Radiofrequency ablation of hepatic tumors: Lessons learned from 3000 procedures


Professor Hyo K Lim, Department of Radiology, Samsung Medical Center, Sungkyunkwan University, School of Medicine, 50 Irwon-dong, Kangnam-ku, Seoul 135-710, Korea. Email:


Radiofrequency ablation has been accepted as the most popular local ablative therapy for unresectable malignant hepatic tumors. For 9 years from April 1999, we performed 3000 radiofrequency ablation procedures for hepatic tumors in our institution. Our results on the safety (mortality, 0.15%/patient) and therapeutic efficacy (5-year survival rate, 58%) are similar to those of previous studies reported, supporting the growing evidence of a clear survival benefit, excellent results for local tumor control and improved quality of life. The most important lesson learned from our 3000 procedures is that the best planning, safe ablation and complete ablation are key factors for patient outcome. Furthermore, multimodality treatment is the best strategy for recurrent hepatocellular carcinoma encountered after any kind of first-line treatment.


Hepatocellular carcinoma (HCC) is one of the most common malignancies in the world and the third most common frequent cause of death of cancer.1–3 Liver transplantation is the best curative option with good survival rates, although its usage is restricted by the shortage of donor organs. Surgical resection may also offer a chance for a cure. However, the tumors in most patients are unresectable because of multifocal disease, poor hepatic reserve or inability to obtain an optimal tumor free margin.4 Therefore, loco-regional treatment is usually the only option for the majority of patients with HCC. Fortunately, there are growing evidences on clinical benefits of loco-regional treatment options. In addition, the fact that HCC has a tendency to remain confined to the liver, until the disease has advanced, makes loco-regional therapies particularly attractive.5,6

Image-guided treatment for patients with unresectable HCC includes catheter-based approaches (transarterial chemoembolization [TACE] or radioembolization) and ablative techniques, chemical (ethanol ablation) or thermal (radiofrequency, microwave, laser ablation). Among them, radiofrequency ablation (RFA) has been used as the most popular method for treating smaller HCC less than 3 cm in diameter. Many clinical studies have proven the safety and efficacy of these loco-regional procedures during the past two decades.6–10

We have performed 3000 RFA procedures at our institution during the past 9 years. Most of the data from our clinical outcomes have been published in more than 25 peer-reviewed articles on various topics.11–36 The purpose of this article is to summarize our 8 years of experience with RFA procedures compared with the results of previous studies, and present the most important lessons learned from the 3000 procedures for treating patients with hepatic tumors.

Overview of RFA procedures

From April 1999 to October 2007, we performed a total of 3000 RFA procedures for the treatment of 2066 patients with unresectable hepatic tumors (HCC, 1804; metastasis, 262). Almost all patients met the criteria for treatment with RFA adopted by our hospital: a single nodular tumor 5 cm or less in maximum diameter; multinodular (n = ≤ 3) tumors 3 cm or less in maximum diameter each; the absence of extrahepatic metastases; prothrombin time ratio of more than 50% and platelet count of more than 50 000 cells/mm3; and a Child–Pugh class A or B liver cirrhosis.

Most procedures were percutaneously (87%) performed under ultrasonography (Sequoia 512, Acuson; Siemens, Munich, Germany) guidance by six experienced radiologists (H. K. L., H. R, D. C, Y. S. K., W. J. L. and S. H. K.). The anesthesia in a percutaneous procedure was local with conscious sedation using lidocaine and pethidine. From July 2000, most (> 95%) procedures were performed using internally-cooled electrodes (Cool-tip RF system; Valleylab, Boulder, CO, USA). The ablation strategy was to include a peripheral margin of 0.5–1.0 cm of normal hepatic parenchyma surrounding the tumor, as well as the entire tumor. The electrode tract ablation technique was used to minimize post-procedural bleeding and tumor seeding along the electrode tract. The therapeutic response was assessed by four-phase dynamic contrast-enhanced computed tomography (CT) immediately and 1 month after the procedure, and then at 3-month intervals.

