Langenbeck's Archives of Surgery

, Volume 398, Issue 1, pp 55–62

Could radiofrequency ablation replace liver resection for small hepatocellular carcinoma in patients with compensated cirrhosis? A 5-year follow-up

Authors

    • Department of Digestive Surgery and Liver UnitUniversity of Perugia
  • Stefano Trastulli
    • Department of Digestive Surgery and Liver UnitUniversity of Perugia
  • Rosario Pasquale
    • Department of Digestive Surgery and Liver UnitSt. Maria Hospital
  • Davide Cavaliere
    • Unit of Surgery and Advanced Oncologic TherapiesForlì Hospital
  • Roberto Cirocchi
    • Department of General and Oncologic SurgeryUniversity of Perugia
  • Carlo Boselli
    • Department of General and Oncologic SurgeryUniversity of Perugia
  • Giuseppe Noya
    • Department of General and Oncologic SurgeryUniversity of Perugia
  • Amilcare Parisi
    • Department of Digestive Surgery and Liver UnitSt. Maria Hospital
Original Article

DOI: 10.1007/s00423-012-1029-2

Cite this article as:
Desiderio, J., Trastulli, S., Pasquale, R. et al. Langenbecks Arch Surg (2013) 398: 55. doi:10.1007/s00423-012-1029-2
  • 368 Views

Abstract

Purpose

Treating hepatocellular carcinoma involves many different specialists and requires multidisciplinary management. In light of the current discussion on the role of ablative therapy, the aim of this study is to compare patients who undergo hepatic resection to those treated with radiofrequency ablation.

Methods

The procedures have been conducted in two institutes following the same methodologies. Ninety-six patients with Child–Pugh class A cirrhosis, single or multinodular hepatocellular carcinoma (HCC) and a diameter less than or equal to 3 cm, have been included in this retrospective study: 52 patients have been treated by surgical resection and 44 by radiofrequency ablation. Patient characteristics, survival and disease-free survival have all been analysed.

Results

Disease-free survival was longer in the resection group in comparison to the radiofrequency group with a median disease-free time of 48 versus 34 months, respectively (P = 0.04, hazard ratio = 1.5, 95 % confidence interval = 0.9–2.5). In the resection group, median survival was 54 months with a survival rate at 1, 3 and 5 years of 100, 98 and 46.2 %. In the radiofrequency group, median survival was 40 months with 1-, 3- and 5-year survival rate of 95.5, 68.2 and 36.4 %.

Conclusion

The current study shows that for small HCC in the presence of compensated cirrhosis, surgical resection gives better results than radiofrequency, both in terms of overall survival, as well as disease-free survival. Further evidence is required to clarify the role of ablative therapy as a curative treatment and whether it can replace surgery.

Keywords

Hepatocellular carcinomaHCCLiver resectionRadiofrequency ablationRFA

Introduction

Hepatocellular carcinoma (HCC) is the most frequently observed primary tumour of the liver and is closely linked to chronic liver disease [1]. In recent years, screening programmes for patients at risk, and follow-up for patients with cirrhosis, have increasingly led to earlier diagnosis [2]. At the same time, new treatments are being developed, making it possible to employ different approaches according to the type of tumour and the severity of the underlying liver disease [3].

Liver transplant and hepatic resection are still the gold standard, yet an ever-decreasing percentage of patients have access to these procedures [4, 5].

As a result, a discussion has developed in recent years regarding the role of radiofrequency ablation (RFA) as an alternative to resection, particularly for small HCC [6].

Several authors have reported excellent results in terms of oncological outcomes [7, 8], to the point where some treatment algorithms have accepted RFA as a curative treatment for single and multinodular HCC, in cases up to three lesions smaller than 3 cm in diameter [9].

The current study investigates this precise category of patient and compares hepatic resection to radiofrequency ablation, analysing the results obtained from our two specialist centres from a sample of patients with Child–Pugh class A cirrhosis and single or multinodular HCC with lesions less than or equal to 3 cm.

The aim is to describe the effects of these two different approaches regarding long-term oncological outcomes.

