Hepatectomy is an established curative treatment for hepatocellular carcinoma (HCC), and now performed worldwide  . However, intrahepatic recurrence of HCC is frequently noted in patients after hepatectomy of HCC. The incidence of intrahepatic recurrence has been reported to be as high as 70% at 5 years after hepatectomy   . The treatment options for intrahepatic recurrent HCC are various, including repeated hepatectomy, radiofrequency ablation (RFA), or transcatheter chemoembolization (TACE)    . Indeed, the treatment algorithm for recurrent HCC varies from center to center. Although repeated hepatectomy would be a curative treatment  , most patients might not be candidates for hepatectomy because of the small remnant liver or inadequate hepatic function. RFA is an effective treatment for small HCC ≤ 3 cm, whose results are comparable to those of hepatectomy. There are a number of studies describing the effectiveness of RFA for recurrent HCC after hepatectomy      .
In our hospital, surgeons and radiologists had consequence that RFA was a first-line treatment for the initial intrahepatic recurrence of small HCC ≤ 3 cm after hepatectomy, because of the good local control and low invasiveness. The purpose of this study was to retrospectively evaluate the effectiveness of RFA for the treatment of intrahepatic recurrent HCC after hepatectomy.
2. Materials and Methods
The institutional review board approved this retrospective study, and written informed waved (E-2127).
Between February 2002 and October 2011, 75 patients who underwent hepatectomy for HCC at our department of surgery and developed initial intrahepatic recurrence during postoperative follow-up periods, were referred to department of radiology for RFA. Our criteria for percutaneous RFA included: 1) nodular size ≤ 3 cm in diameter, 2) number of nodules ≤ 3, 3) absence of tumor thrombus in the portal vein, 4) absence of extrahepatic metastasis, 5) Child-Pugh class A or B, 6) prothrombintime ratio ≥ 50%, 7) platelet counts ≥ 50,000 cell/mm3, 8) depicted nodules with ultrasound (US).
The baseline characteristics of the 75 patients were shown in Table 1. Of the 75 patients, 57 were men and 18 women, whose age ranged from 44 years to 83 years (median, 69 years, mean, 68 years). 69 patients (92.0%) had a solitary nodule and 6 patients (8.0%) had two nodules. The size of total 81 nodules ranged 5 - 30 mm (median, 15 mm). Of 81 nodules, 69 nodules (85.2%) were <2 cm in diameter and 12 nodules (14.8%) were ≥2 cm. Of the 75 patients, 69 patients were classified into Child-Pugh A and eight patients into Child-Pugh B before RFA. Although most patients were believed to be candidates for repeat hepatectomy, surgeons and radiologists in our hospital had consequence that RFA is a first-line treatment for an initial recurrent HCC.
The diagnosis of recurrent HCC was made on the basis of imaging findings: newly presenting tumors at follow-up dynamic multi-detector CT, or dynamic EOB-MRI, and abdominal US after hepatectomy, and characteristic enhancement patterns on the early phase of dynamic multi-detector CT, dynamic EOB-MRI  , or contrast enhanced US using perfluorocarbon microbubbles (Sonazoid, Daiichi-Sankyo, Tokyo, Japan)  in 77 nodules, and confirmed by percutaneous needle biopsy in 4 nodules. Intrahepatic recurrent HCC were diagnosed from 3 to 150 months after hepatectomy (mean, 28.9 months; median, 19.5 months).
Table 1. Patients’ background and tumor characteristics.
2.1. RFA Procedures
An internally cooled electrode system (Cool-tip RF Ablation System, Covidien, MA)  was used for 69 patients. This system consists of a 480-kHz generator; a 20- or 15-cm-long, 17-gauge cooled-tip radiofrequency electrode with a 2- to 3-cm-long exposed metallic tip; and a dispersive pad applied to the patient’s skin. Grounding was achieved by attaching a dispersive pad to each of the patient’s thighs. RF 2000 system (Radio Therapeutics Corporation, Mountain View, CA, USA)  was used for 6 patients. This system consists of a 460-kHz generator; a LeVeen 15-gauge monopolar array electrode with 8 or 10 expandable electrode tines with 2- or 3-cm diameter; and a dispersive pad applied to the patient’s skin. Grounding was achieved by attaching a dispersive pad to each of the patient’s thighs. Three authors (K.S., T.S., and R.I.), who have 8, 29, and 10 years of experience each in interventional radiology, performed RFA under US guidance. 54 nodules were treated under guidance of B-mode US, and 27 nodules treated under guidance of Sonazoid-CE-US  .
