Journal of Gastrointestinal Surgery

, Volume 15, Issue 2, pp 311–320

Radiofrequency Ablation Versus Surgical Resection for Hepatocellular Carcinoma in Childs A Cirrhotics—a Retrospective Study of 1,061 Cases

Authors

  • Jiwei Huang
    • Department of Hepato-Biliary-Pancreatic Surgery, Treatment Centre of Liver Cancer, West China HospitalSichuan University
  • Roberto Hernandez-Alejandro
    • From Department of SurgeryUniversity of Western Ontario, Hepatobiliary and Liver Transplant Surgery, London Health Sciences Centre
  • Kristopher P. Croome
    • From Department of SurgeryUniversity of Western Ontario, Hepatobiliary and Liver Transplant Surgery, London Health Sciences Centre
  • Lvnan Yan
    • Department of Hepato-Biliary-Pancreatic Surgery, Treatment Centre of Liver Cancer, West China HospitalSichuan University
  • Hong Wu
    • Department of Hepato-Biliary-Pancreatic Surgery, Treatment Centre of Liver Cancer, West China HospitalSichuan University
  • Zheyu Chen
    • Department of Hepato-Biliary-Pancreatic Surgery, Treatment Centre of Liver Cancer, West China HospitalSichuan University
  • Pankaj Prasoon
    • Department of Hepato-Biliary-Pancreatic Surgery, Treatment Centre of Liver Cancer, West China HospitalSichuan University
    • Department of Hepato-Biliary-Pancreatic Surgery, Treatment Centre of Liver Cancer, West China HospitalSichuan University
Original Article

DOI: 10.1007/s11605-010-1372-y

Cite this article as:
Huang, J., Hernandez-Alejandro, R., Croome, K.P. et al. J Gastrointest Surg (2011) 15: 311. doi:10.1007/s11605-010-1372-y

Abstract

Introduction

The long-term outcomes of radiofrequency ablation (RFA) vs. surgical resection in cirrhotic patients with hepatocellular carcinoma (HCC) remain controversial. One thousand sixty-one cirrhotic HCC patients were included into a retrospective study. Four hundred thirteen received RFA and 648 received surgical resection.

Results

Overall (OS), recurrence-free (RFS), and tumor-free survival (TFS) were compared between the two groups and in subgroup analyses. The 5-year OS and corresponding RFS as well as DFS were significantly higher in the surgical resection group compared with the RFA group (p < 0.001, p < 0.001, p < 0.001). In subgroup analyses of solitary HCC ≤3 cm, there was no significant difference in RFS between the two groups (p = 0.719). Nonetheless, surgical resection was superior to RFA for OS and TFS in this subgroup as well as for OS, RFS, and TFS in subgroup analyses for solitary lesions 3 cm < HCC < 5 cm and multifocal HCC. Serum AFP was the only significant predicting factor for all survival analyses.

Conclusions

When treating Childs A cirrhotic patients with solitary HCC larger than 3 cm but less than 5 cm, or with two or three lesions each less than 5 cm, surgical resection provides a better survival than RFA. When treating Childs A cirrhotics with solitary HCC ≤ 3 cm, RFA has a comparable RFS to surgical resection, but RFA is less invasive.

Keywords

Radiofrequency ablationHepatectomyHepatocellular carcinoma

Introduction

Hepatocellular carcinoma (HCC) is one of the most common liver neoplasms in the world, with an estimated global incidence of over 500,000 new cases per annum.1 It is prevalent in Asia and Africa, and its prevalence is increasing in the USA and Europe.2,3 Clear guidelines for the treatment of HCC have yet to be established. Surgical resection is the most widely used treatment worldwide.4 Surgical resection can only be used in around 5% of patients in the western world and around 40% of Asian patients due to poor liver functional reserve caused by liver cirrhosis and intrahepatic dissemination.5,6 Radiofrequency ablation (RFA) has emerged as a new effective and reliable therapy modality for small HCC with encouraging outcomes.713 This study was designed to compare long-term outcomes of surgical resection versus RFA for the treatment of HCC patients with Childs A cirrhosis.