Methods

Between January 2004 and January 2012, 96 patients with HCC have been treated and subsequently participated in a follow-up programme, managed by either the Department of Digestive Surgery and Liver Unit of St. Maria Hospital, Terni, Italy (n = 53) or the Unit of Surgery and Advanced Oncologic Therapies, Forlì Hospital, Forlì, Italy (n = 43). The patients have been included in this retrospective study based on the following criteria: no previous treatment for HCC; lesions smaller than or equal to 3 cm; Child–Pugh class A cirrhosis; and single or multiple lesions (Table 1).
Table 1

Inclusion and exclusion criteria of the study

Inclusion criteria

Exclusion criteria

Adult patients

Metastatic disease

Diagnosis of liver cirrhosis from different aetiology

Decompensated liver cirrhosis

Stage A of Child–Pugh

Stage B or C of Child–Pugh

Number of nodules ≤ 3

Number of nodules > 3

Size of lesions ≤ 3 cm

Size of lesions > 3 cm

Patients undergoing liver transplantation

Portal vein thrombosis

Arterioportal fistula

Portal hypertension

Ascites

Biliary dilatation

Platelet count < 100.000 k/μL

The presence of chronic liver disease and the severity of cirrhosis according to the Child–Pugh Score [10], have been diagnosed using clinical, laboratory and ultrasonographic parameters (coarse echo pattern).

HCC diagnosis followed the European Association for Study of the Liver consensus conference criteria [11]. Hypervascularised nodules with a diameter greater than 2 cm in the arterial phase with washout in the portal phase confirmed by two of the following imaging techniques including: four-phase multi-detector CT scan; dynamic gadolinium-based contrast-enhanced MRI; contrast-enhanced ultrasonography with second-generation agents; or the presence of alpha-fetoprotein levels greater than 400 ng/ml and lesion confirmation with just one imaging technique [12, 13].

A biopsy has been performed for tumours smaller than 2 cm when the aforementioned criteria could not be applied. During the study period, 52 patients underwent hepatic resection and 44 were treated with RFA (Table 2). Surgery was performed according to the location of the lesions and based on an accurate assessment of the remnant liver volume, ascertained by evaluating radiological estimation of the functional parenchyma to be resected.
Table 2

Characteristics of patients enrolled in the study

 

Resection (n = 52)

RFA, (n = 44)

P value

Gender

0.4

 Male

37 (71.2 %)

35 (79.5 %)

 

 Female

15 (28.8 %)

9 (20.5 %)

Age

65.6 ± 4.8

64.4 ± 6.5

0.3

Aetiology of cirrhosis

0.07

 Alcol

14 (26.9 %)

10 (22.7 %)

 

 HBV

4 (7.7 %)

11 (25 %)

 HCV

24 (46.2 %)

20 (45.5 %)

 HBV + Alcol

2 (3.8 %)

 HCV + Alcol

4 (7.7 %)

 HBV + HCV

4 (7.7 %)

3 (6.8 %)

Number of lesions

0.8

 1

22 (42.3 %)

19 (43.2 %)

 

 2

18 (34.6 %)

13 (29.5 %)

 3

12 (23.1 %)

12 (27.3 %)

Surgical procedures have been performed by making a bilateral subcostal incision and using retractors to improve accessibility. After dissection of the ligamentum teres hepatis, the liver has been mobilised. The exploratory phase has been conducted with the aid of intraoperative ultrasonography so as to define the precise position of the tumour, its margins and the resection line, in relation to the relationships between vascular and biliary structures.

The Pringle manoeuvre has been applied in 35 patients. In this case, the complete portal triad was clamped using a rubber sling. Portal triad clamping has been set to 15 min of hepatic inflow occlusion followed by 5 min of reperfusion, repeated as many times as necessary, with the final occlusion lasting a full 15 min, even where transection had already been completed.

During reperfusion periods, haemostasis was performed by delicate compression of the parenchyma. Hepatic resection has been performed using bipolar radiofrequency electrocoagulation (Ligasure®, Valley Lab, Covidien). Central venous pressure has been maintained below 5 cm H2O to reduce venous back-bleeding.

Fourteen wedge resections and 38 segmentectomies have been performed. Median blood loss was 450 ml (100–1,250 ml); median duration of surgery was 190 min (120–250 min). Histological examination revealed that all surgical preparations had tumour-free resection margins.

The RFA group comprises patients at higher risk of sacrificing a high percentage of functional liver parenchyma, with significant comorbidities, and at risk of developing severe post-operative complications.

The procedures employed ultrasonography-guided percutaneous treatment, using an expandable, electrode needle-type probe, connected to a radiofrequency generator (RITA Medical System, Mountain View, CA, USA).

An ultrasonography scan has been performed on the second day after the operation to identify areas of necrosis and possible remaining areas of tumour activity, in which case further ablation treatment was performed.

