Percutaneous radiofrequency ablation therapy for patients with hepatocellular carcinoma during occlusion of hepatic blood flow

Comparison with standard percutaneous radiofrequency ablation therapy

Authors


Abstract

BACKGROUND

The therapeutic efficacy of radiofrequency ablation (RFA) for hepatocellular carcinoma (HCC) is limited by the small volume of coagulation necrosis obtained at each activation of the RF system and the sometimes irregular burn shape due to the proximity of large vessels that have a cooling effect. To improve the efficacy of RFA, the authors designed RFA with balloon occlusion of the hepatic artery (balloon-occluded RFA). In this study, we investigated the efficacy of balloon-occluded RFA and compared the coagulation diameters obtained with balloon-occluded RFA and standard RFA.

METHODS

We retrospectively studied 31 patients with 42 HCC lesions measuring less than 4 cm in the greatest dimension. We performed balloon-occluded RFA for 12 patients (n = 15 nodules) and standard RFA for 19 patients (n = 27 nodules). Initial therapeutic efficacy was evaluated with dynamic computed tomography scan performed 2 weeks after one treatment.

RESULTS

There were no significant differences in the ablation conditions such as the frequency of a fully expanded electrode, the number of needle insertions, application cycles, or treatment times between the two groups. However, the greatest dimension of the area coagulated by balloon-occluded RFA was significantly larger (greatest long-axis dimension, 36.6 ± 3.8 mm; greatest short-axis dimension, 30.1 ± 6.0 mm; n = 15 lesions) than that coagulated by standard RFA (greatest long-axis dimension, 26.7 ± 6.4 mm; greatest short-axis dimension, 23.1 ± 5.0 mm; n = 27 lesions; greatest long-axis dimension, P < 0.001; greatest short-axis dimension, P < 0.001).

CONCLUSIONS

Balloon-occluded RFA is superior to standard RFA for the treatment of many hepatocellular lesions, especially when larger volumes of coagulation are required. Cancer 2002;95:2353–60. © 2002 American Cancer Society.

DOI 10.1002/cncr.10966

Percutaneous ethanol injection (PEI) is widely performed as a percutaneous local treatment for small hepatocellular carcinoma (HCC).1 However, this modality is occasionally ineffective because of inhomogeneous distribution within the tumor.2, 3 Thermal methods of ablation including percutaneous microwave coagulation4 and radiofrequency (RF)5–12 were developed as new methods for percutaneous local treatment of small HCC. Both microwave and RF are superior to PEI10, 13, 14 and are equivalent in their therapeutic efficacy.

Radiofrequency ablation (RFA) is performed frequently in Japan.15–17 However, the limited volume of the coagulation necrotic area obtained at each activation of the RF system and the sometimes irregular burn shape due to the proximity of large vessels that have a cooling effect have limited its therapeutic efficacy for the treatment of HCC.18

Both animal and clinical studies have suggested that reduction of blood flow can improve tumor ablation efficacy when using thermal ablation.18–24 To improve the efficacy of RFA, we designed RFA with balloon occlusion of the hepatic artery (balloon-occluded RFA). Based on preliminary results in HCC patients, balloon-occluded RFA increased the area of coagulation necrosis.15

In this study, we investigated the efficacy of balloon-occluded RFA using greater numbers of patients with HCC. In addition, we retrospectively compared coagulation diameters and complications of balloon-occluded RFA and standard RFA in patients with HCC.

MATERIALS AND METHODS

Patients

We retrospectively studied 31 patients with 42 HCC lesions measuring less than 4 cm in the greatest dimension. They were admitted to the Department of Gastroenterology and Hepatology, Yamaguchi University School of Medicine, between March 1999 and January 2001.

Of these patients, 12 (n = 15 nodules) were treated with balloon-occluded RFA and 19 (n = 27 nodules) were treated with standard RFA. The diagnosis of HCC was made by imaging studies and/or histologic findings and was based on elevated serum levels of α-fetoprotein and/or of des-γ-carboxy-prothrombin.

