence treatment strategies and optimize the timing of salvage liver transplantation.
Keywords: Thermal ablation, Hepatocellular carcinoma, Recurrence, Survival
Introduction
Hepatocellular carcinoma (HCC) ranks as the third leading cause of cancer-related mortality globally and remains one of the few malignancies with an increasing mortality rate [1, 2]. Beyond liver transplantation (LT) and surgical resection, percutaneous thermal ablation has emerged as a well-established curative option [3–5]. It achieves a complete response in approximately 90% of cases and yields 5-year survival rates ranging from 60% to 76%, even in patients who are not candidates for surgical resection [6, 7]. Despite these favorable outcomes, tumor recurrence occurs in 60–80% of patients within 5 years following ablation [6–8]. Nevertheless, guidance on optimal management strategies after recurrence remains limited.
In patients with recurrent HCC following ablation, tumor biology, recurrence patterns, and liver function are highly heterogeneous, leading to substantial variation in survival outcomes. Post-recurrence treatment strategies depend largely on the tumor burden, yet evidence-based guidance remains insufficient. According to current American Association for the Study of Liver Diseases (AASLD) guidelines, salvage LT is recommended for recurrence within the Milan criteria [9]. Nevertheless, LT is constrained by organ shortages and the potential for perioperative complications, particularly in countries with a high HCC burden [10]. Moreover, outcomes after recurrence vary widely. While some patients with preserved liver function and limited tumor burden can achieve long-term survival through repeat curative therapies [11], others experience rapid tumor progression or hepatic decompensation, resulting in poor short-term post-recurrence survival (PRS). Therefore, identifying key prognostic factors after HCC recurrence is crucial for guiding treatment selection and determines which patients may benefit most from salvage LT versus repeat locoregional treatments.
For patients with recurrence beyond the Milan criteria, locoregional or systemic therapies are recommended based on tumor stage. When recurrence is confined to the liver and meets downstaging criteria, liver-directed therapies are advised, and salvage LT may be considered if downstaging is successful [9]. Notably, LT following effective downstaging has demonstrated survival outcomes comparable to those in patients initially within the Milan criteria [12]. Nevertheless, survival expectations and the likelihood of achieving successful downstaging in this population remain insufficiently defined. This study aimed to characterize recurrence patterns and long-term PRS in patients who developed recurrence within or beyond the Milan criteria after thermal ablation, and identify key prognostic factors associated with PRS to inform treatment stratification and guide decision-making regarding repeat locoregional therapy and salvage LT.
Patients and Methods
Patients
A total of 1,017 patients diagnosed with HCC who underwent curative radiofrequency ablation (RFA) or microwave ablation at Taipei Veterans General Hospital between October 1, 2007, and March 31, 2023, were retrospectively screened (Fig. 1). Inclusion criteria required participants to be 20 years of age or older and to have undergone RFA or microwave ablation as the primary treatment for Barcelona Clinic Liver Cancer (BCLC) stage 0 or A HCC. Exclusion criteria were as follows: patients with mixed HCC-cholangiocarcinoma or other tumor types (n = 14), lack of confirmation of complete tumor ablation by dynamic imaging (n = 12), failure to achieve complete ablation (n = 95), BCLC stage B or C (n = 46), ablation performed after surgery or transarterial chemoembolization (TACE) (n = 55), concurrent TACE for HCC (n = 1), or death or loss to follow-up within 3 months post-ablation (n = 24). HCC diagnosis was made prior to ablation using contrast-enhanced computed tomography (CECT) or magnetic resonance imaging (MRI), in accordance with the diagnostic guidelines set by the AASLD [3]. In cases where imaging criteria were not met, a percutaneous liver biopsy was performed to obtain pathological confirmation. Complete ablation, defined as the absence of any viable residual tumor, was verified via CECT or MRI 1 month after the procedure. Following successful ablation, patients underwent follow-up evaluations every 3 months during the first 2 years, and subsequently every 4 to 6 months. These assessments included routine checks of liver function, alpha-fetoprotein (AFP) levels, ultrasound, and CECT or MRI scans. Tumor recurrence was confirmed through CECT or MRI. To evaluate the impact of temporal changes in HCC treatment, we stratified patients with recurrence into two time periods: 2007–2016 and 2017–2023. The latter period corresponds to the introduction of immunotherapy (nivolumab) at our institution and represents a new era in HCC management. An independent cohort of 198 patients who received first-line thermal ablation for HCC between April 2009 and December 2024 at Taichung Veterans General Hospital was used as an external validation cohort for the prognostic models.
Fig. 1.
Patient flow of HCC patients undergoing thermal ablation.
Outcome Assessment
The endpoints of this study included time to recurrence, defined as the interval from the date of curative ablation to the time of recurrence, and PRS, defined as the duration from recurrence to death. Local tumor progression (LTP) was defined as the appearance of tumor foci at the margin of the ablation zone, while distant recurrence was defined as the presence of tumor foci outside the ablation zone or the occurrence of extrahepatic metastases, as confirmed by follow-up CECT or MRI after achieving complete ablation [13, 14]. Recurrence status was assessed based on the BCLC stage at recurrence, classification as within or beyond the Milan criteria, within or beyond the United Network for Organ Sharing-Downstaging (UNOS-DS) criteria [3], post-recurrence treatment modalities, and whether successful downstaging to meet the Milan criteria was achieved.
Clinical Information
Clinical parameters were recorded, including age, gender, BMI, hepatitis B surface antigen, nucleos(t)ide analogs (NUC) treatment, anti-HCV antibodies, sustained virological response (SVR) status, diabetes mellitus (DM) status, BCLC staging, tumor size, number of tumors, Child-Pugh score, and serum levels of alanine aminotransferase, aspartate aminotransferase, albumin, creatinine, total bilirubin, platelet count, and AFP. AFP levels were measured using a chemiluminescent microparticle immunoassay (ARCHITECT AFP assay, Abbott Ireland Diagnostics Division, Sligo, Ireland). Biochemical analyses were performed using a multichannel autoanalyzer (Technicon SMAC, Technicon Instruments Corp., Tarrytown, NY, USA). The Fibrosis-4 (FIB-4) score and albumin-bilirubin (ALBI) grade were calculated as previously described [15]. Type 2 DM was diagnosed according to the plasma glucose criteria outlined by the American Diabetes Association [16]. Tumor morphology of post-ablation recurrent tumors was classified into four types: simple nodular, simple nodular with extranodular growth, confluent multinodular, and infiltrative types [17, 18].
Statistical Analysis
Statistical analyses were conducted using IBM SPSS Statistics version 22 (IBM, Armonk, NY). Data are expressed as median (interquartile range) or mean ± standard deviation, depending on the distribution of the variables. For continuous variables, the Mann-Whitney U test was applied, while categorical variables were assessed using Pearson’s chi-square test or Fisher’s exact test, as appropriate. Binary logistic regression analysis was conducted to identify factors associated with recurrence beyond Milan criteria. Kaplan-Meier analysis was employed to estimate recurrence and survival rates, with survival curves compared using the log-rank test. To identify prognostic factors associated with PRS, Cox proportional hazards regression models were utilized. Variables with a p value <0.1 in univariate analysis were subsequently included in a multivariate analysis using a forward stepwise Cox regression model. A two-tailed p value <0.05 was considered statistically significant.
