Short- and Long-Term Outcomes of Radiofrequency Ablation for Early Hepatocellular Carcinoma with Liver Dysfunction

Introduction

Liver cancer is the third most prevalent cause of cancer-related mortality worldwide, with hepatocellular carcinoma (HCC) being the most common form.^1,2 HCC is often associated with cirrhosis, which is also a disease that can be fatal. The long-term survival of patients with HCC is contingent not only on the characteristics of the tumor itself but also on the degree of liver functional reserve.^3,4 Among the various models for non-invasive evaluation of liver function, the Child-Pugh (CP) classification is the most widely used liver function grading system and is adopted by the majority of HCC treatment guidelines.^5,6 Liver function is categorized into class A, B, and C according to the CP classification.^7,8

Patients with early stage HCC and CP class B cirrhosis are generally not considered good candidates for surgical resection. The same is true for patients with CP class C cirrhosis.⁹ In such situations, thermal ablation, represented by radiofrequency ablation (RFA), is a good choice because of its excellent local tumor control and minimally invasive nature, even if the patient is unsuitable for surgical resection due to poor liver function. Jeon et al showed that the five-year overall survival (OS) rate of early HCC patients with CP class B cirrhosis who underwent radical ablation was 41%, which was higher than of patients who received palliative treatment (five-year OS rate: 28% for chemoembolization, 20% for systemic treatment, and 33% for radiotherapy).¹⁰

Previous studies on patients with liver dysfunction were limited, with existing research primarily consisting of small samples and short follow-up periods. Few studies have further analyzed the long-term survival differences among patients with different levels of liver dysfunction following RFA. Therefore, this study reports our center’s decade-long experience with RFA in HCC patients with liver dysfunction (Child-Pugh score ≥ 7 points) and evaluates the safety and efficacy of this treatment in such patients. What’s more, we observed and compared the impact of varying degrees of liver dysfunction on the long-term outcomes.

Method and Material

Patients

This retrospective, single-center study was approved by the Ethics Committee of the First Affiliated Hospital of Army Medical University (No. (B) KY2024182) and conducted in accordance with the Helsinki Declaration, all personal identifying information was anonymized before analysis to protect patient privacy. Informed consent was waived by the Ethics Committee of the First Affiliated Hospital of Army Medical University due to the retrospective nature of the study and anonymized data.

From January 2013 to July 2023, 5360 consecutive patients with HCC who underwent ultrasound-guided percutaneous RFA treatment at the First Affiliated Hospital of Army Medical University and 176 HCC patients with Child-Pugh score (CPS) ≥7 were included in this study (Figure 1). The inclusion criteria were as follows: 1) gender-neutral and aged 18 years or older; 2) diagnosis of HCC based on the diagnostic criteria outlined in the American Association for the Study of Liver Diseases (AASLD) Guidelines for the Diagnosis and Treatment of Liver Cancer and the Chinese Guidelines for the Diagnosis and Treatment of Liver Cancer (2022 Edition); 3). A preoperative liver function CPS ≥ 7. The exclusion criteria were as follows: 1) single tumor diameter >5 cm, multiple tumors >3, or maximum tumor diameter >3 cm; 2) combined with other tumors; 3) lost to follow-up (medical records or follow-up time ≤3 months with no outcome events).

Figure 1 Flowchart shows patient selection. Abbreviations: HCC, hepatocellular carcinoma; RFA, radiofrequency ablation.

To further identify the influence of the level of liver dysfunction on Recurrence-free survival (RFS) and Overall survival (OS), we divided all patients into different groups according to different liver function grading systems: 4 groups according to their CPS (7, 8, 9, and 10&11), 3 groups according to their albumin-bilirubin (ALBI) grade (grade 1, 2, and 3), and 3 groups according to their Model for End-stage Liver Disease (MELD) score (<8, 8–12 and >12).