Safety data

The mortality rate of our series was 0.1% (3/3000 sessions) per session and 0.15% (3/2066 patients) per patient. Two patients died from hepatic infarction followed by hepatic failure (n = 1) and sepsis (n = 1) after intraoperative RFA combined with hepatic resection. One patient with percutaneous RFA died of cardiac tamponade after cardiopulmonary resuscitation for respiratory arrest due to severe side-effects of anesthetics. In our study of 570 patients with 674 HCC, 11 major complications (1.9%/treatment) were found during the follow-up period.27 We experienced four hepatic infarctions, three hepatic abscesses, two tumor tract seeding, one biloma, and one acute cholecystitis in this subset of the patients. A comparison of the complications with previous large series studies is summarized in Table 1.18,26,37–41 Mortality reported ranged 0–0.3%. Major complication rates ranged 1.8–4.0%.

Table 1.  Comparison of complications after radiofrequency ablation of hepatocellular carcinoma in the published work
YearAuthorNo. patientsMortality and morbidity (%)
MortalityMajor complicationMinor complication
  • Multi-center study by Italian Group.

  • Multi-center study by Korean Group.

  • NA, not available.

2003Livraghi et al.23200.32.25
2005Rhim et al.11390.12.4NA
2005Lencioni et al.2060.02.0NA
2005Tateishi et al.3190.04.01.7
2006Chen et al.102660.02.4NA
2007Choi et al.5700.01.83.9

We retrospectively evaluated the frequency and risk factors of liver abscess formation after percutaneous 751 RFA procedures in 603 patients with HCC.24 The frequency of liver abscess was 1.7%/procedure. The risk factors identified were a preexisting biliary abnormality, a tumor with retention of iodized oil from a previous TACE, and treatment with an internally-cooled electrode system. All patients successfully recovered from their liver abscess with parenteral antibiotics and percutaneous drainage of purulent material.18

The frequency of hepatic infarction after our 1120 RFA sessions among 872 patients with HCC was 1.8%.34 The common presenting symptoms were abdominal pain and fever. All infarctions were found at the first follow-up CT examination. Gas collection was noted in 65% of patients. All lesions showed progressive shrinkage. One patient with a lobar infarction died from hepatic failure. An older age and larger tumor were identified as statistically significant risk factors.

Technical effectiveness and survival data

We evaluated the long-term survival results of percutaneous RFA in 570 patients with early-stage 674 HCC treated from April 1999 to May 2005.26 The primary technique effectiveness rate was 96.7%. The cumulative rates of local tumor progression at 1, 2 and 3 years were 8.1%, 10.9% and 11.8%, respectively. The cumulative survival rates at 1, 2, 3, 4 and 5 years were 95.2%, 82.9%, 69.5%, 60.8% and 58%, respectively. Child–Pugh class A cirrhosis, in younger patients (< 58 years), or those with a lower α-fetoprotein (AFP) level (< 100 mg/L) demonstrated better survival results. To date, five clinical studies have reported on the long-term results of the study population in more than 100 patients with HCC treated by percutaneous RFA as the first-line treatment. Comparison with previous studies is summarized in Table 2.37–40 The 1-, 3- and 5-year survival rates of the reviewed studies ranged 94.7–97%, 71–77.7% and 48–68.5%, respectively, which were similar to our study results. The most recent study for early HCC (< 2 cm) reported by Livraghi et al. showed the best survival rates. Based on their results, RFA can be considered the treatment of choice for patients with a single HCC, even when surgical resection is possible.39

Table 2.  Comparison of overall survival rates after radiofrequency ablation of hepatocellular carcinoma in the published work
YearAuthorNo. patientsSize of tumorsOverall survival rates (%)
1 year3 year5 year
  1. NA, not available.

2005Lencioni et al.206< 5 cm976741
2005Tateishi et al.319< 5 cm957854
2006Chen et al.10266< 6.7 cm835843
2007Choi et al.570< 5 cm957058
2007Livraghi et al.216< 2 cmNANA68

We also evaluated the long-term survival results of combined hepatectomy and intraoperative RFA in 53 patients with 66 HCC from April 1999 to November 2006.27 The primary technique effectiveness rate was 98.0%. The cumulative survival rates at 1, 2, 3, 4 and 5 years were 87%, 83%, 80%, 68% and 55%, respectively. Patients with smaller resected tumors (< 5 cm) demonstrated better survival results. Wakabayashi et al. and Raut et al. reported that the cumulative survival rates at 1, 3 and 5 years were 58%, 27% and 22%, and 58%, 45% and 45%, respectively.42,43