Results
Patient Characteristics
A total of 824 patients were included in the analysis of time to recurrence (Fig. 1). Following exclusions of patients who did not experience recurrence during the follow-up period (n = 273), were lost to follow-up after recurrence (n = 12), or underwent LT after recurrence (n = 3), 536 patients remained eligible for PRS analysis. Of these, 449 patients (83.8%) met the Milan criteria, while 87 patients (16.2%) beyond Milan criteria.
Table 1 presents the baseline and recurrence characteristics of patients who experienced recurrence after undergoing ablation for HCC. At baseline, 42.5% of patients were classified as BCLC stage 0, and 57.5% were categorized as BCLC stage A. At recurrence, the majority of the patients had a single tumor and were classified as Child-Pugh class A. Patients with recurrence beyond Milan criteria exhibited significantly larger tumor size, higher BCLC stage, poorer liver function reserve, higher AFP levels, ALBI grade, and FIB-4 scores.
Table 1.
Baseline characteristics of the 536 patients with HCC recurrence after thermal ablation
| Characteristics | Overall (n = 536) | Within Milan criteria (n = 449; 83.8%) | Beyond Milan criteria (n = 87; 16.2%) | p value |
|---|---|---|---|---|
| Baseline characteristics | ||||
| Age, years | 68.4±11.4 | 68.4±11.4 | 68.4±11.5 | 0.905 |
| Sex (male), n (%) | 348 (64.9) | 288 (64.1) | 60 (69.0) | 0.459 |
| BMI, kg/m2 | 25.3±4.2 | 25.4±4.1 | 24.7±4.3 | 0.218 |
| HBsAg-positive, n (%) | 274 (51.1) | 228 (50.8) | 46 (52.9) | 0.810 |
| NUC therapy, n (%) | 217 (78.1) | 184 (80.0) | 33 (68.8) | 0.128 |
| Anti-HCV-positive, n (%) | 199 (37.1) | 171 (38.1) | 28 (32.2) | 0.357 |
| Achieving SVR, n (%) | 87 (43.7) | 79 (46.2) | 8 (28.6) | 0.124 |
| DM, n (%) | 172 (32.3) | 149 (33.4) | 23 (26.4) | 0.251 |
| BCLC stage 0/A, n (%) | 228/308 (42.5/57.5) | 204/245 (45.4/54.6) | 24/63 (27.6/72.4) | 0.003 |
| Tumor size, cm | 2.25±0.84 | 2.22±0.85 | 2.39±0.79 | 0.034 |
| Tumor numbers 1/2/3, n (%) | 466/59/11 (86.9/11.0/2.1) | 399/40/10 (88.9/8.9/2.2) | 67/19/1 (77.0/21.8/1.1) | 0.026 |
| RFA/MWA, % | 525/11 (97.9/2.1) | 441/8 (98.2/1.8) | 84/3 (96.6/3.4) | 0.398 |
| AFP, ng/mL | 13.0 (5.4–64.2) | 12.4 (5.1–56.4) | 18.5 (6.5–97.5) | 0.069 |
| Child-Pugh class A/B, n (%) | 500/34 (93.6/6.4) | 423/24 (94.6/5.4) | 77/10 (88.5/11.5) | 0.057 |
| ALBI grade 1/2/3, n (%) | 261/257/16 (48.9/48.1/3.0) | 223/212/12 (49.9/47.4/2.7) | 38/45/4 (43.7/51.7/4.6) | 0.418 |
| FIB-4 score | 5.82±5.22 | 5.79±5.34 | 5.95±4.55 | 0.607 |
| Characteristic at HCC recurrence | ||||
| Time to recurrence, months | 14.6 (7.4–29.9) | 15.4 (7.6–30.9) | 12.8 (7.0–26.3) | 0.130 |
| Early/late recurrence, n (%) | 361/175 (67.4/32.6) | 297/152 (66.1/33.9) | 64/23 (73.6/26.4) | 0.220 |
| Recurrence site | | | | <0.001 |
| LTP, n (%) | 155 (28.9) | 141 (31.4) | 14 (16.1) | |
| Intrahepatic distant recurrence, n (%) | 309 (57.6) | 277 (61.7) | 32 (36.8) | |
| Extrahepatic distant recurrence, n (%) | 5 (0.9) | 0 (0) | 5 (5.7) | |
| BCLC stage 0/A/B/C/D, n (%) | 257/196/47/34/2 (47.9/36.6/8.8/6.3/0.4) | 257/191/0/0/1 (57.2/42.5/0/0/0.2) | 0/5/47/34/1 (0/5.7/54/39.1/1.1) | <0.001 |
| Macrovascular invasion, n (%) | 21 (3.9) | 0 (0) | 21 (24.1) | <0.001 |
| Extrahepatic metastasis, n (%) | 14 (2.6) | 0 (0) | 14 (16.1) | <0.001 |
| Tumor size, cm | 2.06±1.43 | 1.81±0.76 | 3.30±2.78 | <0.001 |
| Tumor numbers 1/2/3/≥4, n (%) | 396/66/29/38 (73.9/12.3/5.4/7.1) | 373/56/20/0 (83.1/12.5/4.5/0) | 23/10/9/38 (26.4/11.5/10.3/43.7) | <0.001 |
| Tumor morphology type | | | | <0.001 |
| Simple nodular | 407 (75.9) | 370 (82.4) | 37 (42.5) | |
| Simple nodular type with extranodular growth | 55 (10.3) | 42 (9.4) | 13 (14.9) | |
| Confluent multinodular | 44 (8.2) | 28 (6.2) | 16 (18.4) | |
| Infiltrative | 23 (4.3) | 9 (2.0) | 14 (16.1) | |
| Age, years | 70.3±11.3 | 70.4±11.3 | 70.0±11.5 | 0.638 |
| BMI, kg/m2 | 25.5±4.2 | 25.6±4.3 | 25.2±4.0 | 0.566 |
| Child-Pugh class A/B/C, n (%) | 455/70/2 (86.3/13.3/0.4) | 394/48/1 (88.9/10.8/0.2) | 61/22/1 (72.6/26.2/1.2) | <0.001 |
| AFP, ng/mL | 10.5 (4.1–53.8) | 9.5 (3.8–37.6) | 36.4 (5.81–355.7) | <0.001 |
| Albumin, g/dL | 3.81±0.59 | 3.84±0.58 | 3.64±0.61 | 0.006 |
| Total bilirubin, mg/dL | 0.86±1.03 | 0.77±0.54 | 1.37±2.19 | <0.001 |
| ALBI score | −2.536±0.601 | −2.581±0.578 | −2.302±0.666 | <0.001 |
| ALBI grade 1/2/3, n (%) | 282/230/22 (52.8/43.1/4.1) | 251/183/14 (56/40.8/3.1) | 31/47/8 (36/54.7/9.3) | <0.001 |
| Platelet count, 109/L | 126±68 | 127±69 | 121±63 | 0.458 |
| ALT, U/L | 42.5±36.4 | 43.3±38.6 | 38.1±20.6 | 0.346 |
| AST, U/L | 50.1±38.4 | 48.8±39.0 | 56.8±34.6 | <0.001 |
| FIB-4 score | 6.06±4.95 | 5.82±4.82 | 7.30±5.41 | 0.006 |
| Post-recurrence treatment, n (%) | | | | <0.001 |
| Surgical resection | 29 (5.4) | 20 (4.5) | 9 (10.3) | |
| Local ablation | 354 (66.0) | 347 (77.3) | 7 (8.0) | |
| TACE | 118 (22.0) | 68 (15.1) | 50 (57.5) | |
| Systemic therapy | 7 (1.3) | 0 (0) | 7 (8.0) | |
| Others | 28 (5.2) | 14 (3.1) | 14 (16.1) | |
BMI, body mass index; LTP, local tumor progression; NUC, nucleos(t)ide analogs; SVR, sustained virological response; ALBI, albumin-bilirubin; TACE, transarterial chemoembolization; RFA, radiofrequency ablation; MWA, microwave ablation; HBsAg, hepatitis B surface antigen; ALT, alanine aminotransferase; AST, aspartate aminotransferase.