Material and Methods

Diagnostic and Selection Criteria

The study was performed according to the guidelines of the Helsinki Declaration. A written informed consent was obtained from each every patient before intervention. A retrospective review was performed at our center on all patients who presented between July 2000 and October 2005 with liver cirrhosis and who were diagnosed as having HCC. The diagnosis of liver cirrhosis was made by either biopsy or by clear clinical signs of cirrhosis such as: ascites, coagulopathy, or radiological features. The diagnosis of HCC was made according to the diagnostic criteria used by the European Association for the Study of the Liver,14 which based on histopathological confirmation by ultrasound-guided fine-needle aspiration biopsy or noninvasive methods: (1) the concordant classical dynamic radiological features of HCC were represented in two radiologic techniques; (2) one radiologic technique showed typical features of HCC together with an elevated alpha fetoprotein (AFP) level over 400 ng/ml; (3) when a tumor ≤2 cm was found in the cirrhotic liver, magnetic resonance imaging (MRI), and hepatic digital subtraction angiography (DSA) confirmation were both needed before making a diagnosis. Radiologic imaging techniques included ultrasonography, spiral computer tomography (CT), MRI, and hepatic DSA.

Inclusion criteria were as follows: up to three nodules, each ≤5 cm, no extrahepatic metastasis or obvious vascular invasion, well-compensated liver function of Pugh-Child Class A, a platelet count >50 × 109/L and a prothrombin time prolongation ≤5 s, no previous or simultaneous malignancies, HBV-infected surgery-treated patient should have a HBV-DNA-PCR quantitation of less than 105 copies/ml. Patients were excluded if they had undergone other treatments prior to radical resection or RFA.

Follow-up

Patients were followed-up at 3-month intervals after treatment. Abdominal ultrasonography and helical CT, measurement of serum AFP and liver function tests were performed during each visit. When intrahepatic recurrence was suspected, spiral CT, MRI, or contrast-enhanced ultrasonography was performed. When extrahepatic metastases were suspected, thoracic CT, and bone scintigraphy were performed. Once the recurrence was confirmed, the second-time treatment was proposed by a multidisciplinary team of specialists including surgeons, pathologists, and radiologists; however, the patient's opinion was conclusive. The therapies included liver transplantation, RFA, percutaneous ethanol injection, surgical resection, transcatheter hepatic arterial chemoembolization (TACE), and systematic chemotherapy.

Statistical Analysis

Differences between the RFA and surgical resection groups were analyzed using the unpaired t test for continuous variables and by the χ2 test or continuity correction method for categorical variables. Survival curves, recurrence-free survival curves, and tumor-free survival curves were generated using the Kaplan–Meier method and compared by the log-rank test. The relative prognostic significance of the variables in predicting overall survival and overall recurrence were assessed by univariate and multivariate Cox proportional hazards regression models. All variables with a p value <0.05 by univariate comparison were subjected to multivariate analysis. Results of multivariate analysis were presented as relative risk with corresponding 95% confidence intervals (CI). All statistical tests were two-sided, and differences were considered when p < 0.05. The statistical analyses were performed using SPSS 13.0 statistical software (SPSS Company, Chicago, IL, USA).

Radiofrequency Ablation

Procedure

The RFA procedures were performed with a commercially available system (Radionics, Cool-Tip System, Burlington, MA, USA), single/clustered needle electrode(s) with a 2-cm or 3-cm exposed tip and ultrasound guidance (Vivid4, GE, USA; iU22, Philips, USA). The clustered electrodes were systematically applied in lesions larger than 3 cm. The percutaneous RFA procedure was employed when appropriate (n = 361). Open approach was selected when the tumors were located near the subhepatic inferior vena cava or gastrointestinal tract (n = 52). The assessment of response was made according to the modified European Association for the Study of the Liver criteria.14 A spiral triphasic enhanced CT or MRI was performed 1 month after treatment. A complete ablation response was indicated by the absence of enhancing tissue at lesion site. Residual viable tumor was diagnosed if an enhanced area was noted within the treatment zone (n = 42). In order to achieve complete ablation, the treatment course could be repeated with another CT/MRI evaluation 1 month later. If residual viable tissue of the tumor still existed, RFA was considered a failure and the patient was treated with TACE (n = 12).

