Early diffuse recurrence of hepatocellular carcinoma after percutaneous radiofrequency ablation: analysis of risk factors
To evaluate the risk factors affecting early diffuse recurrence within 1 year of percutaneous ultrasound-guided radiofrequency ablation (RFA) for hepatocellular carcinoma (HCC).
Out of 146 patients who received transcatheter arterial chemoembolisation (TACE) for treatment of recurrent HCC after percutaneous ultrasound-guided RFA, we selected 23 patients with early diffuse recurrence. Early diffuse recurrence was defined as three or more new recurrent HCCs within 1 year of initial RFA. As a control group, we selected 23 patients, matched exactly for age and sex, in which there was no local tumour progression or new recurrence after RFA. To analyse the risk factors, we examined patient factors and tumour factors.
Recurrent tumours occurred from 30 to 365 days after RFA (median time, 203 days). Univariate analysis indicated that larger tumour size and poorly defined margin were significant risk factors (P < 0.05). Multivariate analysis indicated that poorly defined margin was a significant risk factor (P < 0.05).
Larger tumour size and poorly defined margin may be risk factors for early diffuse recurrence of HCC within 1 year of RFA. Tumours with such risk factors should be treated with a combination of TACE to minimise the potential for therapeutic failure.
• Ultrasound-guided radiofrequency ablation (RFA) is widely used for hepatocellular carcinoma (HCC).
• Early diffuse recurrence after RFA is an important prognostic factor
• The risk factors for recurrence are larger tumour size and poorly defined margins
• Tumours with such risk factors should be treated with transarterial chemoembolisation.
KeywordsHepatocellular carcinomas Radiofrequency catheter ablation Percutaneous catheter ablation Recurrence Risk factors
Hepatocellular carcinoma (HCC) is the third most common cause of cancer death worldwide, accounting for about 315,000 deaths annually . The optimal curative therapy for patients with HCC is surgical resection or liver transplantation. However, these therapies are often not feasible. There are many limiting factors for successful surgical resection in patients with HCC such as severe impairment of hepatic functional reserve, bilobar distribution of the tumours, extrahepatic metastasis or involvement of the portal vein. Only 9–27% of the patients with HCC are eligible for surgical resection [2, 3]. A shortage of donors limits the possibility of hepatic transplantation. Thus, various local ablation therapies, such as ethanol ablation, microwave ablation and radiofrequency ablation (RFA), have been developed as an alternative to resection  or liver transplantation.
RFA is currently recognised as a minimally invasive and highly curative local treatment for HCC with a diameter of 3 cm or less , and has proven to be superior to ethanol ablation with respect to local tumour control without seriously affecting hepatic reserve [6, 7, 8]. Moreover, Chen et al.  recently showed that the survival rates of HCC patients after RFA were similar to those of patients who underwent surgical resection for small, solitary HCC in a randomised control study.
Nonetheless, even though these ablation therapies can achieve complete necrosis of small HCC, recurrence is still common. Early recurrence, defined as recurrence within a year of liver resection, is one of the most important factors affecting the prognosis and survival of patients with HCC. Regimbeau et al.  reported that the 1-year mortality rate after liver resection for HCC was 22% (123 out of 547 patients). Of these, 53 patients (43%) died of recurrence during the first postoperative year. In addition, a recent study by Park et al.  showed that the 3-year survival rate was 32% in the early recurrence group and 87% in the non-recurrence group, implying that early recurrence is closely related to postoperative survival rates. Among these diverse patterns of early recurrence after surgical resection or liver transplantation, diffuse intra-hepatic recurrence and multiple systemic recurrences are associated with short remaining survival times [12, 13].