HCC Recurrence Pattern
During a median follow-up period of 52.1 months, tumor recurrence occurred in 536 (65%) of the patients after ablation. The cumulative recurrence rates at 1, 2, 5, and 10 years were 29.6%, 47.7%, 70.0%, and 82.8%, respectively (online suppl. Fig. 1A; for all online suppl. material, see https://doi.org/10.1159/000550105). Among patients with recurrence, 155 (28.9%) had isolated LTP, 314 (58.6%) experienced isolated distant recurrence, and 67 (12.5%) developed concurrent LTP and distant recurrence (online suppl. Fig. 1B). Patients with isolated LTP or distant recurrence had comparable PRS, whereas those with concurrent LTP and distant recurrence exhibited significantly poorer PRS (online suppl. Fig. 1C).
Among patients with tumor recurrence during the observation period, 43.5% experienced recurrence within the first year, and 23.9% within the second year. The proportion of early recurrences (within the first 2 years) was 67.4%, with the recurrence rate gradually decreasing in subsequent years (Fig. 2a). Patients with early recurrence had significantly poorer PRS (p = 0.002, Fig. 2d).
Fig. 2.
HCC recurrence pattern and PRS according to timeframe, recurrence tumor burden, and post-recurrence treatment. a Distribution of recurrence at each year after thermal ablation. b Distribution of recurrence tumor burden stratified by BCLC stage and Milan criteria. c Distribution of post-recurrence treatment. d PRS stratified by early and late recurrence. e PRS stratified by Milan criteria and UNOS-DS criteria. f PRS stratified by post-recurrence treatment.
At the time of recurrence, 47.9% were classified as BCLC 0, 36.6% as BCLC A, 8.8% as BCLC B, 6.5% as BCLC C, and 0.4% as BCLC D (Fig. 2b). Overall, 83.8% had recurrence within the Milan criteria, while 16.2% were beyond Milan criteria. PRS was significantly better in patients with recurrence within the Milan criteria, followed by those with recurrence beyond the Milan but within the UNOS-DS criteria, then those with liver-localized recurrence beyond the UNOS-DS criteria, and poorest among patients with macrovascular invasion or extrahepatic metastasis (p < 0.001, Fig. 2e).
After recurrence, 5.4% of patients underwent resection, 66% received ablation, 22% were treated with TACE, and 1.3% received systemic therapy (Fig. 2c). Significant differences in PRS were observed across the treatment groups, with the highest survival rates in patients undergoing resection or ablation, followed by those receiving TACE, systemic therapy, and other treatments (p < 0.001, Fig. 2f). No significant difference in PRS was found between the resection and ablation groups (p = 0.119).
Factors Associated with Recurrence beyond Milan Criteria
In univariate analysis, BCLC stage, tumor size, tumor number, AFP, and bilirubin levels were found to be associated with recurrence beyond Milan criteria. In multivariate analysis, BCLC stage A (odds ratio [OR] = 2.255, p = 0.002), AFP >20 ng/mL (OR = 2.191, p = 0.001), and bilirubin >1.2 mg/dL (OR = 2.170, p = 0.005) were identified as independent predictors of recurrence beyond Milan criteria (online suppl. Table 1).
Factors Associated with PRS in Patients with Recurrence within the Milan Criteria
Among patients who experienced recurrence of HCC within the Milan criteria, univariate analysis identified age, DM, hepatitis B surface antigen positivity, early recurrence, concurrent LTP and distant recurrence, BCLC stage, tumor size, AFP, ALBI grade, FIB-4 score, and tumor morphology type as baseline factors associated with poorer PRS. In multivariate analysis, early recurrence (hazard ratio [HR] = 1.412, p = 0.015), tumor size >2 cm (HR = 1.510, p = 0.002), AFP >20 ng/mL (HR = 1.391, p = 0.009), ALBI grade 2–3 (HR = 1.675, p < 0.001), and FIB-4 score >3.25 (HR = 1.761, p < 0.001) were identified as independent predictors of poorer PRS (Table 2; online suppl. Fig. 2).
Table 2.
Univariate and multivariate analyses of factors associated with PRS in patients with recurrence within Milan criteria
| | Case, n | Univariate | Multivariate | ||||
|---|---|---|---|---|---|---|---|
| HR | 95% CI | p value | HR | 95% CI | p value | ||
| Characteristic at HCC recurrence | |||||||
| Age, >65/≤65 years | 316/133 | 1.465 | 1.112–1.928 | 0.007 | | | NS |
| Male/female | 288/161 | 0.978 | 0.761–1.257 | 0.861 | | | |
| DM, yes/no | 149/297 | 1.327 | 1.034–1.702 | 0.026 | | | NS |
| BMI, >25/≤25 kg/m2 | 212/181 | 1.083 | 0.827–1.418 | 0.562 | | | |
| HBsAg, positive/negative | 231/218 | 0.683 | 0.537–0.870 | 0.002 | | | NS |
| Anti-HCV, positive/negative | 278/171 | 1.121 | 0.879–1.431 | 0.358 | | | |
| Early recurrence, yes/no | 297/152 | 1.460 | 1.110–1.921 | 0.007 | 1.412 | 1.068–1.867 | 0.015 |
| Recurrence site | |||||||
| LTP | 141 | 1 | | 0.042 | | | NS |
| Distant recurrence | 277 | 1.054 | 0.807–1.378 | 0.698 | | | |
| Concurrent LTP and distant recurrence | 31 | 1.789 | 1.123–2.852 | 0.014 | | | |
| BCLC stage, A/0 | 257/191 | 1.319 | 1.036–1.680 | 0.024 | | | NS |
| Tumor size, >2/≤2 cm | 133/316 | 1.417 | 1.100–1.825 | 0.007 | 1.510 | 1.163–1.960 | 0.002 |
| Tumor number, multiple/single | 76/373 | 1.205 | 0.886–1.641 | 0.235 | | | |
| AFP, >20/≤20 ng/mL | 162/285 | 1.618 | 1.268–2.064 | <0.001 | 1.391 | 1.086–1.782 | 0.009 |
| Platelet count, >150/≤150, 109/L | 134/313 | 0.811 | 0.619–1.063 | 0.129 | | | |
| ALT, >40/≤40 U/L | 150/298 | 1.495 | 1.173–1.905 | 0.001 | | | NS |
| AST, >40/≤40 U/L | 180/261 | 1.889 | 1.481–2.409 | <0.001 | | | NS |
| ALBI grade, 2–3/1 | 197/251 | 2.247 | 1.763–2.865 | <0.001 | 1.675 | 1.282–2.189 | <0.001 |
| FIB-4 score, >3.25/≤3.25 | 282/165 | 2.238 | 1.704–2.938 | <0.001 | 1.761 | 1.304–2.379 | <0.001 |
| Tumor morphology type | |||||||
| Simple nodular | 370 | 1 | | <0.001 | | | NS |
| Simple nodular type with extranodular growth | 42 | 1.660 | 1.137–2.422 | 0.009 | | | |
| Confluent multinodular | 28 | 1.492 | 0.932–2.388 | 0.096 | | | |
| Infiltrative | 9 | 3.719 | 1.901–7.277 | <0.001 | | | |
HR, hazard ratio; CI, confidence interval; NS, not significant; BMI, body mass index; LTP, local tumor progression; ALBI, albumin-bilirubin; HBsAg, hepatitis B surface antigen; ALT, alanine aminotransferase; AST, aspartate aminotransferase.