RFA Procedure

All RFA procedures were performed by experienced hepatobiliary surgeons at our center under the guidance of real-time ultrasound guidance. The tumor location was first identified, and the optimal needle path, avoiding large blood vessels, bile ducts, and vital organs, was selected using contrast-enhanced ultrasound (CE-US) before the operation. The puncture point was then marked, and injected with 2% lidocaine for local infiltration anesthesia. Under ultrasound guidance, the largest section of the tumor was selected and one or two electrode needles were punctured directly into the lesion. The ablation power was set at 120 watts and the working time was 3 min. When the tumor was larger than 3 cm and smaller than 5 cm, two ablation needles were used for puncture and ablation. The ablation power was 180 watts and the working time was 5 min. Intraoperative ultrasound was used to monitor the ablation process in real time. Needle-track ablation was performed when the electrode was withdrawn. Oxygen was administered during the procedure and vital signs (blood pressure, heart rate, and blood oxygen saturation) were monitored. CE-US was performed again 20 minutes after the procedure; if any tumor residue was found, ablation was performed again immediately to achieve complete ablation. After treatment, the puncture point was bandaged and the patient returned to the ward. In the evening, ultrasound examination of the chest and abdomen was performed to check for hydrothorax, ascites, or hemorrhage. The day after the operation, serological tests including complete blood count, liver and renal function, coagulation tests, serum electrolytes, and high-sensitivity C-reactive protein were re-examined. On the second postoperative day, CU-US or contrast-enhanced computed tomography (CT) was performed again to ensure complete ablation.

Post-RFA Follow-Up

One month after RFA, contrast-enhanced CT or Primovist-enhanced magnetic resonance imaging (MRI) was performed to check for any abnormal enhancement around the ablation zone. All patients were then advised to attend follow-up every three months for the first five years after RFA and every six months for the following five years, until tumor recurrence or death. These examinations included serological tests and at least one imaging modality: abdominal CE-US, primovist-enhanced MRI or contrast-enhanced CT. If tumor recurrence was found during follow-up, the most appropriate treatment, including RFA, surgical resection, liver transplantation, radiotherapy, intra-arterial chemoembolization, and systemic chemotherapy, was determined according to the tumor characteristics, liver function, and the general condition of the patient.

Short-Term Efficiency

Technical success was defined as follows: on the second day after RFA, CE-US showed that the range of the ablation zone completely covered the tumor lesion, and no abnormal enhancement was observed around the ablation zone. Technical efficacy was defined as follows: one month after ablation, contrast-enhanced CT or primovist-enhanced MRI showed no abnormal enhancement around the ablation zone. According to the Clavien-Dindo (CD) grading system, complications occurring within 3 months after RFA were divided into minor complications (CD grade ≤ II) and major complications (CD grade > II).¹¹

Long-Term Efficiency

RFS was defined as the interval between RFA treatment and initial tumor recurrence after RFA. OS was defined as the interval between RFA treatment and death from any cause. Patients who underwent liver transplantation were censored from the study on the transplantation date. The final follow-up date for this study was 31 December 2023. Patients who were alive at this time point were considered to be censored.

Statistical Analysis

Normally distributed continuous variables were presented as the mean ± the standard deviation (mean ± SD). Non-normally distributed continuous variables were presented as medians with interquartile ranges (IQR). Categorical variables were expressed as the number of cases and percentages. For each variable, data missing more than 20% were deleted, otherwise were interpolated using mean or median. Survival curves were constructed using the Kaplan-Meier method. The Log rank test was used to compare RFS and OS rates. Univariate and multivariate analyses of RFS and OS were performed using the Cox proportional hazard model. Two-tailed p < 0.05 was considered statistically significant. All data analyses were performed using R software (version 4.41).