On contrast harmonic sonography using Sonazoid (Sonazoid-CEUS), early vascular imaging which can show the perfusion features of the microvascular bed of the liver parenchyma and tumor, and late liver parenchymal imaging (Kupffer imaging) can be obtained. After Sonazoid-CEUS was performed, a radiofrequency electrode was introduced into a HCC nodule under guidance of Kupffer imaging.
Combined RFA and TACE   were performed in 22 nodules of 21 patients. First TACE was performed: introducing a microcatheter into a feeding hepatic artery segmentally or sub-segmentally after diagnostic celiac and superior mesenteric arteriography. Then, emulsion of 20 - 40 mg doxorubicin (Epirubicin, Tokyo) and 2 - 4 ml Lipiodol (iodized-oil, Tokyo) was injected via the microcatheter followed by injection of gelatin particles (Gelfoam or Gelpart, Tokyo). Then, RFA was performed within 1 week later.
Dynamic CT was performed within a week after RFA to evaluate therapeutic effectiveness  . Residual unablated lesions were defined as irregular peripheral-enhancing foci in the ablation zones. The residual unablated lesions were treated with additional RFA. When a non-enhancing area larger than the nodule was noted at post-treatment dynamic CT, we diagnosed the therapeutic efficacy as complete.
The patients were followed up in the clinic at 2- or 3-month intervals. At each follow-up visit, blood tests, including those to measure levels of serum a-feto- protein and prothrombin induced by vitamin K absence or agonist II (PIVKA- II), were performed. Dynamic CT or dynamic EOB-MRI was performed at 3- month intervals. Local recurrence was defined as a newly appearing enhancing lesion in or near the treated nodule  . The occurrence of new lesions in the liver was also evaluated with follow-up dynamic CT or dynamic EOB-MRI. Two authors (M.K. and T.K.) assessed these findings. Local recurrence or occurrence of new lesions in the liver was treated with RFA, TACE, or hepatic resection. RFA was performed when three or fewer nodules 3 cm or less in diameter were detected, and TACE was performed if the largest nodule detected was larger than 3 cm or when there were more than three nodules, regardless of their size.
Overall survivals were defined as the interval between the first RFA and the death or the last visit to the outpatient clinic until May 2015. Local recurrence rates were defined as the interval between the first RFA and diagnosis of local recurrence, the last visit to the outpatient clinic, or the death until May 2015. Event-free survival rates were defined as the interval between the first RFA and diagnosis of local recurrence or occurrence of new HCC in the liver, distant metastasis, the last visit to the outpatient clinic, or the death until May 2015. The follow-up periods after RFA for recurrent HCC ranged 3 month to 151 months (median, 55 months, mean, 58.9 months).
The prognostic factors of the baseline characteristics before RFA included patient age (≥70 years or <70 years), gender, Child-Pugh classes (A or B), serum AFP level (>100 ng/mL or <100 ng/mL), PIVKA-2 (>100 ng/mL or <100 ng/mL), nodular size, ablated margin (<5 mm or ≥5 mm), combined TACE or not, local recurrence (present or absent) were assessed with the Cox proportional hazard models. We also evaluated the prognostic factors of the clinico-patho- logic variables related to hepatectomy. These were Child-Pugh classes (A or B), serum AFP level (≥100 ng/mL or <100 ng/mL), PIVKA-II (≥100 ng/mL or <100 ng/mL), and resected tumor size (≥5 cm or <5 cm) before hepatectomy. The factors during and after hepatectomy included positive microvascular invasion or negative, degree of the differentiation of HCC (poorly or moderate/well-differentiated), and recurrence interval after hepatectomy (≥1 year or <1 year). These were assessed by using the Cox proportional hazard models.