Post-recurrence treatment may also have a substantial impact on PRS, particularly for patients who remain eligible for repeated curative therapy. We evaluated the prognostic role of curative versus non-curative treatment after the first recurrence following ablation. In the multivariate analysis, non-curative therapy was independently associated with significantly poorer PRS in patients with recurrence within the Milan criteria (HR = 1.917, p < 0.001), after adjustment for other major predictors (online suppl. Table 2).
Factors Associated with PRS in Patients with Recurrence beyond Milan Criteria
In univariate analysis, DM, tumor size, tumor number, BCLC stage, macrovascular invasion, AFP level, ALBI grade, FIB-4 score, and infiltrative-type HCC were associated with poorer PRS in patients with recurrence of HCC beyond Milan criteria. Non-curative therapy was not associated with PRS in patients with recurrence beyond the Milan criteria (HR = 1.277, p = 0.443). In the multivariate analysis, DM (HR = 4.011, p < 0.001), macrovascular invasion (HR = 2.542, p = 0.001), AFP >400 ng/mL (HR = 1.867, p = 0.039), ALBI grade 2–3 (HR = 2.722, p = 0.002), and FIB-4 score >3.25 (HR = 2.421, p = 0.018) were identified as independent predictors of poorer PRS (Table 3; online suppl. Fig. 3).
Table 3.
Univariate and multivariate analyses of factors associated with PRS in patients with recurrence beyond Milan criteria
| | Case, n | Univariate | Multivariate | ||||
|---|---|---|---|---|---|---|---|
| HR | 95% CI | p value | HR | 95% CI | p value | ||
| Characteristic at HCC recurrence | |||||||
| Age, >65/≤65 years | 57/30 | 0.868 | 0.535–1.408 | 0.565 | | | |
| Male/female | 60/27 | 1.217 | 0.716–2.069 | 0.468 | | | |
| DM, yes/no | 23/64 | 2.343 | 1.389–3.953 | 0.001 | 4.011 | 2.248–7.156 | <0.001 |
| BMI, >25/≤25 kg/m2 | 36/37 | 0.987 | 0.605–1.610 | 0.958 | | | |
| HBsAg, positive/negative | 47/40 | 0.829 | 0.514–1.335 | 0.440 | | | |
| Anti-HCV, positive/negative | 28/59 | 1.121 | 0.879–1.431 | 0.358 | | | |
| Early recurrence, yes/no | 64/23 | 1.328 | 0.747–2.361 | 0.333 | | | |
| Recurrence site | |||||||
| LTP | 14 | 1 | | 0.661 | | | |
| Intrahepatic distant recurrence | 32 | 0.802 | 0.381–1.689 | 0.562 | | | |
| Extrahepatic distant recurrence | 5 | 1.140 | 0.389–3.341 | 0.812 | | | |
| Concurrent LTP and distant recurrence | 36 | 1.124 | 0.549–2.304 | 0.794 | | | |
| Tumor size, >5/≤5 cm | 41/46 | 0.613 | 0.380–0.989 | 0.045 | | | NS |
| Tumor number, multiple/single | 57/30 | 0.625 | 0.380–1.026 | 0.063 | | | NS |
| UNOS-DS criteria, beyond/within | 29/58 | 1.689 | 0.993–2.874 | 0.054 | | | NS |
| BCLC stage, C/A-B | 34/52 | 1.682 | 1.027–2.755 | 0.039 | | | NS |
| Macrovascular invasion, yes/no | 21/66 | 2.410 | 1.419–4.093 | 0.001 | 2.542 | 1.450–4.455 | 0.001 |
| Extrahepatic metastasis, yes/no | 14/73 | 1.183 | 0.633–2.213 | 0.598 | | | |
| AFP, >400/≤400 ng/mL | 21/66 | 2.150 | 1.302–3.549 | 0.003 | 1.867 | 1.033–3.372 | 0.039 |
| Platelet count, >150/≤150, 109/L | 23/60 | 0.422 | 0.228–0.778 | 0.006 | | | NS |
| ALT, >40/≤40 U/L | 36/51 | 1.288 | 0.797–2.081 | 0.301 | | | |
| AST, >40/≤40 U/L | 51/33 | 1.988 | 1.196–3.302 | 0.008 | | | NS |
| ALBI grade, 2–3/1 | 55/31 | 3.004 | 1.742–5.178 | <0.001 | 2.722 | 1.448–5.118 | 0.002 |
| FIB-4 score, >3.25/≤3.25 | 64/22 | 2.911 | 1.547–5.478 | 0.001 | 2.421 | 1.166–5.025 | 0.018 |
| Tumor morphology type | |||||||
| Simple nodular | 37 | 1 | | 0.055 | | | NS |
| Simple nodular type with extranodular growth | 13 | 1.030 | 0.485–2.185 | 0.939 | | | |
| Confluent multinodular | 16 | 1.414 | 0.717–2.787 | 0.318 | | | |
| Infiltrative | 14 | 2.419 | 1.258–4.650 | 0.008 | | | |
HR, hazard ratio; CI, confidence interval; NS, not significant; BMI, body mass index; LTP, local tumor progression; ALBI, albumin-bilirubin; HBsAg, hepatitis B surface antigen; ALT, alanine aminotransferase; AST, aspartate aminotransferase.
Risk Models for Predicting PRS in Patients with Recurrence within and beyond Milan Criteria
Simple risk models for predicting PRS in patients with recurrence within and beyond the Milan criteria were developed based on the independent survival predictors selected by multivariate analyses. In patients with recurrence within Milan criteria, the risk model was comprised of five independent survival predictors. One point was assigned to each of the following factors: early recurrence, FIB-4 score >3.25, ALBI grade 2–3, AFP >20 ng/mL, and tumor size >2 cm, as each factor had a similar beta coefficient (Table 4). Patients were categorized into four risk groups based on their total score: low (0–1), intermediate (2), high (3), and very high (4–5). The corresponding PRS for each risk group was 103.7 months, 65.7 months, 48.7 months, and 28.6 months, respectively, and the corresponding 5-year survival rate was 75.5%, 50.7%, 38.8%, and 14.3%, respectively (Fig. 3a).
Table 4.