We assessed the long-term survival results of percutaneous RFA in 102 patients with recurrent 111 HCC after hepatectomy from April 1999 to October 2005.26 The primary technique effectiveness rate was 93.3%. The cumulative rates of local tumor progression at 1, 3 and 5 years were 6.0%, 8.6% and 11.9%, respectively. The cumulative survival rates at 1, 2, 3, 4 and 5 years were 93.9%, 83.7%, 65.7%, 56.6% and 51.6%, respectively. Patients with a lower AFP level (< 100 mg/L) before the RFA or with a small resected tumor (< 5 cm) demonstrated better survival results. There were several reports on RFA for intrahepatic recurrent HCC after hepatectomy. However, information on the long-term results with 5-year survival rates are limited.44,45

We compared the efficacy of RFA (n = 55) and surgical resection (n = 93) in a group of patients with a Child–Pugh score of 5 and a single HCC less than 4 cm in diameter.25 The local tumor progression rate in the RFA group was significantly higher than in the surgery group, while the incidence of new HCC was similar in the comparison between the two groups. The cumulative and recurrence-free survival rates at 1 and 3 years were not significantly different in comparisons between the two groups. Our results are similar to two recent randomized trials.46,47 The first study included 180 patients with single HCC nodules smaller than 5.0 cm; the 4-year survival rates after surgery and RFA were 67% and 64%, respectively. In the second study, 105 patients with up to three nodules smaller than 3 cm were treated; the 3-year survival rates were 86% in the RFA group and 87% for the surgical resection group.

Lessons learned from 3000 RFA procedures

We learned many lessons from the 3000 RFA procedures for hepatic tumors during the past 9 years. The most important lessons, with consensus of the six operators at our institution, were that successful ablations were based on best planning, safe ablation and complete ablation. In addition, RFA should be considered as one treatment modality among multimodality treatments for management of unresectable hepatic tumors.

Lesson 1: best planning

Best planning is the first step for a successful outcome after RFA of hepatic tumors. Planning includes the following: (i) to assess the feasibility of procedure based on inclusion/exclusion criteria (proper patient selection); (ii) decision of type of approach (i.e. percutaneously, laparoscopically, open), electrodes, guiding modalities and number of ablations; and (iii) to decide whether to apply overlapping ablations or a special technical tip (e.g. artificial ascites) for safe and complete ablation. The feasibility assessment for RFA begins with a review of good-quality abdominal CT or magnetic resonance imaging. These preoperative imaging studies are used to determine the number and size of tumors and their relationship to surrounding structures including blood vessels or vital organs.48,49 An operator should perform planning ultrasonography (US) if a US-guided procedure is considered. An operator should assess three-dimensional (3-D) configuration and size of the tumor, whether there is a safe electrode path to the tumor, whether there is any organ close to the expected RFA zones and whether there is any large vessel close to the tumor.

Currently, there is no consensus on the objective criteria for the feasibility of percutaneous RFA reported in the published work. At our institution, we use a new scoring system for the feasibility assessment (Table 3). The scoring system consists of five categories: (i) safe electrode path; (ii) vital organs adjacent to the RFA zone; (iii) tumor size; (iv) tumor conspicuity; and (v) the heat-sink effect. Each category is assessed on a four-point scale (i.e. 1–4), and a feasibility score is calculated by summation of the points in each category. If a score of 4 is obtained in any category, the patient is considered not to be a candidate for percutaneous RFA. A preliminary study revealed good inter-observer agreement with the new scoring system between two observers (H Rhim, unpubl. data, 2008).

Table 3.  New feasibility scoring system for percutaneous ultrasonography-guided radiofrequency ablation of hepatic tumors
ParameterPoints assigned
  • If any category was judged as ‘4’, the patient was determined to be unfeasible for percutaneous ultrasonography-guided radiofrequency (RF) ablation.

  • Feasibility score (= POSCH score) was calculated by summing each point in each category up. Thus, the feasibility score ranged 5–15.