2.4. Statistical Analysis
SPSS was used. A P value < 0.05 was considered to indicate a significant difference. Data processing and analysis were performed with commercially available software (SPSS for Windows, version 13.0; SPSS, Chicago, IL, USA).
A total of 81 RFA sessions were performed for 81 HCC nodules of 75 patients. Of 81 nodules, 78 were completely ablated with an initial RFA session (Figure 1): primary therapeutic effectiveness, 96.3% (78/81 nodules). In the other three nodules, the second RFA session was added for the residual enhanced foci and complete therapeutic effectiveness was obtained. In all the 81 nodules secondary effectiveness was 100% (81/81 nodules). Of the 81 nodules, 61 nodules (75.3%)
(a) (b) (c) (d)
Figure 1. 75-year-old man with HCV-cirrhosis, who had undergone central bisegmentectomy 1.5 year ago, developed a recurrent HCC. (a) Early-phase dynamic EOB-MRI shows an enhanced nodule sized 1.7 cm in segment 7. (b) Hepatobiliary phase at dynamic EOB-MRI shows a low intense nodule in segment 7. (c) The nodule was invisible with B-mode US, but vascular imaging of Sonazoid-CEUS depicted a highechoic nodule (an arrow). (d) The nodule was noted as low echoic, and RFA was performed under Kupffer imaging. (e) Dynamic CT obtained 3days after RFA showed a low dense area larger than the recurrent nodule. The nodule was completely ablated.
had the circumferential ablative margin with >5 mm (ideal circumferential ablative margin) and in 20 nodules (24.7%) the circumferential ablative margin was <5 mm.
During follow-up periods after RFA (3 to 151 months, median, 55 months), local recurrence was noted in 10 nodules of 10 patients (10/81 nodules = 12.3%). The rates of local recurrence were 1-yr, 3-yr, 5-yr, and 8-yr were 7.6%, 12.0%, 12.0%, and 12.0%, respectively. During the follow-up periods, 36 patients were alive and 39 died. The cumulative overall survival rates of 1-yr, 3-yr, 5-yr, and 10-yr were 97.3%, 79.1%, 56.6%, and 32.2%, respectively (Figure 2). The median survival time was 55 months. Of the 75 patients, 66 patients developed new HCC in the liver, lung metastasis in 5 patients, bone metastasis in 3 patients, lymph node metastasis in 2 patients. Nine patients were no recurrent HCC in the liver or no distant metastasis. The cumulative disease-free survival rates of 1-yr, 3-yr, and 5-yr were 42.7%, 15.8%, and 12.6%, respectively (Figure 3). The median disease-free survival time was 12.3 months. There were significant differences among the subgroups divided by the Child-Pugh class (A or B) before RFA (p = 0.007) as to overall survival rates, and the subgroups divided by the Child-Pugh class (A or B) in clinico-pathologic variables related to hepatectomy as to disease-free survival rates (p = 0.004) (Table 2). There were no significant differences among the subgroups that were divided by other variables.
Current imaging modalities, including dynamic multi-detector CT, dynamic EOB- MRI, or Sonazoid-contrast-enhanced US have high sensitivity for diagnosing HCC    . Especially patients with prior hepatectomy for HCC are high-risk groups of occurring HCC in the remaining liver    . In our
Figure 2. Graphs shows overall survival rates in patients with recurrent HCC nodules treated with RFA (n = 75). The overall survival rates of 1-yr, 3-yr, 5-yr, and 10-yr were 97.3%, 79.1%, 56.6%, and 32.2%, respectively.
Figure 3. Graphs shows disease-free survival rates in patients with recurrent HCC nodules treated with RFA (n = 75). The disease-free survival rates of 1-yr, 3-yr, and 5-yr were 42.7%, 15.8%, and 12.6%, respectively.
Table 2. Prognostic factors of overall survival rate and disease-free survival rates.
study, postoperative follow-up dynamic CT or dynamic EOB-MRI was frequently performed at 3-month intervals in all patients. So early diagnosis of recurrent HCC could be performed. The size of the 81 recurrent HCCs ranged 5 - 30 mm (median, 15 mm), and 69 of 81 nodules (85.2%) were ≤2 cm in diameter.