Risk group classification based on predictors of PRS in patients with recurrence within and beyond Milan criteria
| Risk factors | Beta coefficient | Points | Risk score | Risk group |
|---|---|---|---|---|
| Recurrence within Milan criteria | ||||
| Early recurrence | 0.334 | 1 | 0–1 | Low |
| FIB-4 >3.25 | 0.599 | 1 | 2 | Intermediate |
| ALBI grade 2–3 | 0.534 | 1 | 3 | High |
| Tumor size >2 cm | 0.383 | 1 | 4–5 | Very high |
| AFP >20 ng/mL | 0.327 | 1 | | |
| Recurrence beyond Milan criteria | ||||
| DM | 1.389 | 4 | 0–5 | Low |
| FIB-4 >3.25 | 0.884 | 3 | 6–7 | Intermediate |
| ALBI grade 2–3 | 1.001 | 3 | 8–15 | High |
| Macrovascular invasion | 0.933 | 3 | | |
| AFP >400 ng/mL | 0.624 | 2 | | |
Fig. 3.
PRS stratified by risk models. a PRS in patients with recurrence within the Milan criteria. b PRS in patients with recurrence beyond the Milan criteria. c PRS in the external validation cohort with recurrence within the Milan criteria. d PRS in the external validation cohort with recurrence beyond the Milan criteria.
Similarly, a risk model for predicting PRS in patients with recurrence beyond Milan was developed based on five independent predictors. Points were assigned as follows: four points for the presence of DM, three points each for FIB-4 score >3.25, ALBI grade 2–3, and macrovascular invasion, and two points for AFP >400 ng/mL, based on their respective beta coefficient (Table 4). Patients were classified into low (0–5), intermediate (6–7), and high (8–15) risk categories for poorer PRS. The corresponding PRS for each risk group was 70.1 months, 24.7 months, and 8.1 months, respectively, and the corresponding 2-year survival rate was 80.7%, 50.8%, and 13.3%, respectively (Fig. 3b).
An independent cohort of 198 patients from Taichung Veterans General Hospital, including 160 patients with recurrence within the Milan criteria and 38 patients with recurrence beyond the Milan criteria, was used for external validation. The characteristics of patients in the external validation cohort are summarized in online supplementary Table 3. The Milan-in model demonstrated strong discriminative performance in the external validation cohort, with the corresponding 5-year survival rate of 87.9%, 84.1%, 59.0%, and 40.2%, respectively (p < 0.001, Fig. 3c). The Milan-out model also showed good discriminative ability, with the corresponding 2-year survival rates of 100%, 80.0%, and 44.1%, respectively (p = 0.014, Fig. 3d).
Probability of Achieving Successful Downstaging to Milan Criteria
We further assessed the likelihood of achieving successful downstaging to Milan criteria in patients with recurrence beyond Milan criteria. Among the 87 patients with recurrence beyond Milan criteria, 32 (36.8%) successfully achieved downstaging to Milan criteria following treatment. The probability of downstaging was 62.1% and 20.7% in patients within and beyond the UNOS-DS criteria (p < 0.001, Fig. 4a). The beyond Milan risk model also stratified the likelihood of successful downstaging among patients with recurrence beyond the Milan criteria. The probabilities of downstaging were 66.7%, 39.3%, and 0% in the low-, intermediate-, and high-risk groups, respectively (p < 0.001; Fig. 4b).
Fig. 4.
Probability of achieving successful downstaging in patients with recurrence beyond Milan criteria. a Probability of downstaging stratified by UNOS-DS criteria. b Probability of downstaging stratified by the Milan-out risk model.
Impact of Antiviral Therapy Uptake and Time Period on PRS
In patients with HBV infection, those receiving NUC therapy or with undetectable HBV DNA had significantly better PRS (p = 0.030, online suppl. Fig. 4A), while in patients with HCV infection, achieving SVR was associated with improved PRS (p < 0.001, online suppl. Fig. 4B). The uptake of HBV NUC therapy increased from 73.6% in 2007–2016 to 89.4% in 2017–2023 (p = 0.003), and HCV SVR rates increased from 27.5% to 80.3% over the same periods (p < 0.001, online suppl. Fig. 4C).
Correspondingly, the median PRS in the overall cohort increased from 44.5 months in 2007–2016 to 70.1 months in 2017–2023 (p = 0.001, online suppl. Fig. 5A). Among patients with recurrence within the Milan criteria, the median PRS improved from 47.1 months to not reached (p = 0.003, online suppl. Fig. 5B), whereas in those with recurrence beyond the Milan criteria, it increased from 17.0 to 28.9 months (p = 0.229, online suppl. Fig. 5C).
The Milan-in model demonstrated a significant ability to discriminate PRS across different time periods. Among patients with recurrence between 2007 and 2016, the 5-year PRS rates for the low-, intermediate-, high-, and very high-risk groups were 75.6%, 47.3%, 39.6%, and 13.9%, respectively (p < 0.001, online suppl. Fig. 6A). During 2017–2023, the corresponding rates were 78.6%, 59.1%, 37.7%, and 16.4% (p < 0.001, online suppl. Fig. 6B). Similarly, the Milan-out model effectively stratified PRS, with 2-year PRS rates of 75.0%, 45.0%, and 15.0% for the low-, intermediate-, and high-risk groups, respectively, in 2007–2016 (p < 0.001, online suppl. Fig. 6C), and 90.0%, 70.0%, and 10.0%, respectively, in 2017–2023 (p < 0.001, online suppl. Fig. 6D).
Discussion
Although several studies have examined prognostic factors for long-term outcomes after thermal ablation, most have focused on survival following initial ablation rather than outcomes after post-ablation recurrence. As a result, recurrence patterns and post-recurrence outcomes after thermal ablation remain poorly characterized, and no prior study has developed or evaluated prognostic models specifically for survival following recurrence, particularly in patients whose recurrence falls within or beyond the Milan criteria. This represents a critical knowledge gap, as clinicians currently lack evidence-based tools to stratify prognosis and guide post-recurrence treatment decisions, including the selection of candidates for repeat locoregional therapy, downstaging, or salvage LT. In this long-term follow-up study, we analyzed recurrence patterns and outcomes after ablation and developed risk models to stratify PRS in patients with recurrence both within and beyond the Milan criteria, thereby providing a foundation for individualized post-recurrence treatment planning.
In this study, the overall cumulative recurrence rates at 1, 2, 5, and 10 years were 29.6%, 47.7%, 70.0%, and 82.8%, respectively, which are comparable to previously reported long-term recurrence rates following RFA [6, 7]. Notably, about two-thirds of the observed recurrences occurred within the first 2 years after ablation, with the recurrence rate gradually declining thereafter. This temporal pattern aligns with prior observations indicating that the recurrence risk of HCC after curative treatment decreases over time [19]. These findings underscore the importance of an intensified surveillance strategy, such as dynamic imaging every 3 months during the first 2 years post-ablation, followed by semiannual imaging, in line with surveillance protocols proposed in previous cost-effectiveness studies [20].
Regarding recurrent tumor burden, 84.5% of patients had recurrence at BCLC stage 0 or A, with 83.8% falling within the Milan criteria. Consequently, the majority of patients were eligible for a second curative treatment, such as resection (5.4%) or repeat ablation (66.0%). In line with previous reports [11, 21], no significant difference in PRS was observed between the resection and ablation groups, suggesting that either curative approach provides a comparable survival benefit for first recurrent HCC after initial ablation. Consistent with a previous report [22], we found that 16.2% of patients experienced recurrence beyond the Milan criteria despite regular post-ablation radiological surveillance. We identified baseline BCLC stage, AFP level, and bilirubin level as independent predictors of recurrence beyond the Milan criteria.