Result

Patient Demographics

A total of 176 HCC patients (mean age 55.9±9.6 years, 149 males) with CPS ≥7 and liver dysfunction who underwent RFA at the First Affiliated Hospital of Army Medical University were included in this study. There were 142 (80.7%) patients with single tumor lesion and 34 (19.3%) patients with 2 or 3 lesions, a total of 214 tumor lesions (median diameter 22.0 cm, IQR:17.0–28.0) were ablated. Among them, 110 (62.5%) patients had primary HCC and 66 (37.5%) had recurrent HCC. The median AFP level was 23.1ng/L (IQR: 7.5–95.7). Regarding the etiology of cirrhosis, 154 (87.5%) were found to be infected with HBV, 9 (5.1%) with HCV, and 13 (7.4%) with other factors (alcohol and autoimmune hepatitis). There were 87 (49.4%) patients with CPS 7, 61 (34.7%) patients with CPS 8, 22 (12.5%) patients with CPS 9, 5 (2.8%) patients with CPS 10 and 1 (0.6%) patient with CPS 11. As for ALBI grade, 8 (4.5%) patients were classified as grade1, 130 (74.9%) as grade 2, and 38 (21.6%) as grade 3. The median score of MELD was 9.4 (IQR: 7.6–11.5). Detailed baseline characteristics were shown in Table 1.

Table 1 Baseline Patients Characteristics

Short-Term Efficacy

Twenty minutes after RFA, 6 lesions in 6 patients (3 with CPS 7 and 3 with CPS 8) were identified as residual by CE-US, resulting in a technical success rate of 96.6%. Three patients underwent re-ablation: two immediately after the first ablation and one 2 months after the first ablation. One month after RFA, 4 tumor lesions were found to be residual in 4 patients (CPS 7), giving a technical efficacy rate of 97.7%.

Within 3 months after RFA, one patient (CPS 9) died of esophagogastric variceal hemorrhage on the 55th day after RFA. Postoperative complications mainly included effusion, hemorrhage, and infection. Among these, effusion (84/176) was the most common complication, followed by hemorrhage (17/176), and infection-related complications (9/176) had the lowest incidence. The occurrence of these complications was shown in Figure 2 in different groups divided by CPS.

Figure 2 Complications in different groups divided by Child-Pugh score (CPS). (a). Ascites or hydrothorax in different group (0, none; 1, ascites; 2, hydrothorax; 3, ascites and hydrothorax). (b). Hemorrhage in different groups (0, none; 1, abdominal hemorrhage; 2, gastrointestinal bleeding; 3, hemothorax), the proportion of hemothorax in CPS 7 group was 1.1%, the proportion of gastrointestinal bleeding in CPS 8 group was 1.6%. (c). Infection in different groups (0, none; 1, abdominal infection; 2, pulmonary infection; 3, abscess or wound infection), the proportion of abscess in CPS 7 group was 1.1%, the proportion of wound infection in CPS 8 group was 1.6%.

Among all of the patients underwent RFA, 6 (3.4%) had major complications: one patient (CPS 7) was hospitalized for liver abscess 1 month after operation; one patient (CPS 8) had postoperative abdominal hemorrhage and pleural effusion; One patient (CPS 7) had a large amount of abdominal hemorrhage one day after operation; one patient (CPS 7) had postoperative thoracic bleeding; one patient (CPS 7) had postoperative pleural effusion and one patient (CPS 7) had postoperative pneumonia, type II respiratory failure, and pleural effusion. All the patients were discharged after appropriate treatment.

We compared short-term postoperative data among patients with different CPS. No significant difference was observed between the different CPS for either technical success (p=0.780) or technical efficacy (p=0.293). There were also no significant in 90 days morbidity (effusion, p=0.329; hemorrhage, p=0.115; infection, p=0.546) or mortality (p=0.159) between the different CPS groups. Patients with a higher CPS had a longer hospital stay (p=0.008), but the postoperative length of stay did not increase significantly (p=0.379). Detailed postoperative data were shown in Table 2.

Table 2 Postoperative Data

Long-Term Efficacy

The median follow-up time was 48.3 months (95% CI: 39.2, 53.8), 6 patients underwent liver transplantation after RFA.

