Long-term outcomes of thermal ablation vs surgical resection for single small hepatocellular carcinoma

Abstract

BACKGROUND

Thermal ablation (TA), including radiofrequency ablation and microwave ablation, is a commonly used curative treatment for single small hepatocellular carcinoma (sHCC). The relative advantages of TA and surgical resection (SR) in terms of long-term survival remain controversial.

AIM

To compare their long-term efficacy in this patient population.

METHODS

This population-based retrospective cohort study included 257 patients who received a first diagnosis of single sHCC and underwent SR or TA from January 2012 to September 2017. The primary endpoints were overall survival (OS) and recurrence-free survival (RFS).

RESULTS

The average follow-up duration was 11.4 years. The 1-, 3-, 5-, and 10-year OS rates were 95.8%, 86.0%, 82.5%, and 74.2% in the SR group vs 97.4%, 85.8%, 78.6%, and 65.6% in the TA group, with the median OS not yet reached. The 1-, 3-, 5-, and 10-year RFS rates were 79.8%, 59.6%, 46.2%, and 24.7% in the SR group vs 83.9%, 61.5%, 47.9%, and 41.2% in the TA group, with median RFS values of 3.95 and 4.63 years, respectively. No significant differences in OS or RFS were observed overall (OS: P = 0.244; RFS: P = 0.180), but in patients ≤ 60 years, TA led to a higher RFS than SR (P = 0.021). Multivariate analysis identified age, tumor differentiation grade, and Child-Pugh classification as independent risk factors for OS, whereas age and differentiation grade were significant risk factors for RFS.

CONCLUSION

In patients with single sHCC, SR, and TA offered comparable long-term efficacy. However, TA showed superior RFS in patients ≤ 60 years, suggesting that TA may be a reasonable option for younger patients, pending confirmation by prospective studies.

Key Words: Small hepatocellular carcinoma; Thermal ablation; Surgical resection; Survival analysis; Efficacy

Core Tip: This study evaluated the long-term outcomes of patients with single small hepatocellular carcinoma treated with thermal ablation (TA) or surgical resection. Overall, TA and surgical resection achieved comparable long-term outcomes; however, among patients aged 60 years or younger, TA was associated with better recurrence-free survival. This suggests that for younger patients seeking less invasiveness and faster recovery, TA is not merely an alternative - it may offer an advantage in reducing recurrence. Treatment decisions should still integrate tumor differentiation, liver function, and the experience of the clinical team.

INTRODUCTION

Primary liver cancer ranks sixth in incidence and third in cancer-related mortality worldwide. Hepatocellular carcinoma (HCC) accounts for approximately 75%-85% of cases[1,2]. For patients with adequate liver function reserve and good overall health, surgical resection (SR)[3] is the preferred treatment recommended by international guidelines[4-8], yet only 9%-27% of patients are eligible for surgery[9]. Thermal ablation (TA), most commonly in the forms of radiofrequency ablation (RFA) and microwave ablation (MWA), is a local, potentially curative therapy that induces coagulative necrosis. For cases that are unsuitable for resection, TA has become the preferred radical option[10,11].

The efficacy of SR vs TA in HCC remains debated[9,12-16]. Prior comparative studies had the following limitations: Insufficient follow-up, heterogeneous patient populations (e.g., multiple tumors or prior anti-tumor treatments), and an RFA-dominant evidence base that does not reflect contemporary real-world use of TA modalities[17]. To address these limitations, we focused on a highly homogeneous cohort of newly diagnosed, previously untreated patients with solitary small HCC (sHCC) with a maximum diameter of ≤ 3 cm. We conducted a retrospective cohort study with long-term follow-up comparing the outcomes of SR and TA with respect to overall survival (OS) and recurrence-free survival (RFS). We also evaluated perioperative and periablation safety and analyzed prognostic factors. We aimed to provide evidence for first-line management of solitary sHCC, hypothesizing that TA is noninferior to SR with respect to long-term survival.

MATERIALS AND METHODS

Patients and methods

The study was approved by the Medical Ethics Committee of the Cancer Hospital of the Chinese Academy of Medical Sciences (approval No. 25/017-4963), and the requirement for informed consent was waived. Patients who were initially diagnosed with single sHCC and underwent either SR or TA as their initial treatment from January 2012 to September 2017 at the Cancer Hospital, Chinese Academy of Medical Sciences were included in this study. The inclusion criteria for this study were: (1) A pathologically confirmed diagnosis of HCC; (2) A solitary tumor with a maximum diameter of ≤ 3 cm; (3) Child-Pugh A/B grade; (4) Eastern Cooperative Oncology Group performance status score of 0-1; and (5) Age 18-95 years. The exclusion criteria were: (1) Any major vascular invasion or adjacent organ involvement on pre-treatment imaging evaluation; (2) Presence of distant metastases; (3) Any concurrent malignancy; (4) Any previous anti-cancer treatment; and (5) Incomplete medical data.