Current guidelines recommend salvage LT for patients with recurrent HCC within the Milan criteria [3, 23]. However, limitations such as organ shortage, long waiting times with associated risk of dropout, and surgical complications restrict the applicability of LT, particularly in countries with a high incidence of HCC. Moreover, survival outcomes in patients with recurrent HCC are highly variable. Selected patients with low tumor burden may still achieve long-term survival following a second curative treatment. Therefore, stratifying PRS is essential to guide individualized management strategies and to identify patients with an urgent need for salvage LT.
In this study, we identified early recurrence, tumor size, AFP level, ALBI grade, and FIB-4 score as independent predictors of PRS. Early recurrence of HCC, particularly within 2 years after initial curative treatment, is often associated with aggressive tumor biology and a poorer prognosis [24]. Tumor size and AFP level are well-established prognostic indicators across various tumor stages in HCC [25–27]. Additionally, ALBI grade and FIB-4 score, reflecting hepatic functional reserve and fibrosis severity, respectively, are important host-related factors associated with HCC outcomes [14, 15]. We also evaluated the impact of post-recurrence treatment and found that patients who received repeated curative therapy had significantly improved PRS in those with recurrence within the Milan criteria. Importantly, adjusting for curative treatment did not alter the significance of the original independent prognostic factors, indicating that the Milan-in model maintains its predictive accuracy regardless of subsequent therapy.
Based on these five variables, we developed a simple risk model for patients with recurrence within the Milan criteria, which stratified patients into four risk groups with corresponding median PRS of 103.7, 65.7, 48.7, and 28.6 months, respectively. In the low-risk group, the median PRS exceeding 8 years suggests excellent long-term survival, indicating that a “re-ablate and wait” strategy may be appropriate. In the intermediate-risk group, a median PRS of 65.7 months still reflects relatively favorable survival, supporting a “retreat and reassess” approach. In contrast, patients in the high- and very-high-risk groups exhibited significantly poorer outcomes, with median PRS of approximately 4 and 2 years, and corresponding 5-year PRS rates of 38.8% and 14.3%, respectively. To optimize long-term survival, patients in these higher risk categories may warrant expedited referral for salvage LT, even if liver-directed therapies appear adequate for managing recurrence within the Milan criteria.
Patients with recurrence beyond the Milan criteria typically present with more aggressive tumor behavior and poorer survival outcomes [28]. In this study, we identified macrovascular invasion, AFP, DM, ALBI grade, and FIB-4 score as predictors of PRS. Macrovascular invasion and AFP >400 ng/mL are well-established prognostic factors in patients with recurrent HCC [29]. In addition to ALBI grade and FIB-4 score, DM has also been reported as a negative host-related survival factor in patients undergoing thermal ablation [30, 31]. Based on these five variables, we developed a risk model for patients with recurrence beyond the Milan criteria, stratifying them into three risk groups with corresponding median PRS of 70.1 months, 24.7 months, and 8.1 months, respectively.
The prognostic factors identified in our study, such as tumor burden, AFP level, liver function, fibrosis severity, and recurrence timing, are consistent with previously recognized predictors of HCC outcomes, yet their relevance in the post-recurrence setting has not been well evaluated before. Most prior studies have focused on survival after initial ablation [7, 32], and only a few have specifically examined PRS after ablation. One previous study that included 103 patients reported a median PRS of 22 months and identified the Child-Pugh score, performance status, total tumor diameter at recurrence, and recurrence pattern as independent predictors of PRS [33]. However, that study was limited by its relatively small sample size and lack of detailed analysis of recurrence patterns and patient characteristics. Furthermore, to our knowledge, no prior work has specifically developed or externally validated prognostic models for survival after recurrence following curative ablation, particularly in relation to recurrence within or beyond the Milan criteria. By characterizing survival after recurrence and identifying key determinants of PRS, our study addresses this gap and provides clinically relevant insights to guide individualized post-recurrence treatment planning.
Among patients with recurrence beyond the Milan criteria, 36.8% achieved successful downstaging to within the Milan criteria, primarily through liver-directed therapies such as TACE or transarterial radioembolization, with or without sequential ablation. Consistent with previous studies [34], patients with recurrence within the UNOS-DS criteria had a higher likelihood of successful downstaging (62.1%), whereas those with recurrence beyond the UNOS-DS criteria still retained a modest chance (20.7%). Beyond stratifying PRS, the beyond Milan risk model also predicted the probability of successful downstaging, with downstaging rates of 66.7%, 39.3%, and 0% in the low-, intermediate-, and high-risk groups, respectively. Patients who achieve successful downstaging should be encouraged to undergo evaluation for salvage LT, with the goal of optimizing long-term survival.
Treatment strategies, including ablation techniques, imaging modalities, the introduction of immunotherapy, and second-line systemic therapies, have evolved over the past decade. Antiviral therapy for HBV and HCV also substantially impacts survival in patients with HCC [14, 35, 36], and in our cohort, patients who achieved virologic suppression (for HBV) or SVR (for HCV) had significantly better post-recurrence outcomes than untreated patients. The uptake of antiviral therapy increased markedly after 2017, and the availability of second-line systemic therapies for HCC during this period likely contributed to the observed improvement in PRS over time [37, 38]. Despite these temporal changes, the Milan-in and Milan-out models consistently demonstrated strong discriminative performance, supporting the robustness and clinical applicability of our prognostic framework.
This study has some limitations. First, this study was retrospective in nature, and although the proposed risk models for recurrence within and beyond the Milan criteria were externally validated in an independent cohort of 198 patients, future prospective studies are needed to further confirm their predictive performance and clinical utility. Second, the study cohort was enrolled prior to the widespread use of immunotherapy. PRS in patients who progress to advanced stage HCC may now be improved with effective systemic therapies. Recent advances in immunotherapy, particularly the combination of TACE and immune checkpoint inhibitors, have shown promise in enhancing downstaging success and enabling curative conversion [39–42]. Third, although salvage LT is recommended for patients at high risk of recurrence and those who achieve successful downstaging, survival data following salvage LT were not available in this study due to its underutilization at our country. Consequently, we were unable to directly assess the clinical benefit of our prognostic model on transplant outcomes. Additionally, our study did not evaluate the cost-effectiveness of applying this prognostic framework to guide post-recurrence management, which remains an important topic for future research. Nevertheless, our findings underscore the importance of timely referral for evaluation of salvage LT, particularly for high-risk patients, and provide clinically relevant guidance to optimize treatment prioritization and resource allocation.
In conclusion, PRS in patients with recurrent HCC after thermal ablation is significantly influenced by recurrence patterns, tumor characteristics, and host factors. We proposed simple, clinically applicable risk models to stratify PRS and the probability of successful downstaging in patients with recurrence both within and beyond the Milan criteria. These findings may help guide post-recurrence treatment decisions, optimize the timing of referral for salvage LT, and inform the design of future clinical trials.
Acknowledgment
The authors thank the Clinical Research Core Laboratory, Taipei Veterans General Hospital, for providing their facilities to conduct this study.
