Radiofrequency ablation for hepatocellular carcinoma

Abstract

Hepatocellular carcinoma (HCC) is one of the most common causes of cancer-related mortality worldwide. Unfortunately, only 20% of HCC patients are amenable to curative therapy (liver transplantation or surgical resection). Locoregional therapies such as radiofrequency ablation (RFA), percutaneous ethanol injection, microwave coagulation therapy, and transcatheter arterial chemoembolisation play a key role in the management of HCC. The choice of the treatment modality depends on the size of the tumour, tumour location, anatomic considerations and the number of tumours present and liver function. RFA therapy for HCC can be performed safely using a percutaneous, laparoscopic, or an open approach, even in patients with poor functional reserve. Since the introduction of RFA, several randomised controlled trials and non-randomised studies comparing RFA and other therapies for HCC have been conducted. In addition, in the last decade there have been technical advances in RFA therapy for HCC, resulting in significant improvement in the prognosis of HCC patients treated with this modality. In this review, we primarily focus on percutaneous RFA therapy for HCC and refer to current knowledge and future perspectives for this therapy. We also discuss new emerging ablation techniques.

• Hepatocellular carcinoma
• radiofrequency ablation
• treatment

Introduction

Hepatocellular carcinoma (HCC), which accounts for more than 90% of primary liver cancer, is a major health problem worldwide and is the third most common cause of cancer-related death [1–6]. It is the fifth most prevalent cancer in men and the seventh in women. HCC generally arises in the setting of an existing condition such as chronic hepatitis or liver cirrhosis [1–6]. It is highly prevalent in Asia and Africa and is also increasing in prevalence in Western countries, with an estimated incidence ranging between 500,000 and 1,000,000 new cases annually [1–6]. Curative treatments for HCC consist of surgical resection (SR) and liver transplantation [3–6]. However, in contrast to other solid cancers, SR plays a limited role in the treatment of HCC. SR is precluded in the majority of HCC patients because of the size of the tumour, the number of tumours present, and poor hepatic functional reserve. Only 10–20% of HCC patients are candidates for SR [7–9]. Liver transplantation has been performed in HCC patients who fulfill the Milan criteria, namely those patients who have solitary HCC tumours of <5 cm in diameter or up to three nodules of <3 cm in diameter [10]. However, the availability of liver transplantation is limited by the shortage of organ donors [5,10].

Despite recent advances and technical refinements in HCC therapy, HCC frequently recurs even after curative therapy, resulting in high mortality; HCC recurrence is evident at intrahepatic sites in 50–96% of patients [5,6,11–13]. Conversely, locoregional therapies for HCC such as transarterial chemoembolisation (TACE) and local ablative therapies have evolved over the past 20 years [3,5,6,9,14,15]. Tumour ablation can be achieved by modifying the temperature of cancer cells (microwave ablation, cryoablation, and radiofrequency ablation (RFA)) in the tumour [16–19]. At present, RFA is well established as the standard local ablative therapy for HCC, and can be used as an alternative to surgery because of its superior local control rates and patient survival relative to other local treatments [20–27].

Regarding systemic chemotherapy, numerous clinical trials have been performed in an attempt to establish effective treatment for patients with advanced HCC; however, no agent has reproducibly been shown to achieve a high response rate, and no chemotherapeutic regimens have demonstrated superior survival benefit over placebo for advanced HCC [28–32]. Recently, because of the fact that two randomised studies, namely the SHARP trial and the Asia-Pacific trial, demonstrated the efficacy of sorafenib (a multi-kinase inhibitor that blocks tumour growth and cell proliferation), molecular targeting therapy with this agent has become the first-line systemic chemotherapy for patients with unresectable HCC [33,34]. In May 2009, sorafenib was approved in Japan for the treatment of unresectable HCC [35–37].

Since the introduction of RFA in our institution, we have performed percutaneous RFA in approximately 4800 patients with liver tumours from 1999 to the present time. In this review, we primarily focus on percutaneous RFA therapy for HCC and refer to current knowledge and future perspectives for this therapy. Furthermore, we also discuss new, emerging, promising ablation techniques.

Indications and long-term outcomes for RFA

RFA therapy, an alternative modality to percutaneous ethanol injection (PEI) that has been used in Japan since 1999, has been widely employed as a curative treatment for HCC. Now, RFA is considered the most promising locoregional treatment for HCC [38–42]. This modality induces coagulative necrosis and tissue desiccation by delivering high frequency alternating current via electrodes placed within tissues [38–42]. RFA for HCC is mainly accomplished by a percutaneous approach, although open laparoscopic or thoracoscopic approaches can also be used [38–40]. RFA provides a valuable treatment option for unresectable HCC. As progress in RFA therapy for HCC such as technical advancement and improvement of treatment efficacy continues to be made, it is gradually being performed in patients with resectable HCC [41]. In addition, RFA is a repeatable procedure because it is less invasive than SR, and it can be safely performed in elderly patients with potentially co-morbid diseases [43–45]. Another advantage of percutaneous RFA is that it does not require general anaesthesia [43–45].