Surgical Resection

The hepatectomy procedure was defined according to the Brisbane terminology proposed by Strasberg et al.15 Anatomic resection was defined as resection of the lesion together with the portal vein related to the lesion and the corresponding hepatic territory. Nonanatomic resection was defined as resection of a lesion without regard to segmental, sectional, or lobar anatomy. We performed anatomic partial hepatectomy when appropriate (n = 279). In other cases, we performed nonanatomic resection with a resection margin of 1 cm over the tumor by visual estimation intraoperatively (n = 369).

Results

From July 2000 to October 2005, 2,637 consecutive patients with liver cirrhosis who were diagnosed as having HCC were treated in our center. Of these, 1,576 patients did not meet the aforementioned inclusion criteria. The remaining 1,061 patients consisted of 413 patients treated by RFA and 648 patients received surgical resection, respectively. The diagnosis of HCC and liver cirrhosis of the 648 patients in the surgical resection group were all confirmed by the excised specimens histopathologically. Of the 413 patients treated by RFA, the diagnosis of HCC and liver cirrhosis established noninvasively in 256 patients and was proven by biopsy in 157 patients.

Demographic characteristics as well as the number of patients in each subgroup for both patients in the RFA group and in the surgical resection group can be seen in Table 1. Significant differences between the two groups were seen in three parameters: age, gender, and AFP. The surgical resection group had older, more-female population, and with higher AFP level. In subgroup of solitary HCC ≤3 cm, there were 311 patients in the surgical resection group and 212 patients in the RFA group. Demographic characteristics were comparable except for the surgical resection group had older population. The tumor locations by intervention group can be seen in Table 2. The mean follow-up time was 33.7 ± 17.4 months for surgical resection group and 36.1 ± 12.4 months for RFA group (p = 0.178), respectively. The censored patients were more in the surgical resection group (87/648) than in the RFA group (33/413, p = 0.006).
Table 1

Demographic characteristics of surgical resection group and RFA group

 

Surgical resection group n = 648

RFA group n = 413

p value

Age (years)

46.13 ±16.89

54.67 ±12.18

0.015

Gender:

  

0.000

 Male

489 (75.5%)

361 (87.4%)

 

 Female

159 (24.5%)

52 (12.6%)

 

Viral hepatitis status:

  

0.210

 Hepatitis B viral infected (HBV)

598 (92.3%)

391 (94.7%)

 

 Hepatitis C viral infected (HCV)

28 (4.3%)

15 (3.6%)

 

 Non-HCV/HBV

22 (3.4%)

7 (1.7%)

 

Cause of liver cirrhosis

  

0.189

 Hepatitis-B

598 (92.3%)

391 (94.7%)

 

 Hepatitis-C

28 (4.3%)

15 (3.6%)

 

 Alcoholic

7 (1.1%)

4 (1.0%)

 

 Drugs

13 (2.0%)

2 (0.5%)

 

 Autoimmune hepatitis

0 (0%)

1 (0.2%)

 

 Others

2 (0.3%)

0 (0%)

 

 Tumor category:

  

0.076

 Solitary tumor ≤3 cm

311 (48.0%)

212 (51.3%)

 

 Solitary tumor >3 cm

196 (30.2%)

101 (24.5%)

 

 Multifocal

141 (21.8%)

100 (24.2%)

 

 Mean tumor size

3.56 ±1.47

4.01 ±1.21

0.132

AFP

  

0.000

 ≤ 400

290 (44.8%)

256 (62.0%)

 

 400 < AFP ≤ 1,200

327 (50.5%)

149 (36.1%)

 

 > 1,200

31 (4.8%)

8 (1.9%)

 

Non-HBV/HCV patients negative for both HBV and HCV antibody except HBV surface antibody

Table 2

Tumor locations of surgical resection group and RFA group

Group

Location (segment)

I

II

III

IV

V

VI

VII

VIII

Surgical resection (lesions n = 808)

4

86

62

125

231

168

107

25

RFA (lesions n = 518)