Similarly, cases showing early and diffuse recurrence of HCC after RFA have recently been reported. Kotoh et al.  reported that the incidence of rapid and diffuse recurrent tumours after RFA is also associated with poor prognosis. Within the observation period, the mortality rates of patients with early and diffuse recurrences (within 7 months of RFA) were higher than those of patients without diffuse recurrences, reflecting worse prognoses for patients with early diffuse recurrences. Among the 127 patients without early diffuse recurrences, 40 patients died within the observation period: eight owing to liver failure, two from variceal rupture and the remainder owing to the progression of HCC. On the other hand, eight out of 11 patients with early diffuse recurrences died owing to advanced HCC. Thus, risk factors for the early diffuse recurrence after surgical resection or RFA may be reliable prognostic factors for long-term survival. Therefore, understanding the patterns and risk factors for early diffuse recurrence of HCC after RFA will help clinicians to (1) correct factors that are correctable, (2) tailor post-ablational management and (3) identify patients who may benefit from adjuvant therapy for the improvement of overall survival.
Early diffuse recurrence is rare but, if it occurs, prognosis is often dismal. Although there are several case reports describing the early diffuse recurrence of small HCC after RFA [15, 16], to the best of our knowledge, analyses of risk factors for the early diffuse recurrence after RFA for HCC have not been performed . Therefore, the purpose of this study was to evaluate the patterns and risk factors for early diffuse recurrence of HCC after RFA.
Materials and methods
From April 1999 to October 2009, we treated 1220 patients with HCC using RFA as the first-line treatment. Inclusion criteria for performing RFA in patients with HCC were as follows: the tumour or tumours should be visualised with ultrasound; a single tumour no greater than 5 cm in the largest dimension; multiple tumours (≤ 3) with each tumour measuring no greater than 3 cm; no portal venous thrombosis and extra-hepatic metastasis; prothrombin time ratio over 50% (prothrombin time with international normalised ratio [INR] < 1.7) and platelet count greater than 50,000/μl without transfusion support. We performed RFA under ultrasound guidance with local and conscious sedation.
A total of 32 HCCs in 23 patients (17 men, 6 women; mean age, 65 years) were evaluated (Group A). The diagnosis of HCC was made based on AASLD guidelines . Five tumours were diagnosed by percutaneous ultrasound-guided biopsy for initial diagnosis before RFA. The remaining 27 tumours were diagnosed by imaging criteria. The maximum diameter of the index tumours ranged from 1.2 cm to 4.8 cm (mean, 2.5 cm). The aetiology of hepatitis or liver cirrhosis was viral in all 23 patients (HBV, n = 19; HCV, n = 4). Child–Pugh classifications were A in 19 and B in four patients.
As a control group (Group B), we selected 23 out of 1,074 patients, matched to the cases for age and sex, in whom there was no local tumour progression or new recurrence after the initial RFA during the same study period.
RFA procedure and follow-up
All RFA procedures were performed percutaneously under ultrasound guidance (HDI 5000, Philips Healthcare, Best, The Netherlands; Acouson Sequoia 512, Siemens Medical Solutions, Mountain View, CA, USA; LOGIQ E9, GE Healthcare, Milwaukee, WI, USA). Procedures were performed on an inpatient basis by one of six radiologists (Y.S.K., H.K.L., H.R., M.W.L., D.C. and W.J.L.), each of whom had had at least 7 years’ experience performing this procedure by the end of the study period. We used either internally cooled, multi-tined expandable or perfusion electrode systems from one of six vendors (Cool-tip RF System, Covidien, Mansfield, MA, USA; VIVA RFA System, STARmed, Ilsan, Korea; Starburst RFA System, AngioDynamics, Latham, NY, USA; LeVeen Needle Electrode and RF 3000 Generator, Boston Scientific, Natick, MA, USA; Elektrotom 106 HiTT, Berchtold Corporation, Tuttlingen, Germany; Jet-Tip RF electrode, RF Medical, Seoul, Korea) according to temporal availability or operator preference. When we used internally cooled electrodes, we started from 50 W and continuously increased the power during the initial 2 min to minimise the popping phenomenon. We did not apply stepwise deployment when we used multi-tined expandable electrodes.
All patients were treated with 2% lidocaine hydrochloride at the puncture site and intravenous drip infusion of 50 mg pethidine hydrochloride mixed with 50 ml of 5% dextrose water. Patient cardiovascular and respiratory systems were continuously monitored during the procedures. We treated tumours with complete necrosis by ablating at least 0.5 cm of the normal hepatic parenchyma surrounding the tumour.