For treatment of native HCC ≤ 2 cm in diameter, overall survival rates of patients treated with RFA are comparable to those of hepatectomy   . There have been some reports comparing the outcomes of repeated hepatectomy and RFA for recurrent HCC after hepatectomy     . Chan AC et al.  compared the efficacy of salvage liver transplantation, repeated hepatic resection and RFA in the treatment of intrahepatic HCC recurrence after hepatectomy; salvage liver transplantation and repeated hepatic resection showed comparable survival outcomes, but both treatments were significantly better survival outcomes than RFA. However, Song KD et al.  described that RFA and repeated hepatectomy attained similar survival benefits in the management of recurrent HCC after hepatectomy. They compared the long-term outcomes of repeated hepatic resection and RFA for recurrent HCC by using propensity score matching. They showed that the overall survival rates at 1, 3, 5, and 8 years were 98.7%, 85.7%, 72.1%, and 68.6%, respectively, and the disease-free survival rates at 1, 3, and 5 years were 71.8%, 45.1%, and 39.4% in the RFA group. No statistical differences of overall survivals and disease-free survivals were noted between RFA and repeated hepatectomy groups.
However, the repeated hepatectomy would be technically more difficult, because of intra-abdominal adhesions caused by previous hepatic resections as well as new growth of intrahepatic vascular structures after previous hepatic resection, and there is a potential risk of postoperative hepatic failure after repeated hepatic resection in patients with limited hepatic function reserve  . On the other hand, RFA can be repeatedly performed for local recurrence after both RFA and new HCC appeared in the remaining liver. Moreover, RFA is less morbidity and mortality than hepatectomy. So, in our study RFA has been performed as a first-line treatment for the initial intrahepatic recurrence after hepatectomy, although most of the 75 patients would undergo repeated hepatectomy because of the good hepatic reserve. Our results of the cumulative overall survivals: 1-yr, 3-yr, 5-yr, and 10-yr, 97.3%, 79.1%, 56.6%, and 32.2% and disease-free survivals: 1-yr, 3-yr, and 5-yr, 42.7%, 15.8%, and 12.6%, respectively were comparable to those of the previous reports.
In two studies, long-term survival of patients with native HCC treated with percutaneous RFA as a first-line treatment depended upon the Child-Pugh class, serum AFP level, tumor size, and multiplicity of tumors   . In the study of Choi D, et al.  , high serum AFP level before RFA and resected tumor size at initial hepatectomy were independent significant predictive factors of long-term survival in patients with recurrent HCC after hepatectomy. Lu et al.  reported that the pre-ablation serum AFP level was the only prognostic predictor, and recurrence intervals after hepatectomy and tumor staging were not related to survival. In our study, a prognostic factor affected to overall survivals was Child-Pugh classification (Child A or B) before RFA and that affected to event-free-survivals was Child-Pugh classification (Child A or B) before hepatectomy. Other factors, such as the gender and age etc. did not influence the prognosis.
Takaki et al.  showed that a single tumor at the time of initial hepatectomy and a low a-fetoprotein level (≤100 ng/mL) at recurrence were significantly favorable independent factors affecting overall and recurrence-free survival. Choi et al.  described a lower serum AFP level (≤100 ng/mL) before RFA or with small resected tumors (≤5 cm) demonstrated better survival results  . In our study the most patients had small (median, 15 mm), single recurrent tumor, so tumor size and multiplicity of tumors would not be included into prognostic factors.
This study has several important limitations. First, this was a single-arm retrospective study. RFA was not performed for HCC nodules undetected at US. Moreover we should compare the outcomes between RFA and repeated hepatectomy.
In conclusion, RFA was an effective, useful therapeutic option for treatment of recurrent HCC after hepatectomy. Child-Pugh Class (A or B) before RFA and Child-Pugh classification (Child A or B) in clinico-pathologic variables related to hepatectomy were significant prognostic predictors of long-term survival. The use of RFA was recommended for the treatment of recurrent tumors after hepatic resection.