Recently, an algorithm for HCC treatment has been proposed by the Japanese Society of Hepatology [46]. According to this algorithm, the treatment of HCC depends on tumour size, liver damage as defined by the Child-Pugh classification, the number of tumours detected, and the presence or absence of distant metastasis; a treatment strategy for each HCC case can be recommended using this algorithm [46]. Currently, three or fewer tumours with a diameter of ≤3 cm and no extrahepatic lesions, well-preserved liver function, no intractable ascites, and no vascular invasions are generally considered as indications for RFA therapy; however, the typical treatment algorithms for HCC in Japan, Europe, and North America are each slightly different [42,46–48]. In Europe and North America, the algorithm for the treatment of HCC proposed by the American Association of the Study of Liver Disease (AASLD) recommends local ablative therapy for three or fewer early-stage HCC tumours with diameters of ≤3 cm and very early-stage HCC tumours with diameters of ≤2 cm, with complications such as portal hypertension [48].

In terms of long-term survival after RFA, Shiina et al. conducted a large retrospective study regarding RFA therapy for primary HCC [49]. It included a total of 2,982 RFA treatments for the 1170 primary HCC patients; 422 patients had tumours ≤2.0 cm in diameter, 467 had tumours 2.1–3.0 cm in diameter, 246 had tumours 3.1–5.0 cm in diameter, and 35 had tumours >5.0 cm in diameter (maximum tumour size (mean ± standard deviation), 2.54 ± 1.04 cm). A single tumour was present in 685 patients, two or three tumours in 395, and four or more tumours in 90. These authors reported that with a median follow-up of 38.2 months, 5-year overall survival (OS) and recurrence-free survival (RFS) rates were 60.2% and 22%, respectively, and the incidence rates of complications per treatment was 2.2% (67/2982) and the treatment mortality rate was 0.03% [49]. Furthermore, they also reported that in the Child-Pugh class A or B patients within Milan criteria, the 5-year OS rate was 63.8%, while in those outside Milan criteria, 5-year OS rate was 46.4% [49]. Multivariate analysis in their study demonstrated that age, antibody against hepatitis C virus, Child-Pugh class, tumour size, tumour number, serum des-γ-carboxy-prothrombin level, and serum lectin-reactive α-fetoprotein level were significantly associated with OS [49]. On the other hand, in our study (n = 368: mean tumour size, 2.0 ± 0.7 cm; mean observation period, 3.0 ± 2.0 years; mean number of treatment sessions, 1.25 ± 0.48), the 5-year OS and RFS rates were 59.8% and 18%, respectively, which were comparable to their data, the incidence rate of major complications per treatment was 2.2% (6/368), and the treatment mortality rate was 0% [50]. In our multivariate analysis, tumour size >2 cm, serum albumin >3.5 g/dL, prothrombin time >70%, HCC recurrence within 1 year after initial RFA and an ablative margin ≥5 mm were found to be significant independent factors linked to OS [50]. Prognostic factors were slightly different between these two studies.

Complications associated with RFA

In 2002, Mulier et al. reported that complication rates for percutaneous, laparoscopic, and open RFA of hepatic tumours in 3670 patients were 7.2%, 9.5%, and 9.9%, respectively; the corresponding mortality rates were 0.5%, 0%, and 0%, respectively [51]. In view of previous studies, the frequency of major complications of percutaneous RFA (severe haemorrhage, RFA needle-track seeding, formation of abscess, perforation of gastrointestinal tract, liver failure, biloma, biliary stricture, portal vein thrombosis, haemothorax or pneumothorax requiring drainage, and others) has been reported to range from 0.6 to 8.9% [20–24, 52–55].

In 2007, the Osaka Liver Cancer Study Group in Japan reported that in a total of 3,891 RFA therapies in 2614 HCC patients, complications were observed in 207 of 2614 patients (7.9%), and that nine (0.3%) died within 3 months of undergoing RFA; the departments that treated larger numbers of patients per month had a smaller number of complications and deaths (p-value not shown) [55]. The group concluded that complications associated with RFA can be reduced when institutions have gained sufficient experience in the use of this therapy [55]. Clinicians should be alert to all of the features of RFA-related complications, although the need to obtain a sufficient ablative margin can increase the total number of RFA electrode insertions, and this may have led to an increase in RFA-related complications.