In the balloon-occluded RFA group, eight patients had not received previous treatment for HCC and four had previously undergone treatment for HCC (percutaneous hot water injection and microwave coagulation therapy,25 PEI,1 and transcatheter arterial embolization [TAE]). Of the four patients who had previously undergone treatment for HCC, three had recurrent intrahepatic nodules and one had a recurrent nodule at the margin of pretreatment 1 year after PEI. In the standard RFA group, 13 patients had not received previous treatment. The other six patients had previously undergone treatment for HCC (percutaneous microwave coagulation therapy,4 PEI,1 transcatheter arterial chemoembolization, and surgical resection) and all had recurrent intrahepatic nodules. The histologic diagnosis was confirmed by ultrasonography-guided fine-needle biopsy. The histologic grades of differentiation were defined as well differentiated, moderately differentiated, or poorly differentiated.26 The underlying liver disease was classified as Child–Pugh Class A, B, or C.27

Patients were asked to provide written informed consent to enter the study, which was approved by the Institutional Review Board of Yamaguchi University Hospital.

Techniques

RF system

The RF delivery system (500PA, Radiofrequency Instertitial Thermal Ablation [RITA] Medical System, Mountain View, CA) consisted of an active, expandable needle electrode, an RF generator producing radiofrequency waves with a frequency of 460 kHz, a maximum output power of 50 W; and a monitor that displayed the tissue temperature and impedance around the tip of the needle. The RF needle electrode (model 30), which had four retractable hooks, measured 15 gauge in diameter. Four expandable hooks with a thermometer were deployed from the tip of the needle electrode. The electrode was connected to a generator through a flexible cable. RF power was adjusted to keep the temperature between 80 °C and 110 °C. RF energy was applied for 8–10 minutes (one application cycle). When necessary, after the first ablation, the electrode was pulled out 1.5 cm and a second application was started. These tumors were treated with one to two insertions of the electrode per procedure. One to two application cycles were performed for balloon-occluded RFA and one to four application cycles were employed for standard RFA.

Balloon-occluded RFA

The catheter placement was done by using the Seldinger approach through the femoral artery with a 5.0-Fr catheter (Clinical Supply, Gifu, Japan). Angiography–computed tomography scan (angio-CT)28 was accomplished with a Somatom plus 4F system (Siemens, Erlagen, Germany). Before treatment, CT during arterial portograpy (CTAP) and CT arteriography (CTA) were performed. A catheter was placed in the superior mesenteric artery for CTAP and in the common hepatic artery for CTA. CT scanning for CTAP started 30 seconds after 50 mL of contrast medium, Iomeron 300 (Eisai, Tokyo, Japan) or Proscope 300 (Tanabe, Tokyo, Japan) diluted with saline (1:1 ratio), was injected at a rate of 2 mL per second. CT scanning for CTA started 5 seconds after the injection of 15–20 mL of diluted contrast medium at a rate of 2 mL per second.

The procedure was performed under real-time ultrasound guidance (LOGIX 500MD, GE Yokogawa Medical Systems, Tokyo, Japan, or SSD-5500, ALOKA, Tokyo, Japan) with a 3.5-MHz convex probe. CTAP and CTA were performed before treatment. Using a 21-gauge needle, a biopsy was performed under local anesthesia induced by 5 mL of 0.5% lidocaine injected through the skin into the peritoneum along a predetermined puncture line. Before treatment, 15 mg of pentazocine hydrochloride and 25 mg of hydroxyzine hydrochloride were administered intramuscularly, after which 15 mg of pentazocine hydrochloride was administered intravenously. We then performed balloon-occluded RFA. A balloon with an 11-mm outer diameter attached to a 5-Fr catheter (Clinical Supply) was used for occlusion of the common hepatic artery. During interruption of the hepatic arterial flow, the RF generator was activated. Hepatin sodium (10 U/mL) was sometimes administered to prevent thrombosis of the hepatic artery during the occlusion. After treatment, CTA was performed again to evaluate the effect of the procedure (RFA with angio-CT assistance).15 Patients were instructed to lie quietly in bed for 12 hours after treatment.