RF electrode pathGoodFairPoorNone
Dangerous organNoneEquivocalProbableDefinite
Tumor sizeVery smallSmallIntermediateLarge
Tumor conspicuityGoodFairPoorInvisible
Heat sink effectNoneEquivocalProbableDefinite

The feasibility of any procedure has a broad spectrum depending on the difficulty of the procedure. The feasibility of RFA cannot be determined just as ‘absolutely feasible’ or ‘absolutely unfeasible’. A majority of referred patients will be in the middle of this spectrum. Thus, an operator should consider an alternative approach, alternative guiding modality or alternative treatment modalities (e.g. percutaneous ethanol injection, percutaneous acetic acid injection, TACE) depending on the degree of feasibility or the cause of technical difficulty. HCC generally requires more than a single treatment modality. Hence, we should select the proper treatment modality based on the status of the initial and recurrent tumors. Tailor-made ablation strategies with multimodality treatments may be the best approach for successful ablation outcomes.50,51

Lesson 2: safe ablation

Safe ablation is the second key to a successful loco-regional treatment. Minimal invasiveness is a clear advantage of image-guided ablation over the surgical treatment. Safe ablation can be supported by careful patient selection, close patient monitoring during the ablation and appropriate management of complication after the ablation.52–57 As careful patient selection cannot guarantee safe ablation, early detection during and immediately after ablation is more important for managing the major complications appropriately. To minimize the mortality resulting from the major complication, an interventional radiologist should be aware of the broad spectrum of complications encountered after RFA of hepatic tumors (e.g. bleeding, massive infarction, extensive abscess with sepsis and thermal injury of the colon).18,47,53,54 The frequency of major complications may be correlated with both the experience and aggressiveness of an operator. If an operator has more aggressive posture for achieving complete ablation with enough ablative margin, the rate of major complication will be increased even when treating technically feasible tumors.

We assessed the frequency and clinical significance of thermal injury to the gastrointestinal and biliary tracts in two studies.19,20 Although mild bowel edema or bile duct dilatation adjacent to the RFA zone was often seen at CT performed during or immediately after follow up, a major complication requiring specific treatment was very rare. However, we recommend alternative treatment, especially for tumors broadly abutting colonic loops or major (central) bile ducts. All technical tips to minimize the complication should be considered if there is any possibility of complication at planning phase. There are several technical tips to minimize collateral thermal injury to vital organs including: (i) a different approach;55 (ii) using artificial fluid or air;57–59 (iii) using a balloon catheter; and (iv) pulling back or lifting the electrode.10 Among them, artificial ascites has gained acceptance as a simple and effective technique for successful ablation.60,61

We recently assessed the feasibility, safety and efficacy of percutaneous RFA with artificial ascites for 143 patients with 180 HCC. The reasons for using artificial ascites were to prevent thermal injury of the adjacent organs (n = 146) and to improve the sonic window for hepatic dome lesions (n = 96). We artificially induced ascites before RFA by dripping 5% Dextrose in Water solution (D/W solution; Choonwae Pharma, Seoul, Korea). RFA with artificial ascites was successfully achieved in 130 (90.9%) of 143 patients with a mean of 1010 mL D/W solution. The reasons of technical failure were perihepatic adhesion due to surgery (n = 4), TACE (n = 1), prominent omental fat (n = 1), bare area (n = 2) or unknown (n = 5). Except for one patient with hemoperitoneum, there was no major complication related to artificial ascites. Artificial ascites is a simple and useful technical tip to minimize the collateral thermal injury, to reduce the periprocedural pain and to improve the sonic window (Fig. 1).57

Figure 1.

A 72-year-old man with hepatocellular carcinoma in the right hepatic dome (S8). (a) Diagram of artificial ascites to improve the sonic window and electrode path by displacing the liver downward. Transparent white triangle zone represent an acceptable sonic window, while gray triangular zone represents a poor sonic window due to overlapped lung or ribs. (b) Transverse contrast-enhanced computed tomography (CT) scan before radiofrequency ablation shows a 2-cm hyperattenuating nodule in liver segment VIII (arrow). (c) Intercostal ultrasonographic scan shows a hypoechoic mass after introducing artificial ascites (arrow). The index tumor was not seen on planning ultrasonography due to overlapped lungs. (d,e) Transverse and coronal contrast-enhanced CT scan obtained immediately after radiofrequency ablation (RFA) shows a non-enhancing radiofrequency ablation zone encompassing the tumor (arrow). There is no evidence of thermal damage of the diaphragm. (f) Transverse contrast-enhanced CT scan obtained 1 month after radiofrequency ablation shows complete ablation without local tumor progression. The artificial ascites is completely absorbed without any delayed complication.