Assessment of the treatment efficacy of RFA

Evaluation of the therapeutic effects of RFA is very important. The treatment efficacy of this modality is commonly assessed using findings from contrast-enhanced computed tomography (CT) or magnetic resonance imaging (MRI) [50,56–58]. In our department, to assess the treatment efficacy of RFA for HCC, we have routinely performed dynamic 16-column multi-detector computerised tomography (MDCT) using 3-mm slice scans within 1 week of undergoing RFA. The patients treated with RFA are then classified into four groups according to dynamic CT findings as follows: Grade A (absolutely curative), a ≥5-mm ablative margin around the entire tumour; Grade B (relatively curative), an ablative margin around the tumour but <5 mm in diameter in some places; Grade C (relatively non-curative), only an incomplete ablative margin around the tumour although no residual tumour was apparent; Grade D (absolutely non-curative), the tumour was not completely ablated [58]. We refer to this grading system as radicality grading (R grades: A, B, C, and D) [58]. In a recent study we demonstrated that a more complete and a larger ablative margin was associated with a lower local tumour progression (LTP) rate, and that this grading system was useful for predicting LTP after RFA (cumulative LTP rates at 1, 2 and 3 years: 2.6%, 2.6% and 6.7% for Grade A, 2.4%, 10.2% and 17.6% for Grade B, 19.5%, 44% and 55.8% for Grade C and 46.2%, 53.4% and 82.2% for Grade D (p < 0.001) [58]. In addition, in another study we carried out to evaluate OS, the RFS rate and the distant recurrence rate for each R grade (A–D), and to examine the relationship between clinical outcome and R grading, we demonstrated that our proposed R grading system was useful for predicting not only LTP but also clinical outcomes after initial RFA (cumulative OS rates at 1, 3 and 5 years: 98.6%, 90.3% and 79.5% for Grade A, 95.6%, 75.3% and 66.7% for Grade B, 95.8%, 73.3% and 44.8% for Grade C and 88.5%, 63.8% and 35% for Grade D (p = 0.008); cumulative RFS rates at 1, 3 and 5 years: 88%, 44.2% and 25.9% for Grade A, 88%, 44.2% and 23.5% for Grade B, 69%, 21.9% and 13.3% for Grade C and 49.1%, 8.2% and 4.5% for Grade D (p < 0.001); cumulative intrahepatic and/or extrahepatic distant recurrence rates at 1, 3 and 5 years: 11%, 52.7% and 71% for Grade A, 11%, 54.1% and 72.9% for Grade B, 26%, 72.5% and 83.2% for Grade C and 34.1%, 76% and 88% for Grade D (p < 0.001) [50].

Conversely, there have been several studies regarding the usefulness of MRI for assessing the ablative margin of RFA for HCC [56,57,59]. Koda et al. evaluated the ablative margin of RFA in 101 HCCs using MRI with ferucarbotran administration. These authors reported that the cumulative LTP rates (4.4%, 7.6%, and 7.6% in 1, 2, and 3 years, respectively) in 47 nodules with positive ablative margins were significantly lower than those (13.9%, 33.4%, and 41.8% in 1, 2, and 3 years, respectively) in 36 nodules without positive ablative margins [59]. Multivariate analysis identified contiguous vessels (odds ratio, 12.0) and a positive ablative margin on MRI (odds ratio, 0.19) as independent predictors of LTP; the authors reached the conclusion that assessment of the ablative margin of RFA by MRI using ferucarbotran was a useful method for predicting LTP [59].

Comparisons of clinical outcomes between RFA and other therapies

RFA versus percutaneous ethanol injection

PEI, which involves the injection of absolute ethanol directly into targeted tumours through fine needles under guidance of ultrasonography, has been widely used as a standard local ablative therapy for small HCCs since its introduction in Japan in the 1980s [19–24,52,53,60–62]. However, in many cases its treatment efficacy is unpredictable because the spread of injected ethanol within the targeted tumour is largely affected by the capsule or septa of the tumour, whereas in RFA therapy heat generated around the electrode tip distributes homogenously in all directions, although the energy deposition of RFA is affected by the electron sink effect caused by regional blood vessels [20–24,52,53,60–64].

In several randomised controlled trials (RCTs), the number of treatment sessions, complete therapeutic effect, LTP rate, OS rate, RFS rate, and rate of severe complications were compared between RFA and PEI groups [20–24]. Shiina et al. reported in their largest RCT that the 4-year survival rate was 74% in the RFA group and 57% in the PEI group (p = 0.01), and that RFA was associated with a 46% lower risk of death, a 43% lower risk of overall recurrence, an 88% lower risk of LTP relative to PEI, and the incidence of serious adverse events was no different between the two groups (p = 0.54); they concluded that RFA was superior to PEI for small HCCs in terms of its treatment efficacy and clinical outcomes [23]. Conversely, a meta-analysis by Germani et al. revealed that RFA did not have significantly higher therapeutic efficacy than PEI for tumours of ≤2 cm in diameter, although it was concluded that for tumours of >2 cm RFA was superior ablative therapy to PEI [53].

