Nov 9, 2010 - Abstract. Background Recently randomized controlled trials have been advocated to compare radiofrequency ablation (RFA) and.
J Gastrointest Surg (2011) 15:378–387 DOI 10.1007/s11605-010-1377-6
REVIEW ARTICLE
Radiofrequency Ablation Versus Resection for Liver Tumours: An Evidence-Based Approach to Retrospective Comparative Studies Gianpiero Gravante & John Overton & Roberto Sorge & Neil Bhardwaj & Matthew S. Metcalfe & David M. Lloyd & Ashley R. Dennison
Received: 19 September 2010 / Accepted: 19 October 2010 / Published online: 9 November 2010 # 2010 The Society for Surgery of the Alimentary Tract
Abstract Background Recently randomized controlled trials have been advocated to compare radiofrequency ablation (RFA) and hepatic resection (HR) in resectable tumours and determine whether differences in observed survivals result from the heterogeneity in previous studies between RFA (treating unresectable lesions) and HR (treating lesions deemed resectable). We reviewed the literature that directly compares the treatments and employed an evidence-based approach to examine the data. Materials and Methods All studies comparing RFA and HR were included. Primary outcomes were the overall survival (OS) and disease-free survival (DFS) at 3 and 5 years. A subgroup analysis was conducted for solitary or small tumors (1=35%
1 1 1=76%; >1=24% 1=59%; >1=40% –
Diameter
– – Open/perc. – Lap. – Perc. – – – Perc. – Perc. – –
– Open/Lap./Perc. – Perc. –
Treatment Approach
OS overall survival; DFS disease-free survival, CC case control; Prosp prospective study not randomized, RCT randomized control trial
2006 RCT
Chen34
Santambrogio32 2009 CC
2005 CC
Ogihara28
Number
Type of study Patients Age (years) Sex (males; %) Tumour characteristics
Year
Author
Table 2 Characteristics of comparative studies for hepatocellular carcinoma OS (years)
64 47 36 – – 69 60 37 31 55 40 50 20 – –
– – 56 22 69 – 38 20 – – – – – – – – – – – –
– – 27 22 –
71 92 92 85 66 73 69 77 80 84 73 65 33 – –
65 58 64 42 73
– 80 63 54 41 – – – – – – – – – –
31 39 48 20 –
3 (%) 5 (%) 3 (%) 5 (%)
DFS (years)
J Gastrointest Surg (2011) 15:378–387 381
382
J Gastrointest Surg (2011) 15:378–387
to collect all the relevant data, including patients’ demographics, technical aspects, and outcome measures. Data collection was carried out independently by two researchers (GG and OJ) and then compared. Primary outcome measures were the OS and DFS at 3 and 5 years of follow-up after treatment. A subgroup analysis was conducted when possible for patients affected by solitary tumours or where the maximal tumour diameter was less than 4 cm for CRM and 5 cm for HCC.
RFA to treat unresectable liver tumours including those that presented specific data on solitary lesions or small tumours (Table 3). HR was markedly superior to RFA for 3- and 5year OS as well as 5-year DFS (Table 3, Fig. 1). Results also confirm the superiority of HR over RFA for OS at 5 years of follow-up in tumours smaller than 4 cm and for solitary lesions (Table 3). There were not enough studies to provide results for the DFS at 3 years or for OS at 3 years in small tumours.18,39
Statistical Analysis
Survival of Patients with Hepatocellular Carcinoma
Data analysis was performed using the Statistical Package for the Social Sciences Windows version 13.0 (SPSS, Chicago, IL, USA) and the meta-analysis with Interactive eXplanations (MIX–version 1.6) program. Descriptive statistics for qualitative variables was performed with occurrences and described with relative frequencies, for quantitative parametric variables with the mean and standard deviation and for the quantitative non-parametric variables with the median and range. The odds ratio and 95% confidence intervals in the RFA and HR group were evaluated. Results were considered significant if the probability of a chance occurrence was less than 5% (p< 0.05).
HR was markedly superior to RFA for 3- and 5-year OS as well as 3-year DFS (Table 4, Fig. 2). The subgroup analysis showed better OS at 3 years for HR compared to RFA in solitary lesions and DFS at 3 years for small tumours (Table 4, Fig. 2). No significant differences were found at 3-year OS for small tumours (Table 4, Fig. 2). There were not enough studies to provide definitive results for 5-year DFS29,31 or for the subgroup analysis of OS at 5 years and DFS at 3 years for solitary lesions.
Results Since 2003, numerous articles have compared the results of RFA and HR for CRM and HCC (Tables 1 and 2).18,21–40 All studies were retrospective except for three which were prospective, 27,30,34 including two randomized trials.30,34 Some studies reported specific data for solitary 1 8,2 2, 24, 26 ,2 8,3 2, 34, 36 ,3 9,4 0 or small tumours.18,22,26,30–32,34–36,40 A study based on a theoretical mathematical analysis using the Markov modelling to simulate a randomized trial of RFA and HR for HCC less than 5 cm was also retrieved and discussed, but its numeric data were not included.41 Two studies reported OS and DFS for shorter follow-ups,33,37 one study had a significant higher percentage of patients that underwent previous HR in the RFA group and presented lower RFA survival rates compared to the others studies examined.40 The descriptive data of these three studies are still presented for completeness (Tables 1 and 2) but have not been included (Tables 3 and 4). One article published in a Chinese journal was not available.42 Survival of Patients with Colorectal Metastases At the moment, there are no 5-year survival data available for RFA employed for resectable CRM. All studies used
Discussion Currently, liver resection is the gold standard treatment for resectable liver tumours but is not possible or appropriate in up to 80% of cases due to a low predicted hepatic reserve, significant co-morbidity or technical issues related to the location, number or size of the lesions. Intuitively, RFA presents a valid alternative to hepatic resection on many levels, especially by improving the OS compared to standard chemotherapy or palliative treatments. Despite this, overall survivals at 5 years still do not match those of HR and these outcome differences have been attributed to the fact that HR patients had resectable lesions while those treated by RFA were unresectable.12 It is this explanation, which has been taken by some authors to imply that in matched patients results with HR and RFA would be similar, that has resulted in some units advocating a randomized prospective trial for resectable lesions.12 If proven, the advantage of a minimal invasive technique, with the greater preservation of liver, reduced complications and shorter hospital stays would expand the indications considerably. To date, there are a few sporadic reports of curative rather than palliative treatments with RFA for resectable liver lesions.43,44 Two randomized studies on solitary HCC measuring less than 5 cm appear to confirm similar OS with HR and RFA, although the follow-up is only 3 years.30,34 In the retrospective study of Hasegawa et al. 37 although for HCC higher recurrence rates were found for RFA, OS was similar to HR for tumours of less than 3 cm. We did not include this study due to the length of the
J Gastrointest Surg (2011) 15:378–387
383
Table 3 Results of the meta-analysis for colorectal metastases Parameter
Time interval
Tumour characteristics
RFA
HR
OS
3 years
– Solitary Small (