In patients who underwent balloon-occluded RFA and standard RFA, the following serologic values were measured before treatment and 1 day, 3 days, and 1 week after treatment: transaminases, alkaline phosphatase, bilirubin, electrolytes, hemoglobin, fibrinogen, creatinine, and prothrombin activity, along with a complete blood cell count.

Assessment of Therapeutic Efficacy

Dynamic CT scans were performed 2 weeks after treatment to evaluate the initial therapeutic effect. On dynamic CT scans obtained less than 1 week after treatment, a rim of hyperattenuation around the nonenhanced region was apparent.10, 15 We measured the size of the nonenhanced portion of the liver. Tumor necrosis was considered complete when no areas of enhancement were seen in the tumor or at the periphery on CT scans. If the size of the necrotic area was almost the same as that of the tumor and CT scans showed partial enhancement of a portion of the tumor, we performed an additional cycle of RFA. The second RF procedure was planned as soon as possible for patients who did not show a complete response after the first procedure. Repeated dynamic CT scans were performed every 3–4 months thereafter. In the balloon-occluded RFA group, follow-up lasted for 10.1–22.3 months (mean ± SD, 18.3 ± 3.5 months). In the standard RFA group, follow-up lasted for 12.4–32.4 months (mean ± SD, 19.8 ± 5.6 months).

AFP and DCP assays were performed before treatment, 1 month after treatment, and subsequently every 1–2 months.

Statistical Analysis

The data were expressed as mean ± SD. Statistical analyses were performed using the Student t test for unpaired data, contingency table analysis, and the Mann–Whitney U test as appropriate. A P value of less than 0.05 was considered statistically significant.

RESULTS

Preprocedural Findings

Clinical characteristics of patients are shown in Table 1. There were no significant differences between the two groups for gender, age, Child–Pugh class, viral markers, tumor size, histologic findings, and initial/previous treatment. At the time of analysis, there was no difference in median follow-up times for patients who received balloon-occluded RFA or standard RFA.

Table 1. Clinical Characteristics of Patients with HCC treated by BoRFA or Standard RFA
Clinical characteristicsBoRFA (n = 12, 15 nodules)Standard RFA (n = 19, 27 nodules)P valuea
  • HCC: hepatocellular carcinoma; BoRFA: balloon-occluded radiofrequency ablation; RFA: radiofrequency ablation; HCV: hepatitis C virus; HBV: hepatitis B virus; well: well differentiated; moderate: moderately differentiated; poor: poorly differentiated; NS: not significant.

  • a

    Mann–Whitney U test or Student t test for unpaired data.

  • b

    Data are expressed as mean ± SD.

Gender (male/female)7/512/7NS
Age (yrs)b65.3 ± 11.566.9 ± 6.1NS
Child-Pugh (A/B/C)11/1/014/2/3NS
HCV/HBV11/117/2NS
Tumor size (mm)b20.4 ± 6.420.0 ± 6.6NS
Histologic findings (well/moderate/poor)8/3/412/9/2NS
Initial/previous treatment8/413/6NS

Treatment Efficacy

CTA obtained immediately after balloon-occluded RFA showed a rim of hyperattenuation around the nonenhanced region in all patients. This ring had disappeared on CT scans obtained 2 weeks after balloon-occluded RFA, and the enhanced region of the tumor before treatment changed to a nonenhanced region in all cases (Fig. 1). Treatment was successful after one session of balloon-occluded RFA for 12 nodules (80%). An additional session of standard RFA was performed for three patients (three nodules) because the size of the necrotic area was almost the same as that of the tumor. The results from 31 patients with HCC who underwent balloon-occluded RFA and standard RFA are summarized in Tables 2 and 3, respectively.

Figure 1.

Balloon-occluded radiofrequency ablation with combined angiography and computed tomography (CT) assistance in a patient with hepatocellular carcinoma in the S6 segment of the liver (patient 12). (A) CT angiography (CTA) before treatment showing a 15 × 13-mm tumor (arrow). (B) CTA immediately after treatment. A hyperattenuating ring around the nonenhanced region is apparent. The nonenhanced region is larger than the tumor. It is possible to evaluate the real-time therapeutic effect. (C) Dynamic CT scan taken 2 weeks after treatment shows a 39 × 30-mm necrotic area.