The higher level of local tumour control achieved using RFA as compared with PEI seems to be due to the more expansive coagulative effects of thermal ablation on the targeted HCC nodules and micro-satellites surrounding the nodules [20–23]. The survival rate data indicated a significant benefit for RFA over PEI [20–23]. This higher survival rate may be due to the higher rate of complete tumour response using RFA than using PEI; an initial complete response is an independent predictive factor linked to survival [50,62]. However, in general the rate of major procedure-related complications was higher with RFA than with PEI [54]. We should consider the local ablative therapies as part of the overall risk/benefit profile in each individual case [54]. In selected cases of small HCC, when SR cannot be offered and RFA is not warranted for reasons such as tumour location, PEI should be taken into account. Previously reported RCTs involving comparison of RFA and PEI for the treatment of HCC are listed in Table I.

Table I. Reports of randomised controlled trials of comparison between RFA and PEI for hepatocellular carcinoma.

								OS rate (%)		RFS rate (%)
Author, country	Year	Treatment	No. of patients	Tumour size (cm)	Tumour number	Follow-up (months)	Initial CR (%)	1 year	3 years	1 year	2 years
Lencioni et al. [20], Italy	2003	RFA	52	2.8 (mean)	≤3 nodules	22.9 (mean)	91	100	81	86*	64*
		PEI	50	2.8 (mean)		22.4 (mean)	82	96	73	77	43
Lin et al. [21], Taiwan	2004	RFA	52	2.9 (mean)	≤3 nodules	22.4 (mean)	96	90*	74*	78*	59*
		PEI	52	2.8 (mean)		22.4 (mean)	88	85	50	61	42
		Higher dose PEI	53	2.8 (mean)		22.4 (mean)	NA	88	55	63	45
Lin et al. [22], Taiwan	2005	RFA	62	2.5 (mean)	≤3 nodules	28 (mean)	97	93*	74*	74*	60*
		PEI	62	2.3 (mean)		27 (mean)	89	88	81	70	41
		PAI	63	2.3 (mean)		27 (mean)	NA	90	53	71	43
Shiina et al. [23], Japan	2005	RFA	118	≤2 (38%)	≤4 nodules	37.2 (median)	100	90	80*	NA	NA
				>2 (62%)		NA				NA
		PEI	114	≤2 (50%)		34.8 (median)	100	82	63	NA	NA
				>2 (50%)		NA				NA
Brunello et al. [24], Italy	2008	RFA	70	2.42 (mean)	≤3 nodules	26.1 (median)	96	88	59	NA	NA
		PEI	69	2.25 (mean)		25.3 (median)	66	96	57	NA	NA

RFA, radiofrequency ablation; PEI, percutaneous ethanol injection; CR, complete response; OS, overall survival; RFS, recurrence free survival; NA, not available. *Statistically significant.

RFA versus microwave coagulation therapy

To our knowledge, there have been only one RCT and a few retrospective comparative studies comparing RFA with microwave coagulation therapy (MCT) [65–67]. These studies reported that RFA had a very similar or slightly superior efficacy regarding local tumour control rates, major procedure-related complications, and survival relative to MCT [65–67]. Furthermore, Livraghi et al. recently reported that in 736 patients with 1,037 tumours treated with new technologies for MCT, major complications occurred in 22 patients (2.9%) and minor complications in 54 patients (7.3%); MCT-related complications did not differ from those associated with RFA [68]. However, the ablated area produced by MCT is usually smaller than that produced by RFA, although the advantage of MCT over RFA is that treatment efficacy is less affected by the presence of vessels in proximity to the targeted tumour [69,70]. Hence, MCT requires more treatment sessions to obtain complete therapeutic effects as compared with RFA.

RFA versus surgical resection

An open question is whether or not RFA can compete with SR as a first-line therapy for patients with small HCC. However, no clear consensus has been reached as to which of these two therapies is the best for small HCC eligible for surgery. Three RCTs and several non-RCTs that have compared RFA with SR have been reported [25–27,43,71–81].