A total of 126 (71.6%) patients had tumor recurrence (including 3 patients whose tumor was not completely ablated, their RFS was recorded as 0) after RFA, the median RFS was 1.2 years (IQR: 0.8–1.7). The 1-, 3-, 5-, and 7-year RFS were 54.1%, 28.8%, 16.8% and 10.8%, respectively (Figure 3a and Table 3). RFS was not significantly different between the different CPS, ALBI grade, or MELD scores (Table 3 and Figure 4). Univariate analysis showed 3 of the 24 variables may relevant to recurrence (p < 0.05), male patient (hazard ratio, 2.31; 95% CI: 1.196,4.464; p = 0.013) and recurrent HCC (hazard ratio, 2.347; 95% CI: 1.510,3.648; p < 0.001) were independent predictive factors for RFS in multivariate analysis (Table 4).

Table 3 Cumulative Recurrence-Free Survival in Different Groups Divided by Different Liver Function Models

Table 4 Uni- and Multivariable Cox Proportional Hazard Models of Predictors of Recurrence-Free Survival

Figure 3 Cumulative recurrence-free survival curve (a) and overall survival curve (b) of all hepatocellular carcinoma patients.

Figure 4 Cumulative recurrence-free survival (RFS) curves in different groups divided by different liver function models: (a) Child-Pugh score (CPS) 7/8/9/10&11; (b) albumin-bilirubin (ALBI) grade 1/2/3; (c) model for end-stage liver disease (MELD) grade 1/2/3 (<8/8-12/>12).

During the follow-up period, 63 (35.8%) patients died. The 1-, 3-, 5-, and 7-year OS were 91.3%, 67.1%, 58.5%, and 45.4% respectively (Figure 3b and Table 5). The OS rates differed significantly with ALBI grade (p = 0.015), but there was no significant difference among the different CPS or MELD scores (Table 5 and Figure 5). Univariate analysis showed 1 of the 24 variables may relevant to survival (p < 0.05), but hypertension (hazard ratio, 2.828; 95% CI: 1.037,7.714; p = 0.042) and tumor number (hazard ratio, 2.445; 95% CI: 1.225,4.882; p= 0.011) was independent predictive factor for OS with multivariable analysis (Table 6).

Table 5 Cumulative Overall Survival in Different Groups Divided by Different Liver Function Models

Table 6 Uni- and Multivariable Cox Proportional Hazard Models of Predictors of Overall Survival

Figure 5 Cumulative overall survival (OS) curves in different groups divided by different liver function models: (a) Child-Pugh score (CPS) 7/8/9/10&11; (b) albumin-bilirubin (ALBI) grade 1/2/3; (c) model for end-stage liver disease (MELD) grade 1/2/3 (<8/8-12/>12).

Discussion

In this study, we retrospectively analyzed the short- and long-term efficacy of RFA in 176 patients with early HCC and liver dysfunction (CPS ≥ 7), and demonstrated the safety profile and long-term outcomes of RFA. We found no significant increase in postoperative complication rates among patients with a higher CPS. Furthermore, long-term follow-up revealed no significant differences in RFS among different levels of liver dysfunction, and the ALBI grade was better in differentiating OS than the CPS or MELD scores. Compared to liver dysfunction, tumor-related characteristics have emerged as critical factors influencing long-term outcomes.

The development of HCC has been associated with pre-existing chronic liver disease, and most patients diagnosed with HCC present with advanced liver fibrosis or cirrhosis, limiting the choice of treatment. The safety and efficacy of RFA as a minimally invasive radical treatment for early HCC have been verified.^12,13 In recent years, the indications for RFA have expanded owing to advancements in RFA equipment and technology^14,15 as well as image-guidance modalities.¹³ A shift in CPS from ≤ 8 to ≤ 9 has been noted in several Japanese centers.¹⁶

It was worth noting that most patients with liver dysfunction were accompanied by decompensated liver function symptoms such as coagulation disorders and ascites. Even minimally invasive RFA still had potential risks. To ensure that patients could safely go through the perioperative period, diuretic, hepatoprotective, and other drug treatments were administered before and after RFA according to the specific conditions of the patients, resulting in a relatively long hospital stay (median,12 days; IQR: 8–17). However, there was no significant difference in the incidence of postoperative major complications compared with the previously reported incidence of complications in HCC patients after RFA of approximately 2.92–4.54%.^17–20 Combined with satisfactory long-term survival after ablation, we believe that RFA treatment for HCC patients with liver dysfunction (CP B) was relatively safe and meaningful.