Patients from both groups underwent oncological follow-up with bimonthly clinical examinations and laboratory tests and a CT scan every 6 months. A multidisciplinary team assessed local recurrences for possible further ablative therapies (RFA), transarterial chemoembolization or surgery.

Statistical analysis

Data were analysed with a statistical computer software (MedCalc Software Version 12.2.1). Patient characteristics such as sex, age, cirrhosis aetiology and lesions number were compared in the two treatment arms by using the Chi-square test for categorical variables and the Student t test for continuous variables. P value < 0.05 was considered statistically significant.

Overall survival was computed from the day of surgery or radiofrequency ablation to the day of death or to the most recent follow-up visit. Disease-free survival (DFS) was computed from the day of surgery or radiofrequency ablation to the follow-up visit with evidence of tumour recurrence.

Overall survival and disease-free survival analyses were performed using Kaplan–Meier curves. Comparison between different groups was carried out using the log rank test. Furthermore, hazard ratio (HR) was determined with a confidence interval (CI) of 95 %. A P value < 0.05 has led to consider the results as statistically significant.

Results

Resection group

In this group, perioperative mortality was not observed. Fourteen (26.9 %) patients suffered from post-operative complications; particularly, occurred: hepatic failure (n = 2), biliary fistula (n = 1), hemoperitoneum (n = 2) and ascites (n = 9).

The mean follow-up was 51.7 months (range 28–60 months). Median survival time was 55 months (95 % CI = 47–60), with a survival rate at 1, 3 and 5 years of 100, 98 and 46.2 %, respectively. Median DFS was 48 months (95 % CI = 39–52), with 1-, 3- and 5-year DFS rates of 100, 80.2 and 26.9 %, respectively (Table 3).
Table 3

Percentage values of survival and disease-free survival

 

1 year (%)

3 years (%)

5 years (%)

Overall survival

 Resection

100

98

46.2

 RFA

95.5

68.2

36.4

Survival single HCC

 Resection

100

100

68.2

 RFA

100

89.5

63.2

Survival multiple HCCs

 Resection

100

96.6

30

 RFA

92

52

16

DFS

 Resection

100

80.2

26.9

 RFA

90.9

52.3

22.7

DFS Single HCC

 Resection

100

95.5

45.5

 RFA

100

73.7

42.1

DFS Multiple HCC

 Resection

100

70

13.3

 RFA

84

36

8

RFA group

No mortality was observed as result of the procedure. Issues related to radiofrequency treatment were negligible, and this allowed for rapid discharge of patients (1–2 days of mean hospital stay). In fact, only minor complications occurred, such as pain (12.5 %) and fever (15.2 %).

The mean number of RFA treatments for each patient was 1.5 (range 1–4) and complete necrosis was achieved in 100 % of nodules. Tumour recurred in 34 patients (77.3 %) during 5–year follow-up. Recurrence developed at the site of the treated tumour in 9 and at a different intrahepatic site in 25 of these patients.

Mean follow-up was 47.4 months (range 10–60). Median survival time was 50 months (95 % CI = 44–60) with 1-, 3- and 5-year survival rate of 95.5, 68.2 and 36.4 %, respectively. Median DFS was 48 months (95 % CI = 39–52) with 1-, 3- and 5-year DFS rates of 90.9, 52.3 and 22.7 %, respectively (Table 3).

Resection group vs. RFA group

The selection criteria used for comparison has been restricted to patients eligible for curative treatment, with Child–Pugh class A cirrhosis and lesions with a diameter less than or equal to 3 cm, so as to create two homogenous groups (Table 1). The other patient characteristics are reported in Table 2.

Percentage rates at years 1, 3 and 5 show advantages of surgical treatment over RFA in all comparisons carried out, especially over a longer period of observation (Table 3).

DFS was longer in the resection group in comparison to the RFA group with a median disease-free time of 48 versus 34 months, respectively (P = 0.04, HR = 1.5, 95 % CI = 0.9–2.5) (Table 4, Fig. 1). Analysing the data according to the number of lesions does not show significant differences for single HCC (Fig. 2), with a median disease-free time of 54 months in the resection group and 50 months in the RFA group, even if data are not statistically significant (P = 0.4, HR = 1.3, 95 % CI = 0.5–3).
Table 4

Analysis of disease-free survival

Variable

n

Median (months)

P

HR

95 % CI

Disease-free survival

0.04

1.5

0.9–2.5

 Resection

52

48

 

 RFA

44

34

Single HCC

0.4

1.3

0.5–3

 Resection

22

54

 