Zhou et al. conducted a meta-analysis to assess the efficacy of RFA and SR in the treatment of HCC [81]. They concluded that SR was superior to RFA for patients with small HCC tumours of >3 cm that were eligible for surgery; however, for tumours of ≤3 cm, survival levels did not differ significantly between SR and RFA. Feng et al. reported in their latest RCT that in patients with small HCC with nodular diameters of <4 cm and up to two nodules (n = 168), percutaneous RFA may provide therapeutic effects similar to those of SR (cumulative OS rates at 1 and 3 years: 96% and 74.8% for SR and 90.6% and 76.7% for RFA (p value, not significant); cumulative RFS rates at 1 and 3 years: 93.1% and 61.1% for SR and 86.2% and 49.6% for RFA (P value, not significant) [27]. Conversely, Huang et al. reported in their RCT that SR may provide better survival and lower recurrence rates than RFA for patients with HCC conforming to the Milan criteria (cumulative OS rates at 1, 2, 3, 4 and 5 years: 98.26%, 96.52%, 92.17%, 82.6% and 75.65% for SR and 86.96%, 76.52%, 69.57%, 66.09% and 54.78% for RFA (p = 0.001); cumulative RFS rates at 1, 2, 3, 4 and 5 years: 85.22%, 73.92%, 60.87%, 54.78%, 51.30% for SR and 81.74%, 59.13%, 46.08%, 33.91%, 28.69% and 85.22% for RFA (p = 0.017) [26]. In view of these results it is therefore still unclear whether or not SR can achieve higher survival rates than RFA for patients with resectable HCC. In addition, all of the previous three RCTs were Chinese based [25–27]. In comparing the results of these RCTs with those from Japanese-based studies, the mean age of the Chinese patient population was around 10 years younger than that of the patients in the Japanese studies [43,77,80,81]. In the aetiology of liver disease in the Chinese studies, patients with chronic hepatitis B were in the majority [25–27]. However, in the Japanese studies, patients with chronic hepatitis C were in the majority; hence, the study results did not reflect the actual situation in Japan where Japanese HCC patients consist of many elderly patients, and the aetiology of background liver disease involves chronic hepatitis C, which accounts for about 80% of Japanese HCC patients [43,77,80,81]. We also believe that anti-viral therapy for background liver diseases (i.e. nucleotide analogue therapy for hepatitis B and interferon therapy for hepatitis C) and nutritional supporting therapy for improving liver function such as branched-chain amino acid therapy, as well as tumour-related factors such as HCC stage, tumour size, and tumour number should be taken into account in HCC patients with poor hepatic functional reserve when assessing clinical outcomes after initial therapy for HCC [82–84]. Therefore, we should interpret these study results with caution. In Japan, an RCT (SURF trial) comparing survival between surgery and RFA for patients with resectable HCC of <3 cm in size and up to three nodules is underway [85]. RCTs involving comparison between SR and RFA for resectable HCC are listed in Table II.

Table II. Reports of randomised controlled trials of comparison between SR and RFA for resectable hepatocellular carcinoma.

							OS rate (%)		RFS rate (%)
Author, country	Year	Treatment	No. of patients	Tumour size (cm)	Tumour number	Follow-up (months)	1 year	3 years	1 year	2 years
Chen et al. [25], China	2006	Resection	90	≤3 (46.7%), 3.1–5.0 (53.3%)	Single	29.2 (mean)	93.3	73.4	86.6	76.8
		RFA	71	≤3.0 (52.1%), 3.1–5.0 (47.9%)		27.9 (mean)	94.4	68.6	90.8	68.6
Huang et al. [26], China	2010	Resection	115	>3.0/≤3.0 (44/45, single)	≤3 nodules	46.4 (median)	98.3	92.2*	85.2	73.9*
		RFA	115	>3.0/≤3 (27/57, single)		37.2 (median)	87	69.6	81.7	59.1
Feng et al. [27], China	2012	Resection	84	less than 4	≤2 nodules	36	96	74.8	93.1	67.2
		RFA	84	less than 4		36	90.6	76.7	86.2	66.6

SR, surgical resection; RFA, radiofrequency ablation; OS, overall survival; RFS, recurrence free survival. *Statistically significant.

RFA vs. combination of RFA and other locoregional therapies

RFA versus RFA plus transarterial chemoembolisation

Larger tumour size has been reported to be an adverse predictive factor linked to LTP after RFA [86–93]. However, the limited volume of coagulative necrosis obtained using RFA, and the occasionally irregular ablated shape of the major vessels in proximity to the ablated area due to the heat-sink effect, have prevented the widespread use of RFA for large HCCs [86–93]. To obtain a larger ablative margin in RFA for HCC, RFA has been combined with adjuvant transcatheter techniques such as temporary occlusion of the blood supply of tumours and TACE [86–93]. Because the blood supply to a classical HCC is mainly provided by the hepatic artery, the effect of RFA can be enhanced by occlusion of the hepatic artery [86–94].