All patients underwent serial monitoring of alpha-fetoprotein, chest X-ray, and helical dynamic triple phase CT at 3- to 4-month intervals for the detection of local tumour progression, new intra-hepatic recurrence, and extrahepatic metastasis. Baseline and post-treatment imaging evaluations of HCC were performed using CT with intravenous contrast medium administration. CT was performed with one of two helical systems (Somatom Plus and Somatom Plus 4; Siemens, Erlangen, Germany). A total of 120 ml non-ionic contrast material (Iopromide, Ultravist 300; Schering-Korea, Ansung, Korea) was administered intravenously via the antecubital vein at a rate of 3 ml/s. A triple-phase contrast enhancement CT technique was used with delay times of 30, 60 and 180 s after the initiation of contrast medium injection. Helical CT unenhanced and contrast-enhanced images were acquired using a 5– to 8-mm collimation and 5– to 8-mm/s table speed.
Diagnoses of local tumour progression or new intra-hepatic recurrence were based on CT and/or MRI. Needle biopsies of recurrent tumours were not performed. Recurrent HCC was defined as a new enhancing mass in the arterial phase with washout in the portal or the delayed phase of dynamic liver CT. If the mass showed arterial enhancement without washout, we used the interval growth of the mass on follow-up imaging or lipiodol retention of the tumour after TACE to confirm recurrent HCC. Local tumour progression was defined as tumour recurrence within or at the periphery of the original ablative zone on subsequent CTs. New intra-hepatic recurrence was defined as any new tumour that occurred in the liver separate from the RFA zone. In this study, early diffuse recurrence was defined as three or more new intra-hepatic recurrent tumours within 1 year after initial RFA, regardless of local tumour progression.
Analysis of risk factors for early diffuse recurrence
We compared 11 clinical variables that may be related to the risk of early diffuse recurrence after RFA. These included four patient factors (pre-RFA alpha-fetoprotein [AFP], Child–Pugh classification, hepatitis B surface antigen status, hepatitis C antibody status) and seven tumour factors (size, location, margin, contact with portal vein, hepatic hilum, hepatic capsule and presence of an ablative margin).
For the analysis of patient-related risk factors for early diffuse recurrence, we compared (1) level of baseline serum AFP level at the time of RFA procedure, (2) degree of underlying hepatic reserve (categorised as Child–Pugh class A, B, and C), and (3) presence or absence of hepatitis B surface antigen (HBs Ag) or hepatitis C antibody at the time of RFA between the two groups.
For the analysis of tumour-related risk factors for early diffuse recurrence, we examined (1) maximum diameter of HCCs on ultrasonography, (2) margin of the HCCs after dividing the tumours into poorly marginated and well marginated groups according to the degree of border definition (well or poorly marginated, i.e. greater than or less than 50% of the border sharply defined on CT), (3) segmental location of HCCs according to Couinaud’s classification, (4) whether the tumour had contact with the portal vein, more than 3 mm in diameter at the contact point, (5) whether the tumour was in close proximity to the hepatic hilum (centre of the tumour was located within 1 cm of the hepatic hilum), (6) whether the tumour was located within 5 mm from the liver capsule, and (7) whether the ablative margin, defined as ablated normal surrounding parenchyma more than 5 mm in thickness, compared with the location of tumour before RFA, was established in an immediate three-phase CT scan. These risk factors were assessed by the consensus of two radiologists.
Risk factors for early diffuse recurrence were evaluated with univariate analyses using Cox regression tests. If multiple risk factors were shown to be significant by this test, we performed multivariate analysis using Cox regression tests. All reported P values were two-tailed, and P levels < 0.05 were considered statistically significant. All statistical analyses were performed using SPSS for Windows v. 18 (SPSS, Chicago, IL).