Table 2. Summary of the Results of the Balloon-Occluded Radiofrequency Ablation
Case no.Site/size (mm)Histologic finding (differentiation)No. of treatmentsElectrode insertionstreatment times (min)Coagulation area after BoRFA (mm)AFPDCPDetection of residual fociFollow-up (mos)
Before/AfterBefore/After
  • AFP: α-fetoprotein; DCP: des- γ-carboxy-prothrombin; BoRFA: balloon-occluded radiofrequency ablation.

  • a

    Two treatment sessions are standard for RFA.

1S7/25Poor1110 × 236 × 3696.7/27.827/33(+)22.3
2S5/38Moderate2a1/110 × 2/1039 × 3614.5/12.0184/18(−)22.0
3S6/30Well1110 × 235 × 3219.3/8.066/16(−)21.9
 S3/18Well111042 × 35  (−) 
4S6/14Poor111039 × 36149.9/135.912/0(−)20.5
5S7/16Poor111030 × 2026.0/35.776/94(−)20.1
6S6/18Moderate2a1/110 × 2/10 × 241 × 285.5/6.20/16(−)14.7 dead
7S8/19Moderate111038 × 3034.5/20.338/62(−)19.5
8S6/22Poor2a1/110/1033 × 33215.7/162.132/10(−)18.1
9S5/20Well1210.7 + 9.740 × 2066.2/90.818/16(−)17.0
10S4/17Well111033 × 3358.0/24.112/13(−)16.9
 S7/20Well111036 × 27  (−) 
11S5/14Well111030 × 203.5/4.028/22(−)16.7
1S5/20Well1110 × 238 × 3620.3/26.370/15(−)15.0
12S6/15Well1110 × 239 × 3010.3/10.610/14(−)10.1
Table 3. Summary of the Results of the Standard Radiofrequency Ablation
Case no.Site/size (mm)Histologic finding (differentiation)No. of treatmentsElectrode insertionstreatment times (min)Coagulation area after 1 session of RFA (mm)AFPDCPDetection of residual fociFollow-up (mos)
Before/AfterBefore/After
  1. AFP: α-fetoprotein; DCP: des-γ-carboxy-prothrombin; ND, not determined.

1S2/18Well128 × 2 + 332 × 297.5/8.014/0(+)32.4
2S8/22Moderate1210 × 331 × 3111.2/11.595/23(+)30.5
3S8/24Moderate128 × 233 × 27104.5/11.319/0(+)28.8
4S7/22Moderate21/110/1020 × 201260.7/22.110.7/29(−)23.8
5S4/15ND111023 × 1885.4/14.820/21(−)21.5 dead
 S7/13ND111020 × 20  (−) 
6S8/21Well111024 × 21142.1/62.934/29(−)20.0
7S2/19Moderate111025 × 2076.1/26.6154/42(−)12.4 dead
8S5/17Well21/110/1023 × 2360.8/42.816/ND(−)19.6
9S8/35Moderate1110 × 240 × 2818.1/32.0405/348(+)19.1
10S5/17Moderate111022 × 2026.4/26.4121/31(+)18.8
 S6/17ND111022 × 20  (+) 
11S3/16Well111020 × 1616.1/14.814.8/25(−)18.4
12S5/16Well1110 + 9.828 × 21124.7/ND95/ND(−)18.1
13S2/30Moderate1110 × 232 × 25132.0/52.238/28(−)17.6
14S7/28Moderate125.8 + 10 + 1033 × 307.7/9.311/34(−)17.6
15S6/12Poor111023 × 23129.6/47.123/46(−)17.4
16S8/35Moderate125 + 10 + 10 + 1043 × 3648.4/37.622/29(−)16.7
 S6/17Well111027 × 21  (−) 
 S5/7Well111021 × 20  (−) 
 S6/22Well111018 × 18  (+) 
 S8/21Well111023 × 23  (−) 
17S3/21Poor1210 + 1020 × 17142.4/191.113/15(−)12.7 dead
5S7/17Well111032 × 2820.4/34.715/26(−)16.0 dead
18S7/14Well111030 × 1754.1/15.120/17(−)16.7
19S5/28Well1210 + 9.8 + 1032 × 2979.8/49.7126/36(−)13.7
 S8/16ND118 + 1025 × 22  (−) 