Statement of Ethics
Our study adhered to the prevailing ethical guidelines and principles of the Helsinki Declaration and received approval from the Institutional Review Board of Taipei Veterans General Hospital (IRB number: 2023-03-004CC) and the Institutional Review Board of Taichung Veterans General Hospital (IRB number: CE23139B-2). The Institutional Review Board waived the need for written informed consent due to the retrospective nature of the study.
Conflict of Interest Statement
The authors declare no conflicts of interest.
Funding Sources
The study was supported by grants from Taipei Veterans General Hospital, Taipei, Taiwan (V113C-157, V114C-132), and Ministry of Science and Technology, Taiwan (NSTC 112-2628-B-A49-016-MY3). The funding sources were involved in the establishment of the study database, statistical analysis, and the payment of publication-related fees. The funders had no role in the study design, data interpretation, manuscript conception, planning, writing, or the decision to submit the manuscript for publication.
Author Contributions
Tan C.P.: writing – original draft; visualization; data curation; and formal analysis. Lee I.C.: study concept and design; methodology; funding acquisition; data curation; analysis and interpretation of data; writing – review and editing; and supervision. Lee T.Y., Wu K.C., Liu C.A., Chiu N.C., Hsu S.J., Lee P.C., Wu C.J., Chi C.T., Luo J.C., Hou M.C., and Huang Y.H.: data curation.
Funding Statement
The study was supported by grants from Taipei Veterans General Hospital, Taipei, Taiwan (V113C-157, V114C-132), and Ministry of Science and Technology, Taiwan (NSTC 112-2628-B-A49-016-MY3). The funding sources were involved in the establishment of the study database, statistical analysis, and the payment of publication-related fees. The funders had no role in the study design, data interpretation, manuscript conception, planning, writing, or the decision to submit the manuscript for publication.
Data Availability Statement
The data that support the findings of this study are not publicly available due to privacy and ethical restrictions but are available from the corresponding author upon reasonable request.
Supplementary Material.
References
- 1. Tan DJH, Setiawan VW, Ng CH, Lim WH, Muthiah MD, Tan EX, et al. Global burden of liver cancer in males and females: changing etiological basis and the growing contribution of NASH. Hepatology. 2023;77(4):1150–63. [DOI] [PubMed] [Google Scholar]
- 2. Hwang SY, Danpanichkul P, Agopian V, Mehta N, Parikh ND, Abou-Alfa GK, et al. Hepatocellular carcinoma: updates on epidemiology, surveillance, diagnosis and treatment. Clin Mol Hepatol. 2025;31(Suppl l):S228–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Singal AG, Llovet JM, Yarchoan M, Mehta N, Heimbach JK, Dawson LA, et al. AASLD practice guidance on prevention, diagnosis, and treatment of hepatocellular carcinoma. Hepatology. 2023;78(6):1922–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. European Association for the Study of the Liver . EASL clinical practice guidelines on the management of hepatocellular carcinoma. J Hepatol. 2025;82(2):315–74. [DOI] [PubMed] [Google Scholar]
- 5. Shiina S, Tateishi R, Choi JI, Kim SY, Meng Z, Shen L, et al. Asian conference on tumor ablation guidelines for hepatocellular carcinoma. Liver Cancer. 2025;14(5):651–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Shiina S, Tateishi R, Arano T, Uchino K, Enooku K, Nakagawa H, et al. Radiofrequency ablation for hepatocellular carcinoma: 10-year outcome and prognostic factors. Am J Gastroenterol. 2012;107(4):569–78; quiz 78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Kim YS, Lim HK, Rhim H, Lee MW, Choi D, Lee WJ, et al. Ten-year outcomes of percutaneous radiofrequency ablation as first-line therapy of early hepatocellular carcinoma: analysis of prognostic factors. J Hepatol. 2013;58(1):89–97. [DOI] [PubMed] [Google Scholar]
- 8. Lee IC, Huang YH, Chau GY, Huo TI, Su CW, Wu JC, et al. Serum interferon gamma level predicts recurrence in hepatocellular carcinoma patients after curative treatments. Int J Cancer. 2013;133(12):2895–902. [DOI] [PubMed] [Google Scholar]
- 9. Taddei TH, Brown DB, Yarchoan M, Mendiratta-Lala M, Llovet JM. Critical update: AASLD practice guidance on prevention, diagnosis, and treatment of hepatocellular carcinoma. Hepatology. 2025;82(1):272–4. [DOI] [PubMed] [Google Scholar]
- 10. Lamarque C, Segaux L, Bachellier P, Buchard B, Chermak F, Conti F, et al. Evaluation of a delayed liver transplantation strategy for patients with HCC receiving bridging therapy: the DELTA-HCC study. J Hepatol. 2024;81(2):278–88. [DOI] [PubMed] [Google Scholar]
- 11. Zhong JH, Xing BC, Zhang WG, Chan AW, Chong CCN, Serenari M, et al. Repeat hepatic resection versus radiofrequency ablation for recurrent hepatocellular carcinoma: retrospective multicentre study. Br J Surg. 2021;109(1):71–8. [DOI] [PubMed] [Google Scholar]
- 12. Mehta N, Kelley RK, Yao FY. Refining the approach to down-staging of HCC prior to liver transplantation: patient selection, loco-regional treatments, and systemic therapies. Hepatology. 2024;80(1):238–53. [DOI] [PubMed] [Google Scholar]
- 13. Ahmed M, Solbiati L, Brace CL, Breen DJ, Callstrom MR, Charboneau JW, et al. Image-guided tumor ablation: standardization of terminology and reporting criteria--a 10-year update. Radiology. 2014;273(1):241–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Wu KC, Lee IC, Chi CT, Liu CA, Chiu NC, Hsu SJ, et al. Impact of HCV eradication on recurrence pattern and long-term outcomes in patients with HCV-related hepatocellular carcinoma undergoing radiofrequency ablation. Aliment Pharmacol Ther. 2024;60(7):940–52. [DOI] [PubMed] [Google Scholar]
- 15. Lee IC, Hung YW, Liu CA, Lee RC, Su CW, Huo TI, et al. A new ALBI-based model to predict survival after transarterial chemoembolization for BCLC stage B hepatocellular carcinoma. Liver Int. 2019;39(9):1704–12. [DOI] [PubMed] [Google Scholar]
- 16. American Diabetes Association Professional Practice Committee . 2. Classification and diagnosis of diabetes: standards of medical care in Diabetes-2022. Diabetes Care. 2022;45(Suppl 1):S17–38. [DOI] [PubMed] [Google Scholar]