Risk factors related to early diffuse recurrence of HCC: univariate analysis (n = 46)
Pre-RFA AFP (ng/ml)
B or C
Cause of cirrhosis
Alcohol or idiopathic
Maximal diameter (cm)
Contact of the tumour with the portal vein
Contact of the tumour with the hepatic hilum
Contact of the tumour with the hepatic capsule
Ablative margin at immediate follow-up CT
Independent risk factors related to early diffuse recurrence of HCC: multivariate analysis
The mean values of AFP were 306.8 ng/ml and 263.1 ng/ml for patients with early diffuse recurrence and for patients without recurrence, respectively. There were no significant differences in pre-RFA AFP between Group A and Group B.
Child–Pugh status and cause of cirrhosis
Thirty-six patients had cirrhosis caused by hepatitis B infection, and eight patients had cirrhosis due to hepatitis C. All patients from Group A had cirrhosis caused by viral infection. Two patients from Group B had different causes of cirrhosis: chronic alcohol abuse (n = 1) and idiopathic cirrhosis (n = 1). The causes of cirrhosis and Child–Pugh status were not related to early diffuse recurrence.
Maximal diameter of the tumour
The mean diameter of HCCs was 2.5 cm in Group A and 1.9 cm in Group B. Small tumours had a significantly lower rate of early diffuse recurrence compared with the larger tumours (P = 0.036). However, the difference was not sustained in multivariate analysis (P = 0.666).
Eight tumours in Group A (25.0%) and 0 (0%) in Group B had poorly defined margins. The degree of border definition was significantly related to early diffuse recurrence between two groups (P = 0.001). The poorly defined margin was the only significant risk factor for early diffuse recurrence in multivariate analysis (P = 0.019).
There were no differences between the two groups in terms of the anatomical location of HCC, contact of the tumour with the portal vein, hepatic hilum and hepatic capsule, or ablative margin at immediate follow-up CT.
We demonstrated that larger tumours and poorly defined tumour margins are risk factors for early diffuse recurrence for HCCs after RFA. This result partly agrees with the results of previous studies indicating that large tumour size , absence of tumour capsule, indistinct margins , elevated AFP serum levels  and microscopic vascular invasion  are risk factors for early tumour recurrence after curative surgical resection. Causes may include the presence of a micro-metastatic lesion before surgery or tumour cell metastasis caused by compression or movement of the liver or tumour rupture during hepatectomy [12, 13].
Reported cases of intra-hepatic rapid progression after RFA
Tumour size (cm)
Seki et al.
RFA + TACE
Takada et al.
Takada et al.
Nicoli et al.
Ruzzenente et al.
Ruzzenente et al.
Ruzzenente et al.
Ruzzenente et al.
Izai et al.
Masuda et al.
Masuda et al.
Another hypothesis that may explain early diffuse recurrence after RFA is that rapid heating of a tumour may lead to an unpredicted increase in internal pressure and cause the dislodging and scattering of malignant cells around the ablated tumour [24, 25]. Kotoh et al.  reported that 11 out of 138 patients who were treated with RFA suffered from rapid and scattered recurrences within 7 months of ablation. After evaluating a series of HCC cases treated by RFA with different devices and protocols and switching from the original full-expansion method to the modified stepwise deployment of internal prongs, no scattered recurrences were observed. Therefore, to reduce intra-tumoural pressure during RFA, stepwise deployment of internal prongs may be warranted when using a multi-tined electrode .
We demonstrated that size and poorly defined HCC margins are significant risk factors for early diffuse recurrence of HCC after RFA. Ruzzenente et al.  described four patients (4.5%) with recurrences out of 87 consecutive cirrhotic patients with 104 HCCs and suggested that the following risk factors should be considered: high AFP level (>200 kU/l), location of the tumour near the portal vein branches (< 1 cm from the main or segmental portal branches) and poor differentiation. However, in the current study, only a few tumours (5/32, 15.6%) were confirmed histologically, so risk factors based on the results of surgical pathological features could not be analysed. In addition, Ruzzenente et al.  included only four patients in their sample, whereas we included 23 patients with early diffuse recurrence. Differences in sample size may explain the identification of different risk factors affecting rapid tumour progression.