A combination therapy involving TACE followed by RFA has been performed with the aim of minimising heat loss due to perfusion-mediated tissue cooling and increasing the treatment efficacy of RFA [88–90,92]. Morimoto et al. reported in their RCT that in patients with intermediate-sized HCCs (tumour diameter, 3.1–5.0 cm in the greatest dimension), the mean diameters of the longer and shorter axes of the RFA-induced ablated areas were 50 ± 8 mm and 41 ± 7.1 mm, respectively in the RFA group, and 58 ± 13.2 mm and 50 ± 11.3 mm, respectively in the TACE-RFA group (the mean diameters of the shorter axes in the two groups were significantly different (p = 0.012); the LTP rates at 3 years in the RFA and TACE-RFA groups were 39% and 6%, respectively (p = 0.012) [89]. Thus, they reached the conclusion that RFA combined with TACE was more effective than RFA alone for extending the ablated area and for decreasing the LTP rate [89]. In contrast, Shibata et al. demonstrated in their RCT that in 89 patients with 93 HCC nodules of 0.8–3.0 cm in diameter (46 patients with 49 nodules in the combined RFA and TACE group and 43 patients with 44 nodules in the RFA-alone group), the 1-, 2-, 3-, and 4-year LTP rates were 14.4%, 17.6%, 17.6%, and 17.6%, respectively, in the combined RFA and TACE group and 11.4%, 14.4%, 14.4%, and 14.4%, respectively, in the RFA group (p = 0.797); the corresponding OS rates were 100%, 100%, 84.8%, and 72.7%, respectively in the combined RFA and TACE group and 100%, 88.8%, 84.5%, and 74.0%, respectively in the RFA-alone group (p = 0.515), and the corresponding RFS rates were 71.3%, 59.9%, 48.8%, and 36.6%, respectively in the combined RFA and TACE group and 74.3%, 52.4%, 29.7%, and 29.7%, respectively, in the RFA-alone group (p = 0.365) [92]. These authors concluded that combined RFA plus TACE and RFA alone had equivalent effectiveness in the treatment of small HCCs (≤3 cm) [92]. Thus, RCTs that have evaluated the efficacy of RFA plus TACE have reported that it can improve survival rates relative to RFA alone for patients with HCCs of 3–5 cm in diameter [91]. RCTs that have compared RFA plus TACE with RFA alone for HCC are listed in Table III.

Table III. Reports of randomised controlled trials of comparison between RFA plus TACE and RFA alone for hepatocellular carcinoma.

								OS rate (%)
Author, country	Year	Treatment	No. of patients	Tumour size (cm)	Tumour number	Initial CR (%)	Follow-up (months)	1 year	3 years	5 years
Cheng et al. [88], China	2008	RFA	100	3–7.5	≤3 nodules	37	24.6 (mean)	67*	32*	8*
		TACE	95			5	25.4 (mean)	74*	32*	13*
		RFA + TACE	96			55	35.8 (mean)	83	55	31
Shibata et al. [92], Japan	2009	RFA	43	0.8–3	≤2 nodules	100	30.4 (mean)	100	84.5	NA
		RFA + TACE	46			100	100	84.8	NA
Morimoto et al. [89], Japan	2010	RFA	18	3.7 (mean)	Single	100	32 (mean)	89	80	NA
		RFA + TACE	19	3.6 (mean)		100	30 (mean)	100	93	NA
Peng et al. [90], abstract only, China	2012	RFA	95	<7	NA	NA	7 to 62 months	85.3	59*	NA
		RFA + TACE	94			NA		92.6	66.6	NA

RFA, radiofrequency ablation; TACE, transcatheter arterial chemoembolization; CR, complete response; OS, overall survival; NA, not available. *Statistically significant.

RFA versus RFA plus PEI

To our knowledge, there has been only one RCT that has compared RFA plus PEI with RFA alone [95]. The authors demonstrated that 1-, 2-, 3-, 4-, and 5-year OS rates were 95.4%, 89.2%, 75.8%, 63.3%, and 49.3%, respectively using RFA-PEI, and 89.6%, 68.7%, 58.4%, 50.3%, and 35.9%, respectively using RFA alone (p = 0.04) [95]. They also reported that the OS rate for the RFA-PEI group with 3.1–5.0 cm tumours was significantly higher than that for the RFA-alone group with 3.1–5.0 cm tumours (p = 0.03), but that the OS was not significantly different between the two modalities in patients with tumours that were ≤3.0 cm (p = 0.44) or 5.1–7.0 cm (p = 0.70) in diameter [95]. In addition, Kurokohchi et al. reported that RFA plus PEI might be effective for the treatment of HCCs located in regions that were difficult to treat with RFA alone [96]. However, because of the limited published data available, further larger prospective studies will be needed to clarify if RFA combined with PEI can be more effective than RFA alone for local ablative HCC therapy.

Limitations of RFA therapy for HCC

There are several limitations associated with RFA therapy for HCC despite its many potential favourable effects [18,97,98]. These include a limited ablation volume, technical limitations, expected complications dependent on tumour location, the heat-sink effect, intravascular tumour spreading by intratumoral pressure during RFA, and tumour seeding [18,97,98,99].