Six HCC patients with CP C cirrhosis did not experience major complications, with a median OS of 3.9 years. Although the finding differ from previous reports indicating major complication rates of 0–12.6% and median OS of 12–26 months.^21–24 It is noteworthy that all of these reports concurred that ablation therapy provides a survival benefit. However, constrained by the level of evidence, we contend that no universally applicable treatment exists for HCC patients with CP C. A more meticulous, individualized assessment of cirrhosis severity is warranted, alongside leveraging each center’s ablation expertise to formulate the most balanced and comprehensive treatment strategy. This approach should consider the potential of multiple therapeutic options, not merely the most readily available treatment.

In this study, 126 (69.9%) patients experienced tumor recurrence after RFA. These findings align with the previously reported high range of postoperative recurrence rates (50–80%). Interestingly, RFS did not decrease as the level of liver dysfunction increased in our study, male and recurrent HCC remained the independent predictive factors for RFS. Considering that the disease naturally progressing to a more aggressive state over time. We believe that, for patients with liver dysfunction, tumor-related factors are more likely to lead to postoperative recurrence than liver dysfunction.

Although the postoperative recurrence rate was relatively high and RFS was short, the OS after RFA was encouraging. The 5-year OS reacheed 58.5%, surpassing the previously reported 5-year OS for CP B early HCC patients, which ranged between 25.9% and 46.5%.^25–27, This outcome stems from two factors: First, our center has performed over 10,000 RFA procedures for HCC patients over the past two decades, resulting in highly refined ablation techniques. Second, we analyzed data from patients enrolled within the last decade, with the final follow-up date set to December 31, 2023. The shortest follow-up period was only 5 months, indicating that some patients’ mortality outcomes could not be observed, potentially leading to an overestimation of the 5-year overall survival rate.

Previous studies have confirmed that a higher CPS correlates with lower OS in patients with HCC after RFA.²⁸ To further clarify whether therapeutic efficacy differs among patients with CPS ≥ 7, based on CPS level, we subdivided all patients into CPS 7, 8, 9, and 10&11 groups. We found no significant differences in OS among the four groups (p=0.690). Using the more objective ALBI classification, we grouped patients into ALBI grade 1, 2, and 3.²⁹ Results revealed a statistically significant difference in OS among these three groups (p=0.015). Although the CP classification was widely accepted for liver function grading, it had limitations: the CTP score was constrained by the equal weighting of five parameters with arbitrary cutoff values and it relied on relatively subjective assessments of ascites and hepatic encephalopathy, which may compromise interobserver reproducibility and not accurately suited for retrospective calculations for research. In contrast, ALBI score can be easily calculated retrospectively, as the included variables were available in almost all patients. Related studies indicate that even within the CP A,³⁰ the ALBI score may identify different survival probabilities after surgical and non-surgical therapies. This aligns with our findings within CP B, confirming its ability to provide a more granular assessment of hepatic reserve capacity than the Child-Pugh score.

This study had the following limitations: First, as a retrospective, single-center study, it suffered from missing clinical data and recall bias. Second, this study focused on the short- and long-term efficacy of RFA in patients with liver dysfunction who were not candidates for surgical resection. Liver function was quantified using the Child-Pugh score, enrolling patients with a CPS ≥ 7. However, the CPS system lacks standardized quantification of ascites and hepatic encephalopathy, introduces subjective elements. Whether this system can more accurately identify patients with poor liver function compared with other objective non-invasive liver function assessment systems warrants further investigation. Third, this study did not examine post-ablation changes in the liver function. Given the high recurrence rate, whether liver function after recurrence supports repeat RFA requires further investigation. Fourth, because only six patients with CP C liver function were included in this study, it was not possible to accurately assess the safety and long-term postoperative efficacy of RFA in treating such patients.

Conclusion

Radiofrequency ablation is safe and effective for early hepatocellular carcinoma patients with liver dysfunction who cannot tolerate surgical resection. The impact of tumor characteristics on long-term efficacy may be more significant than that of liver dysfunction.