In a previous study by Yu et al. , tumour size, poor pathological differentiation of tumour cells and advanced tumour staging were shown to be risk factors for early local tumour recurrence within 1 year of RFA. However, Yu et al.  defined early recurrence as local tumour recurrence within 1 year of initial RFA, and included only three patients with early local tumour recurrence and eight patients with early intra-hepatic metastasis after RFA of single small HCC (tumour size: 3.2 ± 0.7 cm in the early local recurrence group, 2.4 ± 0.8 cm in the early intra-hepatic metastasis group) in their sample. In addition, they did not clarify the patterns of early recurrence (number of tumours, presence of extrahepatic metastasis). Based on accumulative mortality rates (2/27 patients, 7.4%), we hypothesised that the patterns of early recurrence were probably not diffuse as in our cases. Therefore, it is not reasonable to compare the series by Yu et al. with our series.
Large tumour size , absence of tumour capsule, indistinct margins , elevated serum levels of AFP  and microscopic vascular invasion  have been reported to increase the risk of early tumour recurrence after curative resection. Except for AFP levels, our results are in line with those of previous studies of the early diffuse recurrence after curative liver resection. Previous studies have indicated that increased levels of serum AFP are associated with intra-hepatic recurrence after RFA for HCC. In our study, the relationship between serum AFP and intra-hepatic recurrence was not significant (P = 0.645).
To our knowledge, this is the first study to analyse risk factors of early diffuse recurrence of HCC after RFA as a first-line treatment. Moreover, this is the largest series to date comparing patients with early diffuse recurrence with those without any recurrence. However, the present study has several limitations. First, this study was retrospective and was not a randomised or controlled trial. Second, only a small number of the tumours in this study (5/32, 15.6%) were histologically confirmed. Finally, data regarding the duration and number of sessions of RFA were unavailable.
In addition, during the study period, we have used a few different types of electrodes. More than 90% of cases were performed with Cool-tip RF system (Covidien, Boulder, CO, USA) in both groups. Because of small number of other electrodes used and missing data regarding the type of electrode used in several cases, it was impossible to perform any statistical analysis to clarify the relationship between the electrode type and early diffuse recurrence. The lack of the record on the RFA electrode, RFA technique and duration of ablation made it impossible to compare the rates of early diffuse recurrence between the two groups using different RFA technique and type of electrodes.
In conclusion, our results indicate that poorly defined margin was the only independent factor predicting early diffuse recurrence of HCC. Therefore, patients at risk of early diffuse recurrence should be treated with a combination of TACE [28, 29] to minimise the potential for therapeutic failure and should be closely monitored during the first year after treatment.
- 2.(1980) Primary liver cancers in Japan. Cancer 45:2663–2669Google Scholar
- 15.Seki T, Tamai T, Ikeda K, Imamura M, Nishimura A, Yamashiki N, Nakagawa T, Inoue K (2001) Rapid progression of hepatocellular carcinoma after transcatheter arterial chemoembolization and percutaneous radiofrequency ablation in the primary tumour region. Eur J Gastroenterol Hepatol 13:291–294PubMedCrossRefGoogle Scholar
- 24.Kotoh K, Nakamuta M, Morizono S, Kohjima M, Arimura E, Fukushima M, Enjoji M, Sakai H, Nawata H (2005) A multi-step, incremental expansion method for radio frequency ablation: optimization of the procedure to prevent increases in intra-tumor pressure and to reduce the ablation time. Liver Int 25:542–547PubMedCrossRefGoogle Scholar
- 25.Mori Y, Tamai H, Shingaki N, Moribata K, Shiraki T, Deguchi H, Ueda K, Enomoto S, Magari H, Inoue I, Maekita T, Iguchi M, Yanaoka K, Oka M, Ichinose M (2009) Diffuse intrahepatic recurrence after percutaneous radiofrequency ablation for solitary and small hepatocellular carcinoma. Hepatol Int 3:509-515Google Scholar