0

41

45

87

174

107

54

10

P = 0.099 by Pearson χ2 test

Register per lesion

Lesion between segments, registered as major location

The RFA Group

Two hundred twelve of 413 RFA-treated patients had single lesion ≤3 cm; the mean treatment session was 1.68 ± 0.81/lesion. One hundred and one of 413 patients had solitary HCC >3 cm, but <5 cm, the mean treatment session was 3.07 ± 0.23/lesion. 93/413 patients had two lesions, the treatment session were 2.13 ± 0.34/lesion. Seven patients had three lesions, and the mean treatment session was 1.94 ± 0.57/lesion. The complete ablation response rate after first treatment was 89.83% (371/413) and 97.09% (401/413) after the second-time treatment. Twelve patients resulted in treatment failure by RFA, and they all treated by TACE and other palliative therapies.

The Surgical Resection Group

Two hundred seventy-six of 648 operated patients received anatomic partial hepatectomy, the remaining 372 received nonanatomic resection. Ninety-one of 276 patients were treated with monosegmentectomy; 152/276 patients received bisegmentectomy; 23/122 underwent hemihepatectomy; 9/276 were treated with tri-sectorectomy, and 1/122 received right lateral sectorectomy plus bisegmentectomy of segment II+III. Pringle maneuver was performed in 217/648 patients. Mean blood-loss during operation was 501.2 ± 214.3 ml (range, 200–2,750 ml), 111/648 patients required intraoperative blood transfusion.

Fifty-nine new lesions were found in 57 patients intraoperatively, 58 by IOUS during the operations, and one by specimen examination. The smallest “safe margin” from the lesion measured in nonviable specimens were from 0.5 to 3.5 cm, and insufficient resection were revealed in 23 patients. There were 126 specimens poorly differentiated, 295 moderately differentiated, and 227 well differentiated. Twenty-four specimens were found microsatellites while 12 found microvascular invasion. Fifty-nine patients (23 with insufficient resection, 24 with microsatellites, and 12 with microvascular invasion) each underwent one session of TACE.

Hospitalization Length, Mortality, Complications, and Adverse Events

The hospitalization length was significantly longer in the surgical resection group (17.83 ± 3.25 day) than in RFA group (6.12 ± 2.98 day; p < 0.001).

There was one death related to acute pulmonary embolism on postoperative day 7 in the surgical resection group. No patient died within 30 days after treatment in the RFA group.

Frequency of complications was significantly higher in the surgical resection group than in the RFA group (71/648 vs. 19/413, p < 0.001). Complications in the surgical resection group were as follows: acute pulmonary embolism (one case), hepatic failure (four), refractory ascites (24), encapsulated effusion needing percutaneous drainage (21), bile leakage (12), postoperative bleeding (six), and gastrointestinal bleeding (three). Complications in the RFA group were: gastric perforation (one case), procedure-related hemorrhage (11), malignant seeding (three) hepatic infarction (one), skin burn (three). A significantly higher number of patients required analgesics after treatment in the surgical resection group than in the RFA group (339/648 vs. 131/413, p < 0.001).

Survival

One hundred seventy-seven patients in the RFA group died during the follow-up. Causes of death were cancer recurrence (148 cases), liver failure (12), upper gastrointestinal hemorrhage (11), and miscellaneous (six). One hundred forty-one patients in the surgical resection group died during the follow-up, causes of death were cancer recurrence (114cases), liver failure (23), and miscellaneous (four).

The 1-, 2-, 3-, 4-, and 5-year overall survival rates for the RFA group, and the surgical resection group were 86.19%, 75.06%, 63.20%, 56.54%, 53.34%, and 94.14%, 87.89%, 83.26%, 79.48%, 76.47%, respectively. The surgical resection group had significantly better overall survival than the RFA group (p < 0.001 by log-rank test, Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs11605-010-1372-y/MediaObjects/11605_2010_1372_Fig1_HTML.gif
Fig. 1

Overall survival of RFA group and resection group

The corresponding recurrence-free survival rates were 73.25%, 54.02%, 44.89%, 31.78%, and 26.51% for the RFA group and 80.19%, 67.01%, 57.13%, 49.09%, and 42.97% for the surgical resection group, respectively. The recurrence-free survival of the surgical resection group was significantly better than the RFA group (p < 0.001by log-rank test, Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs11605-010-1372-y/MediaObjects/11605_2010_1372_Fig2_HTML.gif
Fig. 2