Data collection and study endpoints

Baseline data for all patients were retrospectively collected from the hospital’s electronic records, including gender, age, maximum tumor diameter, Child-Pugh grade, tumor differentiation (obtained from pathology reports), alpha-fetoprotein level, hepatitis history, surgical method, ablation method, transfusion status, operation time, postoperative hospital stay, and postoperative complications.

Within 3 months postoperatively, all patients underwent contrast-enhanced computed tomography or magnetic resonance imaging, and treatment efficacy was assessed using the modified response evaluation criteria in solid tumors to confirm complete resection or ablation of the lesion. Recurrence and survival data were obtained from hospital electronic records or through follow-up. For patients with available medical records, data were extracted directly from the hospital database. For those without accessible records, follow-up was conducted systematically. The hospital medical records department performed centralized follow-up and documentation at the 5-year timepoint. Beyond 5 years, patients were contacted at least annually via telephone to update recurrence and survival status.

The primary endpoints of the study were OS and RFS, while the secondary endpoints included treatment safety (e.g., incidence of severe complications, transfusion rate, and postoperative hospital stay). RFS was defined as the time from the index treatment (SR or TA) to the first radiologic or pathologic confirmation of recurrence. Patients without a documented recurrence were right-censored at the date of their last disease assessment. If death occurred before any documented recurrence, observations were censored at the date of death. Patients without any post-baseline tumor assessment were pre-specified as excluded from the RFS analysis set.

Statistical analysis

All data were analyzed using R version 4.4.2. The normality of quantitative data was assessed with the Kolmogorov-Smirnov test. Differences in normally distributed data were analyzed with the independent samples t-test, while differences in non-normally distributed data were analyzed with the Wilcoxon signed-rank test. Differences in categorical data were analyzed using the χ² test. Survival analysis was performed using the Kaplan-Meier method, with differences evaluated by the log-rank test. Univariate and multivariate Cox regression analyses were conducted to identify factors influencing OS and RFS, with results reported as hazard ratios (HRs) and corresponding 95% confidence intervals (CIs). Subgroup analysis was conducted to explore prognostic differences among distinct patient cohorts. Statistical significance was defined as a two-tailed P < 0.05.

RESULTS

Clinical characteristics of patients

A total of 257 patients were included in this study, among whom 143 underwent SR and 114 underwent TA as primary treatment for single sHCC (Figure 1). In the SR group, 87.41% of patients (125/143) underwent traditional open surgery, while 12.59% (18/143) underwent laparoscopic surgery. In the TA group, 33.33% of patients (38/114) received RFA and 66.67% (76/114) were treated with MWA. The mean age of patients in the TA group was significantly higher than that of patients in the SR group (60.27 years vs 54.03 years, P < 0.001), with a notably higher number of elderly patients (≥ 80 years) in the TA group compared with the SR group (6 cases vs 0 cases). Although sHCC was pathologically confirmed for all patients, more than half of the patients (51.8%) in the TA group lacked an accurate differentiation grading, compared to only 3.5% in the SR group. No differences in any other clinical characteristics were observed between the two groups. The detailed baseline data for each group are presented in Table 1.

Figure 1 Patient screening flowchart. HCC: Hepatocellular carcinoma; ECOG: Eastern Cooperative Oncology Group; SR: Surgical resection; TA: Thermal ablation.

Table 1 Baseline characteristics and periprocedural metrics of patients with single small hepatocellular carcinoma, n (%)/mean ± SD.