In general, the ablation zone that can be achieved using the currently available RFA device is limited to 4–5 cm in maximum diameter [97]. For obtaining a larger ablative margin, the employment of multiple electrodes or overlapping ablations with a single electrode is useful. TACE combined with RFA is also promising, as described above [88–90]. Conversely, the treatment for HCC tumours in the hepatic hilar location or adjacent to the gall bladder using RFA increases the risk of incomplete ablation and serious complications [98]. Cooling of bile ducts using an endoscopic nasobiliary drainage tube can prevent biliary complications induced by RFA of HCC located in the hepatic hilar site close to major bile ducts [100]. The presence of large vessels in proximity to the targeted tumours also has an adverse effect on thermal ablation, which is called the ‘heat sink effect’ [63]. The Pringle manoeuvre is a well-known method for minimising the heat sink effect when treating the targeted HCC abutting the major vessel, although open laparotomy is needed for this method [101]. To minimise intravascular tumour spreading by intratumoral pressure during RFA, the multi-step ablative method by incremental, stepwise expansion of the needle may be effective [102]. Moreover, neoplastic tumour seeding is well known to be one of the complications of the RFA procedure, although this complication is extremely rare in Japan [55,98,103]. This may be because when the RFA electrode is removed, the needle tract is sufficiently coagulated, unlike when performing tumour biopsy.

Advanced techniques of RFA

Artificial ascites method

Although various methods of RFA treatment can be performed, from the percutaneous approach to open surgery, most operators prefer the percutaneous approach because it is easy to perform and cost effective [18,42,98]; however, this approach is not always possible to achieve and depends on the characteristics of the tumour. When the tumour is located in the right subphrenic region, B-mode ultrasound (US)-guided percutaneous RFA alone is technically challenging because of partial tumour visibility and a poor electrode path due to overlapping lungs or ribs [104,105]. Minami et al. demonstrated the safety and feasibility of artificial intrathoracic pleural effusion in percutaneous RFA using a 5% glucose solution to separate the lung and the liver [104]. These authors reported that by using this method, complete tumour necrosis was achieved in a single session of RFA in 27 (96.4%) out of the 28 lesions [104]. In addition, Minami et al. reported the usefulness of percutaneous RFA guided by contrast-enhanced harmonic sonography, with concurrent use of artificial pleural effusion for LTP of HCC located in the right subphrenic region [105].

Conversely, Kondo et al. reported on the usefulness of the artificial ascites technique for percutaneous RFA of liver tumour adjacent to the gastrointestinal tract in avoiding RFA-induced thermal injury [106]. To avoid burns, artificial ascites were useful for creating a space between the surface of the liver and the skin, or gastrointestinal tract or diaphragm [106]. The artificial ascites technique has advantages over artificial pleural effusion because it can avoid interference with the lungs and gastrointestinal tract at the same time [106,107]. However, in patients with a past history of abdominal surgery, which can cause intra-abdominal adhesions, the use of this technique may be limited; in such patients artificial pleural effusion may be more effective [107].

Fusion imaging technique

The fusion imaging technique, which is referred to as real-time virtual sonography (RVS), a virtual CT sonography system with magnetic navigation (Hitachi Medico, Tokyo, Japan), can synchronise images obtained using B-mode ultrasonography with those obtained using multiplanar reconstruction (MPR) CT on the same screen in real time [108]. This system was composed of the main unit, the body of the magnetic location detector unit, the magnetic field generator, and the magnetic sensor attached to the sonographic transducer; it can show any cross-sectional image of CT volume data in real time corresponding to the angle of the transducer in the magnetic field [108,109]. This imaging technique is a useful procedure for detecting unclear nodules on B-mode ultrasonography [108,109].

Minami et al. have reported that the technical success rate after a single RFA session was significantly higher for patients treated with RVS-assisted RFA than for those treated with RFA using B-mode ultrasonography alone (p = 0.017); the number of treatment sessions was significantly lower for the RVS-assisted group (p = 0.021) [109].