Recurrence-free survival of RFA group and resection group

The corresponding tumor-free survival (defined by the absence of a detectable tumor at the endpoint) rates were 78.69%, 60.77%, 46.97%, 36.83%, and 31.71% for the RFA group and 87.80%, 75.31%, 63.88%, 53.09%, and 43.67% for the surgical resection group, respectively. The tumor-free survival of the surgical resection group was significantly better than the RFA group (p < 0.001by log-rank test, Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs11605-010-1372-y/MediaObjects/11605_2010_1372_Fig3_HTML.gif
Fig. 3

Tumor-free survival of RFA group and resection group

Subgroup Analyses

In subgroup analyses of overall survival performed in solitary HCC ≤3 cm, surgical resection was superior to RFA in overall survival (p < 0.001by log-rank test, Fig. 4); however, there was no statistical difference in recurrence-free survival (p = 0.719 by log-rank test, Fig. 5). The tumor-free survival result was concordant with overall survival between the two groups (p < 0.001 by log-rank test, Fig. 6). In subgroup analyses for solitary lesions, 3 cm < HCC <5 cm and multifocal HCC surgical resection was superior to RFA for overall survival, recurrence-free survival, and tumor-free survival.
https://static-content.springer.com/image/art%3A10.1007%2Fs11605-010-1372-y/MediaObjects/11605_2010_1372_Fig4_HTML.gif
Fig. 4

Overall survival of RFA group and resection group with solitary HCC ≤3 cm

https://static-content.springer.com/image/art%3A10.1007%2Fs11605-010-1372-y/MediaObjects/11605_2010_1372_Fig5_HTML.gif
Fig. 5

Recurrence-free survival of RFA group and RES group with solitary HCC ≤3 cm

https://static-content.springer.com/image/art%3A10.1007%2Fs11605-010-1372-y/MediaObjects/11605_2010_1372_Fig6_HTML.gif
Fig. 6

Tumor-free survival of RFA group and RES group with solitary HCC ≤3 cm

The main causes of liver cirrhosis in this study were HBV and HCV infection. There were 989/1061 patients infected by HBV and 43 infected by HCV (Table 1). The 1-, 3-, and 5-year overall survival rates for the HBV-infected patients were 91.27%, 75.45%, and 66.67% and 83.33%, 54.72%, and 26.20% for the HCV-infected patients, respectively. The corresponding recurrence-free survival rates were 75.61%, 62.47%, and 35.17% for the HBV-infected patients and 76.19%, 47.61%, and 14.29% for the HCV-infected patients, respectively. The corresponding tumor-free survival rates were 80.14%, 71.25.47%, and 40.01% for the HBV-infected patients and 78.57%, 52.38%, and 19.05% for the HCV-infected patients, respectively. The HCV-infected patients had significantly poorer outcomes in overall, recurrence-free, and tumor-free survival analyses (p < 0.001, p < 0.001, p < 0.001 by log-rank test).

The univariate and multivariate analyses of predictors of overall survival, recurrence-free survival, and tumor-free survival for all 1,061 patients were shown in Table 3. Significant predicting parameters for the overall survival were: intervention (surgical resection vs. RFA), recurrence intervention (non-radical vs. radical), serum AFP, tumor size, and tumor number. The corresponding relative risks were 3.471 (95% CI 1.217–5.574), 2.012 (95% CI 0.098–4.221), 9.041(95% CI 3.764–25.133), 1.893 (95% CI 1.231–5.865), and 2.981(95% CI 2.010–4.351), respectively. For recurrence-free survival, significant predicting parameters were the same as overall survival except recurrence intervention. The corresponding relative risks were 2.611 (95% CI 1.521–3.894), 9.033 (95% CI 4.510–24.315), 2.031 (95% CI 1.049–4.508), and 4.151(95% CI 2.186–10.481), respectively. However, for tumor-free survival, intervention (surgical resection vs. RFA), recurrence intervention (non-radical vs. radical), and serum AFP were the significant predicting parameters, and the relative risks were 2.079 (95% CI 1.144–3.586), 1.559 (95%CI 0.751–2.874), and 4.656 (95% CI 2.688–9.147), respectively.
Table 3