 RFA

19

50

Multiple HCC

0.006

2

1.1–3.7

 Resection

30

37

 

 RFA

25

30

Comparative results including different variables

N number of patients, HR hazard ratio, CI confidence interval

https://static-content.springer.com/image/art%3A10.1007%2Fs00423-012-1029-2/MediaObjects/423_2012_1029_Fig1_HTML.gif
Fig. 1

Kaplan–Meier curves of Disease-free survival (months)

https://static-content.springer.com/image/art%3A10.1007%2Fs00423-012-1029-2/MediaObjects/423_2012_1029_Fig2_HTML.gif
Fig. 2

Kaplan–Meier curves of Disease free-survival (months) for single nodular HCC

On the other hand, in the treatment of multinodular HCC (Fig. 3), there is a clear advantage of surgery in DFS (P = 0.006, HR = 2, 95 % CI = 1.1–3.7).
https://static-content.springer.com/image/art%3A10.1007%2Fs00423-012-1029-2/MediaObjects/423_2012_1029_Fig3_HTML.gif
Fig. 3

Kaplan–Meier curves of Disease free-survival (months) for multinodular HCC

Overall survival (Fig. 4) has been significantly longer in patients treated surgically (P = 0.02, HR = 1.7, 95 % CI = 1–2.9). Of particular note was the median survival of 54 months in the resection group vs. 40 months in the RFA group (Table 5). A survival analysis of the subgroups has also been conducted based on the presence of single (Fig. 5) or multinodular HCC (Fig. 6). In the former, the results were not statistically significant (P = 0.6), probably due to the small sample size in both groups, while comparative analysis for multinodular tumours showed excellent results for resection compared to RFA (P = 0.001, HR = 2.4, 95 % CI = 1.2–4.6).
https://static-content.springer.com/image/art%3A10.1007%2Fs00423-012-1029-2/MediaObjects/423_2012_1029_Fig4_HTML.gif
Fig. 4

Kaplan–Meier curves of overall survival (months)

Table 5

Analysis of survival

Variable

n

Median (months)

P

HR

95 % CI

Overall survival

0.02

1.7

1–2.9

 Resection

52

54

 

 RFA

44

40

Single HCC

0.6

1.2

0.4–3.7

 Resection

22

60

 

 RFA

19

56

Multiple HCC

0.001

2.4

1.2–4.6

 Resection

30

46

 

 RFA

25

38

Comparative results including different variables

N number of patients, HR hazard ratio, CI confidence interval

https://static-content.springer.com/image/art%3A10.1007%2Fs00423-012-1029-2/MediaObjects/423_2012_1029_Fig5_HTML.gif
Fig. 5

Kaplan–Meier curves of survival (months) for single nodular HCC

https://static-content.springer.com/image/art%3A10.1007%2Fs00423-012-1029-2/MediaObjects/423_2012_1029_Fig6_HTML.gif
Fig. 6

Kaplan–Meier curves of survival (months) for multinodular HCC

Discussion

Treating hepatocellular carcinoma in patients with chronic liver disease has always presented a challenge because of the clinical complexity of managing these patients and the potential risks associated with post-operative complications [14]. The approach to treating HCC has become multidisciplinary, involving a team of experts from institutes dedicated to the study, treatment and follow-up of hepatic tumours [15]. The aim is to prevent liver failure, which can occur after a wide or narrow surgical resection with related morbidity and mortality. This has led to the development of new technologies for the treatment of hepatocellular carcinoma and to the introduction of percutaneous ablation in clinical practise [16]. Radiofrequency has proven to be reliable, safe and easy to use, with negligible post-operative complications and offering the possibility to effect complete necrosis of the lesion with one single procedure [17].

Which treatment is most suitable for patients with relatively unimpaired hepatic function (Child–Pugh class A) is currently subject to much debate. Surgery has shown excellent results in long-term outcomes but can lead to significant mortality rates of up to 5 % in specialised centres (none recorded in our study) and to the development of post-operative complications (26.9 % in our study) [18, 19]. By contrast, radiofrequency is a far less invasive procedure, only requires a brief hospital stay and has fewer severe complications associated with the procedure and a low mortality risk [20].

Although the role of RFA when hepatic resection is not possible or as a bridge to liver transplantation is well defined, its precise indications and its use as a replacement for surgery are still open to debate [21]. In particular, RFA has recently been suggested as an alternative to resection for class A patients (Child–Pugh Score), with lesions up to 3 cm in diameter [22].

However, there are few studies that report long-term results, and results that compare hepatic resection with RFA even fewer.