Contrast-enhanced ultrasonography

Sonazoid® (Daiichi Sankyo, Tokyo, Japan) is a second-generation contrast agent used in ultrasonography. Levovist (Levovist®; Bayer Co., Ltd., Tokyo, Japan), a first-generation US contrast agent, improved the localisation of sonographically undetectable hypervascular tumours of the liver. However, because of the fragility of its microbubbles, images must be obtained intermittently, while the post-vascular phase must be obtained by a single sweep scan of the liver [110–112]. Conversely, Sonazoid produces stable non-linear oscillations in a low-power acoustic field and provides great detail regarding the second harmonic signals in real time [110–112]. Masuzaki et al. have reported that in 716 HCC nodules treated with RFA that were detected on dynamic CT, the detectability rate for tumour nodules was significantly higher using contrast-enhanced ultrasonography (CEUS) with Sonazoid than using conventional ultrasonography (p = 0.04); the number of RFA sessions using CEUS with Sonazoid was significantly lower relative to conventional ultrasonography (1.33 ± 0.45 for CEUS with Sonazoind versus 1.49 ± 0.76 for conventional ultrasonography alone, p = 0.0019) [110]. In our department we have also used this method to treat many HCC nodules using RFA, and in our experience, tumours located on the surface of the liver can be identified most clearly using CEUS with Sonazoid. Although in clinical practice there are no specific indications of fusion image guidance using RVS versus CEUS with Sonazoid, proper use or combination use of these two methods depending on tumour location and tumour visibility on ultrasonography is essential.

Future perspective of RFA and new emerging ablative techniques

The recent addition of molecular-targeted agents to the treatment armamentarium for HCC has prompted the design of clinical trials with the purpose of investigating the synergies between locoregional therapies and systemic chemotherapy. Combining RFA and targeted systemic therapy including sorafenib may be a novel approach for optimising clinical outcomes in HCC patients [35]. Xu et al. demonstrated that relative to the RFA-alone group, hypoxia inducible factor-1α and vascular endothelial growth factor A expression, which were associated with promoting rapid progression of the residual tumour after RFA through hypoxia, were significantly decreased in the group that received RFA combined with sorafenib therapy (p < 0.05); they concluded that sorafenib was able to increase the time to recurrence when used as an adjunct to RFA [113]. Currently, a large global study (the Sorafenib as Adjuvant Treatment in the Prevention of Recurrence of Hepatocellular Carcinoma (STORM) trial) is underway [35]. If favourable results are obtained in this trial, the treatment strategy for HCC will be drastically changed.

Currently, several new promising ablative techniques have emerged as shown in Table IV. Lyso-thermosensitive liposomal doxorubicin (LTLD) consists of heat-enhanced cytotoxic doxorubicin within a heat-activated liposome [114]. LTLD is infused intravenously prior to performing RFA. When the ablated area is heated to >39.5 °C, LTLD releases doxorubicin in high concentrations into the targeted HCC and its margins, leading to a larger ablative margin [114]. Irreversible electroporation (IRE) is a non-thermal ablative technique [115,116]. IRE is a method for the induction of irreversible disruption of cell membrane integrity resulting in cell death without the need for additional pharmacological injury [115,116]. Because IRE is a non-thermal technique, issues related to perfusion-mediated tissue cooling or heating are not relevant [115,116]. High intensity focused ultrasound (HIFU) ablation is an extracorporeal non-invasive ablation mode using focused ultrasound energy, which is capable of causing coagulative necrosis of the targeted HCC via intact skin without the need for surgical incision or insertion of instruments [117,118]. HIFU can provide a potential therapeutic method for the precise ablation of entire liver tumours without damaging vital structures [117,118]. Other new promising ablative therapies for HCC include oncolytic virus therapy using the oncolytic effect of the virus itself [119], photodynamic therapy [120], robot-assisted ablation therapy [121], microwave ablation therapy [122], cryoablation therapy [123] and laser ablation therapy [124].

Table IV. New emerging ablative techniques for hepatocellular carcinoma.

1.	Radiofrequency thermal ablation using lyso-thermosensitive liposomal doxorubicin (LTLD).	[114]
2.	Irreversible electroporation (IRE).	[115,116]
3.	High-intensity focused ultrasound ablation (HIFU).	[117,118]
4.	Oncolytic virus therapy.	[119]
5.	Photodynamic therapy.	[120]
6.	Robot-assisted ablation therapy.	[121]
7.	Microwave ablation therapy.	[122]
8.	Cryoablation therapy.	[123]
9.	Laser ablation therapy.	[124]

Conclusions

RFA has become the standard locoregional ablative therapy for small HCC because of its more favourable survival and local tumour control characteristics relative to other locoregional therapies. RFA-related complications can be reduced by practitioners gaining sufficient experience in the application of this therapy. Assessment of the treatment efficacy of RFA is essential because it is associated with clinical outcomes after RFA. Whether or not RFA therapy is as effective as SR for the treatment of HCCs remains controversial. Combination therapy involving RFA and PEI, or TACE, can be performed in large tumours for the enhancement of anticancer effects. Technical advancements in RFA have improved treatment outcomes in HCC patients treated with this modality. Adjuvant therapies, such as molecular-targeted therapies following RFA, may improve prognosis after RFA. Several new promising ablative techniques have emerged; however, further studies are needed to confirm the treatment efficacy of these therapies.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.