- 17. Kudo M, Kawamura Y, Hasegawa K, Tateishi R, Kariyama K, Shiina S, et al. Management of hepatocellular carcinoma in Japan: JSH consensus statements and recommendations 2021 update. Liver Cancer. 2021;10(3):181–223. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Hung YW, Lee IC, Chi CT, Lee RC, Liu CA, Chiu NC, et al. Radiologic patterns determine the outcomes of initial and subsequent transarterial chemoembolization in intermediate-stage hepatocellular carcinoma. Liver Cancer. 2024;13(1):29–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Lee HA, Lee YS, Kim BK, Jung YK, Kim SU, Park JY, et al. Change in the recurrence pattern and predictors over time after complete cure of hepatocellular carcinoma. Gut Liver. 2021;15(3):420–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Chen QF, Chen S, Yi JZ, Wang JL, Zhong SX, Jiang XY, et al. Recommended 10-Year Follow-Up strategy for small hepatocellular carcinoma after radiofrequency ablation: a cost-effectiveness evaluation. Am J Gastroenterol. 2024;119(10):2052–60. [DOI] [PubMed] [Google Scholar]
- 21. Xia Y, Li J, Liu G, Wang K, Qian G, Lu Z, et al. Long-term effects of repeat hepatectomy vs percutaneous radiofrequency ablation among patients with recurrent hepatocellular carcinoma: a randomized clinical trial. JAMA Oncol. 2020;6(2):255–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Doyle A, Gorgen A, Muaddi H, Aravinthan AD, Issachar A, Mironov O, et al. Outcomes of radiofrequency ablation as first-line therapy for hepatocellular carcinoma less than 3 Cm in potentially transplantable patients. J Hepatol. 2019;70(5):866–73. [DOI] [PubMed] [Google Scholar]
- 23. Mehta N, Bhangui P, Yao FY, Mazzaferro V, Toso C, Akamatsu N, et al. Liver transplantation for hepatocellular Carcinoma. Working group report from the ILTS transplant oncology consensus conference. Transplantation. 2020;104(6):1136–42. [DOI] [PubMed] [Google Scholar]
- 24. Kang TW, Lim HK, Cha DI. Aggressive tumor recurrence after radiofrequency ablation for hepatocellular carcinoma. Clin Mol Hepatol. 2017;23(1):95–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Hung YW, Lee IC, Chi CT, Lee RC, Liu CA, Chiu NC, et al. Redefining tumor burden in patients with intermediate-stage hepatocellular carcinoma: the seven-eleven criteria. Liver Cancer. 2021;10(6):629–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Tsilimigras DI, Moris D, Hyer JM, Bagante F, Ratti F, Marques HP, et al. Serum alpha-Fetoprotein levels at time of recurrence predict post-recurrence outcomes following resection of hepatocellular carcinoma. Ann Surg Oncol. 2021;28(12):7673–83. [DOI] [PubMed] [Google Scholar]
- 27. Lee IC, Lei HJ, Wang LC, Yeh YC, Chau GY, Hsia CY, et al. M2BPGi correlated with immunological biomarkers and further stratified recurrence risk in patients with hepatocellular carcinoma. Liver Cancer. 2025;14(1):68–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Fuster-Anglada C, Mauro E, Ferrer-Fàbrega J, Caballol B, Sanduzzi-Zamparelli M, Bruix J, et al. Histological predictors of aggressive recurrence of hepatocellular carcinoma after liver resection. J Hepatol. 2024;81(6):995–1004. [DOI] [PubMed] [Google Scholar]
- 29. Moazzam Z, Alaimo L, Endo Y, Lima HA, Woldesenbet S, Rueda BO, et al. A prognostic model to predict survival after recurrence among patients with recurrent hepatocellular carcinoma. Ann Surg. 2024;279(3):471–8. [DOI] [PubMed] [Google Scholar]
- 30. Nakatsuka T, Tateishi R. Development and prognosis of hepatocellular carcinoma in patients with diabetes. Clin Mol Hepatol. 2023;29(1):51–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Wu KC, Lee IC, Liu CA, Chiu NC, Hsu SJ, Lee PC, et al. Diabetes mellitus negatively impacts outcomes of HBV-related hepatocellular carcinoma following thermal ablation. J Hepatocell Carcinoma. 2024;11:2257–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Kao WY, Su CW, Chiou YY, Chiu NC, Liu CA, Fang KC, et al. Hepatocellular carcinoma: nomograms based on the albumin-bilirubin grade to assess the outcomes of radiofrequency ablation. Radiology. 2017;285(2):670–80. [DOI] [PubMed] [Google Scholar]
- 33. Facciorusso A, Del Prete V, Antonino M, Crucinio N, Neve V, Di Leo A, et al. Post-recurrence survival in hepatocellular carcinoma after percutaneous radiofrequency ablation. Dig Liver Dis. 2014;46(11):1014–9. [DOI] [PubMed] [Google Scholar]
- 34. Tan DJH, Lim WH, Yong JN, Ng CH, Muthiah MD, Tan EX, et al. UNOS down-staging criteria for liver transplantation of hepatocellular carcinoma: systematic review and meta-analysis of 25 studies. Clin Gastroenterol Hepatol. 2023;21(6):1475–84. [DOI] [PubMed] [Google Scholar]
- 35. Lee TY, Lin JT, Zeng YS, Chen YJ, Wu MS, Wu CY. Association between nucleos(t)ide analog and tumor recurrence in hepatitis B virus-related hepatocellular carcinoma after radiofrequency ablation. Hepatology. 2016;63(5):1517–27. [DOI] [PubMed] [Google Scholar]
- 36. Lee IC, Lei HJ, Chau GY, Yeh YC, Wu CJ, Su CW, et al. Predictors of long-term recurrence and survival after resection of HBV-related hepatocellular carcinoma: the role of HBsAg. Am J Cancer Res. 2021;11(7):3711–25. [PMC free article] [PubMed] [Google Scholar]
- 37. Lee IC, Chao Y, Lee PC, Chen SC, Chi CT, Wu CJ, et al. Determinants of survival and post-progression outcomes by Sorafenib-Regorafenib sequencing for unresectable hepatocellular carcinoma. Cancers. 2022;14(8):2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Lee IC, Lee PC, Chao Y, Chi CT, Wu CJ, Hung YP, et al. Application and impact of antiviral therapy for patients with HBV-related hepatocellular carcinoma receiving sorafenib and lenvatinib treatment. Viruses. 2022;14(11):2355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Kudo M, Aoki T, Ueshima K, Tsuchiya K, Morita M, Chishina H, et al. Achievement of complete response and drug-free status by Atezolizumab plus Bevacizumab combined with or without curative conversion in patients with transarterial chemoembolization-unsuitable, intermediate-stage hepatocellular carcinoma: a multicenter proof-of-concept study. Liver Cancer. 2023;12(4):321–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Kudo M, Ren Z, Guo Y, Han G, Lin H, Zheng J, et al. Transarterial chemoembolisation combined with lenvatinib plus pembrolizumab versus dual placebo for unresectable, non-metastatic hepatocellular carcinoma (LEAP-012): a multicentre, randomised, double-blind, phase 3 study. Lancet. 2025;405(10474):203–15. [DOI] [PubMed] [Google Scholar]
- 41. Lee IC, Wang HW, Teng W, Lin TJ, Chen CH, Lai HC, et al. Taiwan liver cancer association management consensus guidelines for intermediate-stage hepatocellular carcinoma. Clin Mol Hepatol. 2025;31(4):1213–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Sangro B, Kudo M, Erinjeri JP, Qin S, Ren Z, Chan SL, et al. Durvalumab with or without bevacizumab with transarterial chemoembolisation in hepatocellular carcinoma (EMERALD-1): a multiregional, randomised, double-blind, placebo-controlled, phase 3 study. Lancet. 2025;405(10474):216–32. [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 data that support the findings of this study are not publicly available due to privacy and ethical restrictions but are available from the corresponding author upon reasonable request.