Univariate and multivariate analysis of the predictors for survival of all patients

Variable

Univariate analysis (p value)

Multivariate analysis

Relative risk (95%CI)

p value

Survival

 Intervention (RFA vs RES)

0.000

3.471 (1.217–5.574)

0.000

 Recurrence intervention (RFA vs RES)

0.038

  

 Recurrence intervention (non- vs radical)

0.001

2.012 (0.098–4.221)

0.001

 Age (years) (>65 vs ≤65)

0.032

  

 Underlying liver disease

   

 HBV vs non-HBV

0.518

  

 HCV vs non-HCV

0.945

  

 Albumin (g/L) (≤35 vs >35)

0.521

  

 Total bilirubin (mmol/L) (>10 vs ≤)

0.493

  

 Serum AFP (ng/ml) (≥400 vs <400)

0.015

9.041 (3.764–25.133)

0.000

 Prothrombin time (<15′ vs >15′)

0.917

  

 Tumor size (cm) (>3 vs ≤3)

0.001

1.893 (1.231–5.865)

0.001

 Tumor number (multifocal vs single)

0.000

2.981 (2.010–4.351)

0.000

Recurrence-free survival

 Intervention (RFA vs RES)

0.000

2.611 (1.521–3.894)

0.000

 Age (years) (>65 vs ≤65)

0.531

  

 Underlying liver disease

   

 HBV vs non-HBV

0.032

  

 HCV vs non-HCV

0.067

  

 Albumin (g/L) (≤35 vs >35)

0.241

  

 Total bilirubin (mmol/L) (>10 vs ≤10)

0.971

  

 Serum AFP (ng/ml) (≥400 vs <400)

0.012

9.033 (4.510–24.315)

0.000

 Prothrombin time (<15′ vs >15′)

0.651

  

 Tumor size (cm) (>3 vs ≤3)

0.001

2.031 (1.049–4.508)

0.023

 Tumor number (multifocal vs single)

0.001

4.151 (2.186–10.481)

0.000

Tumor-free survival

 Intervention (RFA vs RES)

0.017

2.079 (1.144–3.586)

0.001

 Recurrence intervention (RFA vs RES)

0.041

  

 Recurrence intervention (non- vs radical)

0.031

1.599 (0.751–2.874)

0.012

 Age (years) (>65 vs ≤65)

0.211

  

 Underlying liver disease

   

 HBV vs non-HBV

0.081

  

 HCV vs non-HCV

0.121

  

 Albumin (g/L) (≤35 vs >35)

0.322

  

 Total bilirubin (mmol/L) (>10 vs ≤10)

0.856

  

 Serum AFP (ng/ml) (≥400 vs <400)

0.018

4.656 (2.688–9.147)

0.010

 Prothrombin time (<15′ vs >15′)

0.557

  

 Tumor size (cm) (>3 vs ≤3)

0.167

  

 Tumor number (multifocal vs single)

0.133

  

Non-HBV/HCV patients negative for both HBV and HCV antibody except HBV surface antibody

In the subgroup of solitary HCC ≤3 cm, univariate and multivariate analyses of predictors of overall survival, recurrence-free survival, and tumor-free survival were shown in Table 4. Intervention (surgical resection vs. RFA) was not related to overall survival, recurrence-free survival, and tumor-free survival. Nonetheless, recurrence intervention (non-radical vs. radical) was related to overall survival and tumor-free survival. Serum AFP was a significant predictor for all survival analyses.
Table 4

Univariate and multivariate analysis of the predictors for survival of patients with solitary HCC ≤3 cm

Variable

Univariate analysis (p value)

Multivariate analysis

Relative risk (95%cl)

p value

Survival

 Intervention (RFA vs RES)

0.041

  

 Recurrence intervention (RFA vs RES)

0.047

  

 Recurrence intervention (non- vs radical)

0.004

1.510 (0.311–3.023)

0.001

 Age (years) (>65 vs ≤65)

0.341

  

 Underlying liver disease

   

 HBV vs non-HBV

0.087

  