Complications

In the balloon-occluded RFA group, localized warmth and pain occurred in all cases (100%). Transient fever occurred in 10 cases (77%). In the standard RFA group, localized warmth and pain occurred in 14 of 25 cases (56%, P = 0.005) and transient fever occurred in 10 of 25 cases (40%, P = 0.031). Transient increases in serum asparate aminotransferase (AST) concentration were observed in all patients. However, 1 week after treatment, the concentration decreased almost to the pretreatment level. There were no significant differences between the two groups in serum AST concentration before treatment and at 1 day, 3 days, and 1 week after treatment.

In the balloon-occluded RFA group, one patient had intraperitoneal hemorrhage on the day of the procedure and required blood transfusion. A skin burn occurred in one patient. In the standard RFA group, inflammation of the intestine occurred in one patient.

Balloon-Occluded RFA Versus Standard RFA

We compared the coagulation diameters for balloon-occluded RFA and standard RFA on CT scans obtained 2 weeks after one treatment. There were no significant differences in ablation conditions such as the frequency of a fully expanded electrode, the number of needle insertions, application cycles, or treatment times between the two groups. The number of needle insertions for standard RFA increased (balloon-occluded RFA, 1.07 ± 0.26; standard RFA, 1.26 ± 0.44, P = 0.086).

The greatest dimension of the area coagulated by balloon-occluded RFA was significantly larger (greatest long-axis dimension, 36.6 ± 3.8 mm; greatest short-axis dimension, 30.1 ± 6.0 mm; n = 15 lesions) than that coagulated by standard RFA (greatest long-axis dimension, 26.7 ± 6.4 mm; greatest short-axis dimension, 23.1 ± 5.0 mm; n = 27 lesions; greatest long-axis dimension, P < 0.001; greatest short-axis dimension, P < 0.001; Table 4).

Table 4. Comparison of BoRFA and Standard RFA Groups
CharacteristicsBoRFA (n = 15)Standard RFA (n = 27)P valuea
  • BoRFA: balloon-occluded radiofrequency ablation; RFA: radiofrequency ablation.

  • a

    Student t test for unpaired data or contingency table analysis.

  • b

    Data are expressed as mean ± SD.

Frequency of fully expanded electrode14/15 (93%)24/27 (89%)0.638
No. of needle insertionsb1.07 ± 0.261.26 ± 0.440.086
Application cyclesb1.47 ± 0.521.63 ± 0.880.463
Treatment time (min)b14.7 ± 5.215.3 ± 7.60.764
Coagulation area   
 Long-axis diameter (mm)b36.6 ± 3.826.7 ± 6.4P < 0.001
 Short-axis diameter (mm)b30.1 ± 6.023.1 ± 5.0P < 0.001

Detection of residual foci of unablated tumors occurred in one patient (one nodule; [7%]; Case 1) 7 months after treatment in the balloon-occluded RFA. We performed the same treatment again. Five (42%) showed intrahepatic recurrence. Of the five patients who showed intrahepatic recurrence, one was treated with balloon-occluded RFA, one was treated by operation, two were treated with standard RFA, and one was not treated because of severe liver damage.

In the standard RFA group, detection of residual foci of unablated tumors occurred in six patients (seven nodules [26%]) at 2–30 months (2, 8,10,10,23,30 months) after treatment. Three patients were treated with RFA, one was treated by surgery, and two were treated with chemotherapy. Significant differences were not found in the detection of residual foci of unablated tumors between the two groups (P = 0.13).

During follow-up, one patient in the balloon-occluded RFA group and three patients in the standard RFA group died. The cause of death was liver failure in the balloon-occluded RFA group and cancer progression in all patients in the standard RFA group.