The many variables to consider make it difficult to conduct studies on the treatment of hepatocellular carcinoma in the presence of cirrhosis because tumour characteristics are invariably linked to chronic liver disease characteristics.

It is a very controversial topic, and conflicting results are emerging from the published literature based on the findings of different clinical trials. Vivarelli [23] has retrospectively analysed the results obtained from 158 cirrhotic patients. The findings are favourable towards surgical treatment with a survival rate after 1 and 3 years of 83 and 65 % vs. 78 and 33 %, respectively, and a disease-free survival rate after 1 and 3 years of 79 and 50 % vs. 60 and 20 %, respectively. In particular, the author has noted major advantages to surgery in patients with Child–Pugh class A cirrhosis. A study by Abu Hilal [24] suggests that resection is the best treatment for HCC below 5 cm, and our study supports this conclusion, at least for the more restricted range of patients with lesions less than or equal to 3 cm.

By contrast, Wakai [25] has noted that results obtained from ablative therapy are comparable to surgery for the treatment of small hepatocellular carcinoma. However, the maximum diameter considered for this study is 2 cm, and the role of radiofrequency is not entirely clear because it is studied together with other percutaneous ablation treatments (PEI n = 37, RFA n = 21, microwave coagulation n = 6).

Our study confirms that surgery presents better results than RFA, and that these differences are also evident for HCC smaller than 3 cm. This study is limited, however, by the fact that it is retrospective with a relatively small patient sample. However, we do not believe that these factors have influenced the statistical analysis. In contrast to the majority of studies presented in the literature on this topic, we have created two uniform groups—by dimension (<3 cm) and severity of cirrhosis (Child class A)—with the aim of producing a focussed study and contributing to the current discussion on the management of hepatocellular carcinoma.

We have thus decided to include patients undergoing surgical or ablative curative treatments, choosing a diameter of 3 cm as the cut-off point, as RFA results have demonstrated the highest tumour necrosis rates on tumours below this size [26].

The results highlight that, in terms of overall survival, surgery is better than RFA. This difference is particularly evident over the long term. One year after treatment, survival within the resection group and within the RFA group was 100 and 95.5 %, respectively. However, after 3 and 5 years, survival was 98 vs. 68.2 % and 46.2 vs. 36.4 %, respectively.

From an analysis of the subgroups, the advantage gained from surgical treatment in patients with multinodular HCC is clear and statistically significant (P = 0.001, HR = 2.4, 95 % CI = 1.2–4.6). The difference between the two treatments is further confirmed by an analysis of the DFS

(P = 0.006, HR = 2, 95 % CI = 1.1–3.7). These significant findings could result from the possibility of removing, through hepatic resection, small peritumoral, satellite lesions and microvascular invasion.

Interpreting data from a subgroup comprising single HCC is more complex. The analysed data show a modest difference between the two treatments, both in terms of DFS (HR = 1.3, 95 % CI = 0.5–3), as well as survival (HR = 1.2, 95 % CI = 0.4–3.7). However, this comparison has not yielded statistical significance (P > 0.05), probably because of the small patient sample available (resection group n = 22 vs. RFA group n = 19). Nevertheless, the general trend identified supports the findings of other studies presented in the literature.

In Nishikawa’s retrospective study [27], which focused on patients with HCC < 3 cm, no significant difference between the two treatment groups was identified, either in terms of overall survival or DFS. The authors concluded that RFA should be considered a potential alternative for the treatment of single and small HCC. For patients with Child–Pugh class A cirrhosis, Huang [28] reports better results from surgery in cases of single HCC larger than 3 cm but less than 5 cm, or with two or three lesions each less than 5 cm, and adds that in patients treated with single HCC < 3 cm, RFA has a comparable disease-free survival to surgical resection.

In conclusion, hepatic surgery has made great progress over the past 10 years, thanks to improvements in the instruments used and anaesthetic techniques, with a dramatic reduction in perioperative mortality. As a result, resection has remained the treatment of choice for HCC where residual hepatic functional allows.

Our study has shown that RFA is a valid and effective treatment, especially for small and single HCC, but at present, there is not sufficient evidence to propose radiofrequency ablation for those patients who could benefit from resection.

Conflicts of interest

The Authors state that none of the authors involved in the manuscript preparation has any conflicts of interest towards the manuscript itself, neither financial nor moral conflicts. Besides, none of the authors received support in the form of grants, equipment and/or pharmaceutical items.

Copyright information

© Springer-Verlag Berlin Heidelberg 2012