 HCV vs non-HCV

0.189

  

 Albumin (g/L) (≤35 vs >35)

0.351

  

 Total bilirubin (mmol/L) (>10 vs ≤10)

0.344

  

 Serum AFP (ng/ml) (≥400 vs <400)

0.000

11.552 (5.853–31.568)

0.000

 Prothrombin time (<15′ vs >15′)

0.035

  

Recurrence-free survival

 Intervention (RFA vs RES)

0.013

  

 Age (years) (>65 vs ≤65)

0.487

  

 Underlying liver disease

   

 HBV vs non-HBV

0.158

  

 HCV vs non-HCV

0.264

  

 Albumin (g/L) (≤35 vs >35)

0.544

  

 Total bilirubin (mmol/L) (>10 vs ≤10)

0.762

  

 Serum AFP (ng/ml) (≥400 vs <400)

0.031

5.122 (1.587–8.994)

0.001

 Prothrombin time (<15′ vs >15′)

0.632

  

Tumor-free survival

 Intervention (RFA vs RES)

0.035

  

 Recurrence intervention (RFA vs RES)

0.056

  

 Recurrence intervention (non- vs radical)

0.021

2.077 (0.988–4.556)

0.006

 Age (years) (>65 vs ≤65)

0.117

  

 Underlying liver disease

   

 HBV vs non-HBV

0.122

  

 HCV vs non-HCV

0.207

  

 Albumin (g/L) (≤35 vs >35)

0.410

  

 Total bilirubin (mmol/L) (>10 vs ≤10)

0.799

  

 Serum AFP (ng/ml) (≥400 vs <400)

0.045

3.114 (0.877–8.7557)

0.003

 Prothrombin time (<15′ vs >15′)

0.845

  

Non-HBV/HCV patients negative for both HBV and HCV antibody except HBV surface antibody

Recurrence

Till the endpoint date of this study, recurrence was observed in 282 patients in the RFA group and 346 in the surgical resection group. Two hundred nine recurrent patients in the RFA group and 275 in the surgical resection group were amenable for radical treatment (i.e., excision, ablation, or transplantation). The other patients were either with extensive intrahepatic tumor dissemination or extrahepatic metastasis. They were treated by TACE or other palliative therapies. In the RFA group, 161/209 patients were treated by iterative RFA, 46 were given hepatic resection, and two received transplantation. In the surgical resection group, RFA was performed in 217/275 patients, and a repeat resection was applied in 58 patients. There was no significant difference in proportion of recurrence amenable to radical treatment between the two groups (p = 0.112). The frequency of applying RFA was significantly higher than excision for the recurrent HCC patients amenable to radical treatment (p < 0.05).

However, in the subgroup of solitary HCC ≤3 cm, there were 112 patients in the RFA group, and 194 in the surgical resection group were amenable to radical treatment. There were significantly more recurrent HCC patients amenable to radical treatment in the surgical resection group (p = 0.030).

Discussion

Several different treatment modalities exist for patients with HCC and cirrhosis. Liver transplantation has been shown to have superior results with regards to overall and recurrence-free survival; however, with organs in limited supply other modalities must be fully examined.16 Radiofrequency ablation has gained support in recent years due to its postulated decrease in complication rates when compared with surgical resection. Previous studies have compared RFA with surgical resection with mixed results. Several observational studies have shown that surgical resection is superior to RFA in terms of overall and recurrence-free survival.17 Other observational studies and one randomized prospective trial have suggested that with smaller HCC lesions (<4 cm) there are no significant differences in overall and recurrence-free survival between RFA and surgical resction.12,18,19 To our knowledge the current study is the largest published to date comparing RFA and surgical resection. Our results had demonstrated superior survival benefit for the Childs A cirrhotic patients undergoing surgical resection as compared with radiofrequency ablation. However, in subgroup analysis of lesions ≤ 3 cm, we found no significant difference in recurrence-free survival between RFA and surgical resection. This corresponds with the findings of the aforementioned studies.12,18,19