Lesson 3: complete ablation

Complete ablation is the last key to a successful local treatment. The ablative margin surrounding the index tumor is an established prognostic factor for local tumor control. Local tumor progression can develop at the margin of an ablation zone if the ablative margin is inadequate compared to acceptable ablative margins.62–68 So that there is enough ablative margin, three factors should be considered prior to or during the ablation: (i) the tumor size with 3-D configuration; (ii) the configuration and size of the RFA zone made by a specific electrode; and (iii) the direction of the RF electrode path related to the tumor configuration. We believe that a complete ablation depends on achieving a symmetrical ablative margin rather than simply increasing the volume of the ablation zone (Fig. 2).

Figure 2.

A 76-year-old man with hepatocellular carcinoma in the right hepatic dome (S8). (a) An ablative margin (black zone) of 5–10 mm surrounding the index tumor (white zone) is the essential factor to achieve local tumor control. (b) Transverse contrast-enhanced computed tomography (CT) scan before radiofrequency ablation (RFA) shows a 2.5-cm hyperattenuating nodule in liver segment VIII (arrow). (c) Transverse contrast-enhanced CT scan obtained immediately after RFA shows a 3.5 cm non-enhancing RFA zone with sufficient symmetrical ablative margin (arrow).

Regarding the tumor configuration, 3-D measurement of the index tumor must be done accurately. An effective direction for the RF electrode path depending on the expected RFA zone made by a specific electrode, should be selected. Finally, targeting of the intended site of the tumors (the location determined during the planning phase) must be performed accurately because position changes of the electrode become difficult due to a poor sonic window caused by microbubbles. Hence, some experience in image-guided liver biopsy is mandatory for those involved in an RFA procedure. To solve this problem, multiple electrodes can be used, either at monopolar or bipolar modes, for achieving the desired ablative margin.69–71

Another inevitable obstacle in achieving a complete ablation is the heat-sink effect caused by abutting large vessels of more than 3 mm in diameter.70–72 There are many options to deal with this technical challenge. One can choose a more aggressive posture using an intraoperative Pringle maneuver, or combined therapy with an ethanol injection or TACE for a perivascular tumor.73,74 At our institution, we recommend TACE rather than RFA if more than 90° of the circumference of the central tumor abuts large vessels (> 3 mm in diameter), especially at the hepatic hilum, because the possibility of an incomplete ablation and biliary complications are relatively high.

We retrospectively assessed the morphological pattern and exact site of local tumor progression after RFA in 86 patients with 89 hepatic tumors (72 HCC; 17 metastases) who developed local tumor progression after RFA using internally-cooled electrodes. The most common morphological pattern was the peripheral nodular type. The exact site of the local tumor progression was concordant with an insufficient ablative margin in 85.0%, a contiguous vessel in 42.4% and a subcapsular location in 46.9% (H Rhim, unpubl. data, 2008).

To sustain complete ablation for index tumor, regular imaging follow up is essential. Most local tumor progression at the RFA zone are usually detected at the 3–4-month interval imaging studies, and can be treated by additional RFA. If the local tumor progression is very small (< 5 mm) or very subtle on planning ultrasonography, one can use a contrast-enhanced US-guided procedure.75


Our results on the safety and therapeutic efficacy are similar to those of previous studies, supporting the growing evidence of a clear survival benefit, excellent results for local tumor control and improved quality of life. The most important lesson was that best planning, safe ablation and complete ablation are important factors for patient outcome. We believe that one can achieve successful RFA in most cases if one keeps in mind the three basic lessons and the implication of multimodality treatment.

Many sophisticated studies are ongoing to enhance these basic lessons in the field of interventional oncology76–84 such as: refinement of RF devices; technical tips for safe ablation; fusion imaging for accurate targeting, better monitoring and precise assessment of the treatment response; and combined treatment with new chemotherapeutic agents or TACE (Fig. 3). Although HCC should not be approached with only a single treatment modality, image-guided ablation currently plays an important role in the management of patients with unresectable HCC.85,86

Figure 3.

The most important lessons learned from 3000 radiofrequency ablation (RFA) procedures. The most important keys for successful ablation are best planning, safe ablation and complete ablation. Many technical advances are ongoing to enhance these basic factors in the field of interventional oncology such as: refinement of radiofrequency devices; technical tips for safe ablation; fusion imaging for accurate targeting, better monitoring and precise assessment of the treatment response; and combined treatment with new chemotherapeutic agents. RCT, randomized clinical trial.