The present study was a single center and retrospective study. Prior to angiography and RFA, written informed consent was obtained from all the patients. The protocols for angiography and RFA were approved by the ethics committee of our department (Fig. 1).

We assessed the electronic medical records of patients registered in our database. The primary endpoint was post-RFA IDR-free survival, and the secondary endpoints were local tumor progression (LTP) and overall survival (OS).

We performed RFA in 269 consecutive treatment-naive patients diagnosed with solitary and hypervascular HCC in the Department of Gastroenterology and Hepatology, Osaka Red Cross Hospital, Japan, between January 2004 and October 2010. During the period from January 2004 to April 2005 and from December 2008 to October 2010, we routinely performed TAI-EML prior to RFA. Between May 2005 and October 2008, only lipiodol infusion was performed prior to RFA, as we were using the protocols of several clinical studies that we had participated in. Lipiodol (Lipiodol Ultra-Fluid, Schering Japan, Osaka, Japan) accumulates selectively in tumors when infused intra-arterially (14). Therefore, lipiodol infusion was performed to improve tumor visibility when assessing the effectiveness of RFA using dynamic computed tomography (CT). A representative case is shown in Fig. 2.

Of the 269 patients mentioned above, patients who met the following inclusion criteria were analyzed: a) complete ablation of RFA; and b) no evidence of other malignancies. We excluded patients who had: a) incomplete ablation of RFA; b) interferon treatment for chronic hepatitis C; c) nucleoside analogue treatment for chronic hepatitis B; and d) only segmental TAI before RFA due to the decision of attending physicians (Fig. 1). Complete ablation was defined as no apparent residual tumors on dynamic CT performed within 7 days after RFA and was determined by three radiologists (not blinded) experienced in liver imaging. In patients with incomplete RFA, TACE was performed after RFA.

HCC was diagnosed using abdominal ultrasound and dynamic CT scans (hyperattenuation during the arterial phase in all or some part of the tumor and hypoattenuation in the portal-venous phase) and/or magnetic resonance imaging (MRI), mainly based on the recommendations of the American Association for the Study of Liver Diseases (15). Arterial and portal phase dynamic CT images were obtained at approximately 30 and 120 sec, respectively, after the injection of the contrast material. In our department, abdominal angiography combined with CT (angio-CT) assistance was performed on all patients before RFA; CT during hepatic arteriography (CTHA) was performed with the catheter tip in the proper hepatic artery and CT during arterial-portography (CTAP) was performed with the catheter tip in the superior mesenteric artery, as Yamasaki et al reported that this technique was useful for detecting small satellite nodules (16). Then, we confirmed the presence of single and hypervascular HCC with no vascular invasion using CTHA and CTAP. Immediately after evaluation using angio-CT, TAI-EML or lipiodol infusion-alone via the proper hepatic artery was performed in the same session.

A catheter was introduced into the proper hepatic artery using the Seldinger technique. This was followed by an intra-arterial infusion via the proper hepatic artery according to tumor size and liver function, of an emulsion containing epirubicin (mean dose, 28.2 mg; range, 10–40 mg), mitomycin (mean dose, 6.2 mg; range, 4–10 mg) and lipiodol (mean dose, 2.1 ml; range, 1–3 ml) in the TAI group and of lipiodol alone (mean dose, 1.9 ml; range, 1–3 ml) in the lipiodol-alone group, respectively. When patients had poor liver function, the dosage of the anticancer agents and lipiodol were reduced.

Epirubicin (Farmorubicin; Kyowa Hakko Kirin Company, Ltd., Tokyo, Japan) is an anthracycline-based anticancer drug and is the 4′-epimer of doxorubicin. Epirubicin, as well as doxorubicin, binds to the deoxyribonucleic acid (DNA) in tumor cells, leading to suppression of DNA synthesis (17). Mitomycin (Mitomycin C; Kyowa Hakko Kirin Company, Ltd., Tokyo, Japan) is changed into activated metabolites by various enzymes, and these block DNA replication in tumor cells leading to antitumor effects (18). Due to the fact that epirubicin easily undergoes glucuronidation, its toxicity is milder than that of doxorubicin. In 1986, the Japan Epirubicin Study Group for Hepatocellular Carcinoma reported that in 53 HCC patients who had received TAI with epirubicin, eight of the patients had an objective response. A retrospective comparison of this result with that of patients that were treated using TAI with doxorubicin demonstrated that epirubicin was more effective than doxorubicin in terms of survival rate (17). Since that time, epirubicin has been conventionally used in TAI for HCC in Japan, and additional clinical trials have been performed in several faculties. Epirubicin alone or in combination with other chemotherapeutic agents such as mitomycin C has been used in TAI for HCC in Asian countries, including Japan (17). Considering this background, we have conventionally performed TAI for HCC in our department with an epirubicin-mitomycin-lipiodol emulsion.

In all patients, the first session of RFA was performed within 7 days after angiography. We routinely used a cool-tip needle (Radionics Corp., Burlington, MA, USA) while performing RFA. Under local anesthesia, using the intercostal or subcostal approach, an electrode was inserted under real-time ultrasound guidance. The initial treatment was planned with one ablation for tumors of <2 cm in diameter, and two or more ablations with the overlapping technique for tumors of ≥2 cm in diameter. When tumor ablation was complete, thermal ablation was performed along the needle track. We did not use the coaxial technique. All patients were carefully observed for treatment-related complications. The procedures were all conducted under ultrasound guidance by one of five operators who had at least 3 years of experience in performing RFA.