According to the results of this study, recurrence was the main reason of death which directly affected the survival analyses (148/177 in the RFA group and 114/141 in the surgical resection group). The difference of local tumor clearance between the two modalities might be the essential factor that affected recurrence. HCC mainly disseminates through portal veins and hepatic veins. The tumor embolus could shed in the neighboring branches of vessels and form the microsatellite.2023 Partial hepatectomy especially anatomic resection removed at least 1 cm rim of normal liver parenchyma together with the original lesion macroscopically, and thus theoretically eliminated both the primary tumor and possible venous tumor thrombi.24,25 This was impossible to be achieved by any local ablation modalities. Furthermore, in the RFA procedure, repeated insertion and overlapping the ablation areas were necessary when encountering tumors larger than one single session ablative area. Via the guidance of 2D ultrasonography, a viable seam could be possibly left undetected in the actual lesion area which existed in a three-dimensional formation during the process of overlaying the ablation sessions. This hypothesis had actually been proved by colleagues from Japan.24

In cases of solitary HCC ≤3 cm, overlaying ablation was usually not necessary because the necrosis area produced by one session of a single-needle electrode was closed to a sphere with a diameter of 3 cm.26 The viable tumor nest was consequently hard to survival due to homogeneously heat effect. This might at least in part explain why no significant difference in recurrence-free survival between RFA and surgical resection for HCC less than 3 cm was found. However, in the subgroup of solitary HCC ≤3 cm, the overall survival and tumor-free survival of surgical resection group were significantly better than the RFA group. This might be due to the recurrent patients in the surgical resection group had a larger proportion amenable to radical treatment, which might confound the overall and tumor-free survival results. Furthermore, we found that, in multivariate analyses (radical or non-radical), recurrence intervention was a significant predicting factor for overall and tumor-free survival.

The patients' liver cirrhosis in our study were mainly caused by hepatitis virus. In subgroup survival analyses of HBV and HCV infection, HCV-infected patients had significantly inferior survival in overall, recurrence-free, and tumor-free survival when compared with HBV-infected patients. This might reveal a trend that HCC patients with HCV-related cirrhosis had poorer prognosis and higher incidence of tumor relapse than HBV-infected patients.

The current study showed a lower incidence of complications and adverse events in the RFA group. In addition the length of hospital stay was significantly shorter in the RFA group. These results were likely explained by the less invasive nature of RFA compared with surgical resection.

When facing recurrent HCC amenable for radical treatment, the frequency of applying RFA was significantly higher than excision. The contraindications for surgical resection might increase in patients with recurrent HCC with histories of surgery because of poor liver functional reserve or inadequate liver remnant, especially in distant intrahepatic recurrence cases. In contrast, RFA could be applied in most of these patients. As concluded by other groups,27,28 RFA might even be superior to surgery when facing iterative recurrences. These results were likely due to the less invasive nature of RFA compared with surgical resection as well.

In multivariate analyses, intervention was an independent predictor for overall, recurrence-free, and tumor-free survival for all the 1,061 patients. Interestingly, in the subgroup of solitary HCC less than 3 cm, it was completely not related to any. This might reveal surgical resection was superior to RFA when treating HCC larger than 3 cm or multiple lesions, but when facing solitary small HCC, the effectiveness of the two modalities were comparable.

Serum AFP was the only significant predicting factor for all survival analyses in this study. A high AFP level was usually observed in a high-grade malignancy HCC, which was featured by high incidence of recurrence and poor prognosis.29 Moreover, high AFP level was proved to be a risk factor for HCC occurrence in cirrhotic patients.30,31

This study clearly has limitation as a result of its retrospective cohort design. Baseline characteristics of the two groups showed slight differences in age of patients, gender distribution, and baseline AFP level. Furthermore, the rates of censorship were high in 87/648 in the surgical resection group and 33/413 in RFA group.

Conclusion

The current study demonstrated that when treating Childs A cirrhotic patients with solitary HCC larger than 3 cm but less than 5 cm, or with two or three lesions each less than 5 cm, surgical resection provides better overall survival, recurrence-free survival, and tumor-free survival. When treating Childs A cirrhotics with solitary HCC ≤ 3 cm, RFA has a comparable recurrence-free survival benefit to surgical resection, but RFA is less invasive.

Copyright information

© The Society for Surgery of the Alimentary Tract 2010