		Surgical resection group, n = 143	Thermal ablation group, n = 114	P value	Test
Age (years)		54.0 ± 9.2	60.3 ± 10.2	< 0.001	T
	≤ 60	109 (76.2)	62 (54.4)
	60-80	34 (23.8)	46 (40.4)
	≥ 80	0 (0)	6 (5.3)
Sex	Male	119 (83.2)	85 (74.6)	0.121	χ2
	Female	24 (16.8)	29 (25.4)
Residence	Urban	53 (37.1)	53 (46.5)	0.127	χ2
	Rural	90 (62.9)	61 (53.5)
BMI		24.91 ± 3.4	24.75 ± 3.4	0.709	T
AFP (ng/mL)	≤ 20	81 (56.6)	72 (63.2)	0.252	χ2
	20-400	35 (24.5)	29 (25.4)
	> 400	27 (18.9)	13 (11.4)
HBV	Positive	115 (80.4)	92 (80.7)	1	χ2
	Negative	28 (19.6)	22 (19.3)
HCV	Positive	10 (7.0)	12 (10.5)	0.435	χ2
	Negative	133 (93.0)	102 (89.5)
Child-Pugh class	A	140 (97.9)	108 (94.7)	0.191	Fisher exact
	B	3 (2.1)	6 (5.3)
Hypertension	Yes	33 (23.1)	37 (32.5)	0.124	χ2
	No	110 (76.9)	77 (67.5)
Diabetes	Yes	21 (14.7)	27 (23.7)	0.093	χ2
	No	122 (85.3)	87 (76.3)
Alcohol	Yes	58 (40.6)	47 (41.2)
	No	85 (59.4)	67 (58.8)	1	χ2
Tumor size (cm)		2.3 ± 0.5	2.3 ± 0.8	0.524	T
Differentiation grade	Well-differentiated	19 (13.3)	33 (28.9)	< 0.001	χ2
	Moderately-differentiated	89 (62.2)	13 (11.4)
	Poorly-differentiated	30 (21.0)	9 (7.9)
	Unknown	5 (3.5)	59 (51.8)
Operation time, minute, median (IQR)		175.0 (133.0, 212.0)	85.0 (70.0, 101.3)	< 0.001	Wilcoxon
Hospital stay, days, median (IQR)		14.0 (13.0, 17.0)	7.0 (8.0, 11.0)	< 0.001	Wilcoxon
Blood transfusion	Yes	30 (21.0)	0 (0)	< 0.001	χ2
	No	113 (79.0)	114 (100.0)
Blood transfusion volume, IU/U, median (IQR)		4.0 (2.0, 5.8)	0 (0,0)	< 0.001	Wilcoxon

BMI: Body mass index; AFP: Alpha-fetoprotein; HBV: Hepatitis B virus; HCV: Hepatitis C virus; IQR: Interquartile range.

Analysis of efficacy and prognostic factors

Treatment-related parameters: In the SR group, 30 patients (21%) received intraoperative or postoperative blood transfusion, with a mean transfusion volume of 4.0 U. No transfusion events occurred in the TA group, and the difference in the percentage of patients who received transfusion between the groups was statistically significant (P < 0.001). Post-treatment, the median postoperative hospital stay was significantly longer in the SR group (8 days) than in the TA group (5 days, P < 0.001). In the SR group, 8 patients (5.6%) experienced serious complications (Clavien-Dindo grade ≥ III[18]), including 4 cases of ascites with abdominal infection or bile leakage, 2 cases of pleural infection with pleural effusion or atrial fibrillation, 1 case of anaerobic sepsis with intestinal fistula, and 1 case of intestinal obstruction. In the TA group, only 1 patient (0.9%) developed postoperative pneumothorax. The difference in the incidence of complications between the groups was statistically significant (P = 0.041).

Survival and recurrence outcomes: In the SR group, 3 patients (2.1%) experienced early recurrence within 3 months postoperatively, whereas 5 patients (4.3%) in the TA group had early recurrence. Among these, 3 cases (2.6%) in the TA group were classified as incomplete ablation, but complete ablation was achieved upon reassessment 1 month later. The mean follow-up duration was 11.4 years in both groups (P = 0.88). The respective 1-, 3-, 5-, and 10-year OS rates were 95.8% (95%CI: 92.6%-99.1%), 86.0% (95%CI: 80.5%-91.9%), 82.5% (95%CI: 76.4%-88.9%), and 74.2% (95%CI: 67.1%-82.0%) after SR, and 97.4% (95%CI: 94.4%-100.0%), 85.8% (95%CI: 79.6%-92.5%), 78.6% (95%CI: 71.3%-86.6%), and 65.6% (95%CI: 55.8%-77.0%) after TA; the median OS was not reached in either group. The respective 1-, 3-, 5-, and 10-year RFS rates were 79.8% (95%CI: 72.3%-88.1%), 59.6% (95%CI: 50.7%-70.1%), 46.2% (95%CI: 37.3%-57.2%), and 24.7% (95%CI: 16.8%-36.4%) after SR, and 83.9% (95%CI: 77.0%-91.5%), 61.5% (95%CI: 52.5%-72.1%), 47.9% (95%CI: 38.7%-59.3%), and 41.2% (95%CI: 32.0%-53.1%) after TA (Figure 2). The median RFS was 3.95 years (95%CI: 3.04-6.75) for SR and 4.63 years (95%CI: 3.77 to not estimable) for TA. In multivariable Cox models with TA as the reference, SR showed a HR of 1.324 (95%CI: 0.825-2.126; P = 0.244) for OS and a HR of 0.785 (95%CI: 0.550-1.119; P = 0.180) for RFS, consistent with the absence of a statistically significant between-group difference.