Follow-up after RFA consisted of monthly blood tests and monitoring of tumor markers, including α-fetoprotein (AFP) and des-γ-carboxy prothrombin (DCP), which were measured using a chemiluminescent enzyme immunoassay (Lumipulse PIVKAII Eisai, Tokyo, Japan). Abdominal ultrasonography and dynamic CT scans were obtained every 3–4 months following RFA.

We defined IDR as the recurrence of a new lesion (excluding extrahepatic metastasis) at a distant site from the ablation zone detected using a dynamic CT scan. When a new lesion appeared in the same segment as the original tumor, we defined it as IDR if it was not adjacent to the ablation zone. We defined local tumor progression (LTP) as the presence of a hypervascular nodule adjacent to the ablation zone after RFA revealed using a dynamic CT scan (19). Recurrence that was distant from the ablation zone in the same segment was not included in the definition of LTP. IDR and LTP were determined by the same three radiologists (not blinded) as described above.

Intercomparison of the lipiodol-alone group and the TAI group was performed using the unpaired t-test or Fisher’s exact test. Data were expressed as the mean ± standard deviation and were analyzed using univariate and multivariate analyses. The cumulative post-RFA IDR-free survival rate, the cumulative LTP rate and the cumulative OS rate were calculated using the Kaplan-Meier method, and tested using the log-rank test. Regarding factors contributing to IDR, the Cox proportional hazard model was used for multivariate analyses of factors that were considered significant in univariate analysis. These statistical methods were used to estimate the interval from RFA treatment. Data were analyzed using SPSS software, version 9.0 (SPSS Inc., Chicago, IL, USA) for Microsoft Windows. P-values <0.05 were considered to indicate statistically significant differences.

In 94 of the 182 patients analyzed in the present study, between January 2004 and April 2005 and between December 2008 and October 2010, TAI-EML was performed prior to RFA. Over the period from May 2005 to October 2008, infusion involving lipiodol-alone via the proper hepatic artery was performed prior to RFA in 88 patients.

The baseline characteristics of the two groups are shown in Table I. The mean tumor size was 2.06 cm (range, 0.9–3.2 cm) in the TAI group and 1.97 cm (range, 0.9–3.3 cm) in the lipiodol-alone group, respectively. There were no significant differences between the two groups in terms of baseline characteristics. At a median follow-up interval of 36.4 months (range, 6.2–83.6 months), 55 (62.5%) of the 88 patients in the lipiodol-alone group and 31 (33.0%) of the 94 patients in the TAI group had IDR. The cumulative post-RFA IDR-free survival rates at 1, 2 and 3 years were 74.0, 50.8 and 34.9%, respectively, in the lipiodol-alone group, and 90.8, 74.8 and 70.0%, respectively, in the TAI group (P<0.001). There was a significant difference in the post-RFA IDR-free survival rate between the two groups (P<0.001) (Fig. 3).

Using univariate analysis, age ≥70 years (P=0.046), platelet count ≥10×10⁴/μl (P<0.001), des-γ-carboxy prothrombin level (DCP ≥100 mAU/ml) (P=0.030) and TAI-EML before RFA (P<0.001) were found to be significant factors contributing to IDR (Table II). In the multivariate analyses involving the four factors that were found to be significant in the univariate analysis, hazard ratios (HRs), 95% confidence interval and P-values are detailed in Table III. A platelet count of ≥10×10⁴/μl and TAI-EML were found to be significant independent factors linked to IDR.

During the observation period, 16 patients (17.0%) in the TAI group and 28 patients (31.8%) in the lipiodol-alone group, respectively, had LTP. The 1-, 2- and 3-year LTP rates were: 8.4, 17.4 and 22.2%, respectively, in the TAI group and 8.6, 19.6 and 34.9%, respectively, in the lipiodol-alone group. In terms of the LTP rate, there was no significant difference between these two groups (P=0.145) (Fig. 4).

During the observation period, 23 patients (24.5%) in the TAI group and 30 patients (34.1%) in the lipiodol-alone group, respectively, succumbed to the disease. Four patients (4.3%) in the TAI group and six patients (6.8%) in the lipiodol-alone group, respectively, were lost to follow-up. The causes of mortality were HCC recurrence (17 patients), liver failure (four patients) and miscellaneous causes (two patients) in the TAI group, and HCC recurrence (23 patients), liver failure (four patients) and miscellaneous causes (three patients) in the lipiodol-alone group, respectively. The 1-, 3- and 5-year OS rates were: 96.0, 83.7 and 71.1%, respectively, in the TAI group and 96.0, 79.6 and 51.0%, respectively, in the lipiodol-alone group. In terms of the OS rate, there was a significant difference between these two groups (P=0.048) (Fig. 5).

Patients in the TAI group experienced transient fever, abdominal pain, appetite loss, nausea and elevation of liver enzymes after TAI. However, these side-effects were all grade 1 (as determined using the National Cancer Institute Common Terminology Criteria for Adverse Events, version 3.0) (20). Side-effects improved within 3 or 4 days folowing TAI. In the lipiodol-alone group, no side-effects were observed.

In terms of complications associated with RFA itself, pneumothorax in one patient, biloma in one patient and intra-abdominal bleeding in one patient in the TAI group, and biloma in one patient, refractory ascites in one patient and retroperitoneal bleeding in one patient in the lipiodol-alone group were observed. All patients with complications related to RFA improved during the same hospitalization period. However, one patient in the TAI group required re-admission due to biloma and improved after percutaneous drainage. There were no needle tract implantations and there were no mortalities due to RFA related complications.