Figure 2 Kaplan-Meier curves for overall survival and recurrence-free survival in the surgical resection and thermal ablation groups. A: Overall survival; B: Recurrence-free survival. OS: Overall survival; RFS: Recurrence-free survival.

Potential prognostic factors: Univariate Cox regression analysis identified four potential influencing factors for OS: Age, tumor diameter, pathological differentiation grade, and Child-Pugh classification. Multivariate analysis further indicated that age, perioperative blood transfusion, pathological differentiation, and Child-Pugh grade were independent prognostic factors for OS. On the univariate analysis for RFS, age was identified as a potential influencing factor (Figure 3). Multivariate analysis further identified both age and pathological differentiation as independent risk factors for recurrence.

Figure 3 Multivariable Cox regression forest plots for overall survival and recurrence-free survival (thermal ablation as reference). A: Overall survival; B: Recurrence-free survival. OS: Overall survival; TA: Thermal ablation; RFS: Recurrence-free survival; AFP: Alpha-fetoprotein; AIC: Akaike information criterion.

Influence of age on survival outcomes: To test the influence of age on OS and RFS, patients were divided into two subgroups based on age: ≤ 60 years and > 60 years (Figure 4). The analysis revealed that for patients aged > 60 years, no significant differences in OS and RFS were detected between the two treatment groups (OS: P = 0.250; RFS: P = 0.875). Consistently, in the subgroup aged ≤ 60 years (109 patients in the SR group and 62 patients in TA group), no difference in OS was observed between the two groups. However, a significant difference in RFS was detected between the treatment groups for the younger patient subgroup (P = 0.021). The 1-, 3-, 5-, and 10-year RFS rates for the TA subgroup were 80.8%, 66.8%, 60.5% and 58.4%, respectively, and those for the SR subgroup were 78.9%, 61.8%, 49.8% and 26.8%, respectively. To examine the clinical factors that could contribute to this difference in RFS between treatment groups only among patients ≤ 60 years, we compared the clinical characteristics between the treatment subgroups. This analysis showed that among the patient subgroup aged ≤ 60 years, the TA subgroup had a higher proportion of well-differentiated tumors (29.0%) compared with the SR subgroup (9.17%), although more than half of cases (51.6%) in the TA subgroup lacked a definitive pathological grade.

Figure 4 Subgroup analysis of overall survival and recurrence-free survival. A: Overall survival for patients aged ≤ 60 years; B: Recurrence-free survival for patients aged ≤ 60 years; C: Overall survival for patients aged > 60 years; D: Recurrence-free survival for patients aged > 60 years. OS: Overall survival; RFS: Recurrence-free survival.

DISCUSSION

The comparative effectiveness of SR vs TA for solitary sHCC remains uncertain[19]. The present study was a single-center, large-sample cohort with long-term follow-up (mean: 11.4 years) of newly diagnosed, previously untreated patients with solitary sHCC. All cases were pathologically confirmed. Both RFA and MWA were included, and adjunct techniques were used when indicated, reflecting real-world TA practice. The consecutive and clinically homogeneous cohort enabled robust long-term comparisons of SR and TA for OS, RFS, and safety, including 10-year outcomes.

In this study, the 5-year OS and RFS rates in the TA group were 78.6% and 47.9%, respectively, compared with 82.5% and 46.2% in the SR group, with no significant differences between the groups - findings consistent with previous randomized controlled trials[17,20,21]. The outcomes in the ablation group were superior to previously reported 5-year OS (50.0%-66.7%) and RFS (20.0%-46.3%) rates[17,22,23], which may be attributable to the use of an ablation safety margin greater than 1.0 cm in this study[24].

The incidence rates of severe postoperative complications were 5.6% in the SR group and 0.9% in the TA group, which are consistent with the incidence rates reported by previous studies and still relatively low[25-27]. These low complication rates may be attributed to the inclusion of only solitary small lesions in our study. The absence of cases requiring blood transfusion and the shorter postoperative hospital stay in the TA group further highlight the minimally invasive nature of TA. Additionally, three cases of incomplete ablation were successfully treated with further ablation, underscoring the high repeatability of TA[26].

Multivariate Cox regression analysis confirmed that treatment modality (SR vs TA) had no significant effect on OS and RFS. However, age, perioperative blood transfusion, pathological differentiation, and Child-Pugh classification were found to be independent risk factors for reduced OS, while age and pathological differentiation were independent risk factors for reduced RFS.

In this study, patients in the TA group were significantly older than those in the SR group, with a notably higher proportion of elderly patients (≥ 80 years). Despite this, the two groups showed comparable long-term outcomes. These findings suggest that ablation may serve as the preferred curative treatment option for elderly patients or those unable to tolerate surgery. In the subgroup < 60 years, the TA group had a significantly superior RFS compared with the SR group (P = 0.021), suggesting that younger patients may derive greater benefit from the minimally invasive approach and reduced systemic trauma of TA. The TA group was older at baseline, an unfavorable prognostic factor for survival outcomes[28-30]. Despite this disadvantage, no significant differences in OS or RFS were observed between the groups treated with TA and SR. Accordingly, we consider TA comparable to SR for solitary sHCC, and that this conclusion is conservative rather than an overstatement of TA’s effectiveness.

Perioperative blood transfusion has been established as an independent risk factor for diminished OS[31,32]. Blood transfusion indicates greater surgical invasiveness and is linked to an increased risk of postoperative infection, hepatic impairment, and alloimmune reactions[33]. These complications may compromise the survival benefit. Moreover, transfusion may increase the risk of tumor recurrence and negatively impact long-term survival by inducing immunosuppression and promoting tumor angiogenesis[34,35]. In this study, no patients in the TA group required transfusion, thereby avoiding transfusion-related adverse effects and highlighting the minimally invasive nature of TA.

The histopathological differentiation grade of HCC serves as a significant predictor for both OS and RFS. Poorly differentiated tumors are better able to adapt to their microenvironment and typically exhibit stronger proliferative abilities, invasiveness, and metastatic tendencies, which increase the risk of residual or recurrent tumor cells after treatment[36,37], ultimately affecting long-term survival. Therefore, for patients with poorly differentiated HCC tumors, even if still small at first diagnosis, close follow-up and aggressive intervention strategies are warranted.

The Child-Pugh classification directly reflects liver function and treatment tolerance, serving as a crucial indicator for predicting the prognosis of liver cancer patients[38]. Patients with liver dysfunction often exhibit reduced tolerance to surgery or ablation treatments, and the rate of postoperative adjuvant therapy tends to be lower[39,40], which could further influence their survival outcomes. The minimally invasive nature of TA results in less liver function impairment than SR, reducing the prognostic burden of hepatic injury.

The results of the present study suggest that in clinical practice, in addition to the choice of treatment modality, the patient’s baseline characteristics and indicators of tumor biological behavior also should be fully considered. For example, elderly patients or those with Child-Pugh B/C status may benefit more from TA due to its minimally invasive nature, while patients with poorly differentiated tumors may require more aggressive treatment approaches and closer follow-up.

This study has several limitations. First, as a single-center retrospective cohort analysis, it is prone to selection and information bias, and unmeasured confounding cannot be excluded. Second, because pathologic sampling before ablation yields limited tissue, some patients in the TA group lacked histologic differentiation; this constrained comparisons of tumor biology. Third, the ≤ 60-year subgroup in which TA appeared to lower the RFS hazard was underpowered (TA, n = 60). Approximately half of these TA cases had unknown differentiation, leaving about 30 graded cases. This small effective sample increases estimate instability and makes the result sensitive to a few events. Subgroup comparisons were not matched, so residual confounding and selection bias remain possible. Younger patients selected for TA may have had unmeasured favorable features (e.g., easier lesion access or operator preference) that were not captured. Alpha-fetoprotein levels were broadly similar between SR and TA in this subgroup, which reduces - but does not eliminate - concern about systematic selection of less aggressive tumors. Fourth, the TA method was not technically uniform: Both RFA and MWA were used, and adjunct techniques were applied at operator discretion. This heterogeneity may influence ablation margins, thermal dose, and local control, and could attenuate or inflate the observed differences between TA and SR. Prospective, preferably multicenter studies with standardized TA protocols (preset device settings, predefined ablation margins, protocolized use of adjuncts), uniform pathology sampling and reporting, and prespecified analyses stratified by age and TA modality are needed to address these gaps.

CONCLUSION

The results of the present study demonstrate that the long-term efficacy of TA is comparable to that of SR in patients with single sHCC. Additionally, TA offers significant minimally invasive advantages over SR. Therefore, in clinical practice, TA is preferentially recommended for elderly patients and those with poor liver function reserve. In patients with sHCC aged ≤ 60 years, TA resulted in a higher RFS compared with SR.