DISCUSSION

RF has been successful in destroying tumors in patients with HCC. It has many advantages compared with other methods of thermal ablation. Rossi et al.8 reported that the maximum diameter of coagulation necrosis was 2.5 cm. This was achieved with a single electrode insertion using an RF generator with a frequency of 480 kHz, a maximum output power of 50 W (electrosurgical generator, RITA Medical System), and four expandable needle electrodes. In our study, the greatest long-axis dimension of standard RFA was 26.7 ± 6.4 mm. using an RITA 500PA with a frequency of 460 kHz, a maximum output power of 50 W, and four expandable needle electrodes. A necrotic area of up to 3 cm in dimension produced with a cooled RF single electrode was observed in tumors10 and cluster electrodes achieved a much larger necrotic area.9, 12, 29 However, the limited volume of coagulation necrosis obtained at each activation of the RF system and the sometimes irregular burn shape due to the proximity of large vessels that have a cooling effect have limited the therapeutic efficacy of RFA for the treatment of HCC.18

To improve its efficacy, we performed RFA using only occlusion of the hepatic artery. In a preliminary trial, balloon-occluded RFA (n = 5 lesions) increased the area of coagulation necrosis.15 In the current study, we investigated the efficacy of balloon-occluded RFA using a larger number of patients with HCC. In addition, we retrospectively compared the coagulation diameters after balloon-occluded and standard RFA in patients with HCC.

The greatest dimension of a coagulation area achieved with standard RFA using the RITA 500PA and an RF needle electrode (Model 30) was 26.7 ± 6.4 mm (greatest long-axis dimension). In the current study, the average tumor size was 20.0 ± 6.6 mm. Therefore, the coagulation area for one session of standard RFA was insufficient. Multiple electrode insertions were required to destroy small tumors because of the limited size of the necrotic area created with RF system. Conversely, the greatest long-axis dimension of a coagulation area with balloon-occluded RFA was 36.6 ± 3.8 mm. Although standard RFA required more needle insertions (balloon-occluded RFA, 1.07 ± 0.26; standard RFA, 1.26 ± 0.44, p = 0.086), balloon-occluded RFA increased the area of coagulation necrosis.

In animal studies, the reduction of blood flow during RF application increased the volume of necrosis.18–22 Rossi et al.22 reported that occlusion of the hepatic artery during the RF procedure did not significantly increase the coagulation area, whereas occlusion of the portal vein or both vessels (both hepatic artery and portal vein) resulted in larger coagulation areas. However, Chinn et al.18 also reported that the effect of vascular occlusion on RFA could be accomplished with hepatic artery occlusion alone. However, these studies used a normal liver. Therefore, results cannot be compared because of the vascular difference between the normal liver and HCC nodules. In fact, the hepatic artery mainly nourishes the HCC nodules. The surrounding hepatic tissue has a dual blood supply and is nourished mainly by the portal vein and hepatic artery, which provide about two-thirds and one-third of the blood flow, respectively.30 When the hepatic artery was occluded in HCC nodules, in which the portal blood flow decreased on CTAP and arterial blood flow increased on CTA, the cooling effect decreased because of the lack of blood supply (from both the hepatic artery and portal vein) to the tumor. We believed that the effect of hepatic artery occlusion on the HCC nodules might be similar to that of the Pringle maneuver (for both the hepatic artery and portal vein) in the normal liver. However, the exact mechanism needs further investigation. The development of a suitable experimental model that can reproduce HCC nodules is required.

In clinical studies, several methods have reported a significant difference in the increase in the RFA-induced coagulation necrosis area obtained with occlusion of the portal19, 23 or arterial flow.24 Rossi et al.24 reported that RFA was safe and effective in nonresectable HCCs measuring 3.5–8.5 cm in diameter. In their study, RFA was performed after interruption of the tumor arterial supply by occluding either the hepatic artery with a balloon catheter or feeding the arteries with gelatin sponge particles. However, they did not refer to the coagulation area. It is also difficult to compare the data of our study and their report because of the difference in the RF system. Our study provides the first comparison of balloon-occluded and standard RFA in patients with a single disease process, namely, HCC. We did not perform occlusion of the HCC feeding arteries with gelatin sponge particles because angio-CT15 cannot be be employed when TAE using gelatin sponge particles is being performed. The advantages of RFA with angio-CT assistance are that it is possible to detect small satellite nodules and to evaluate the real-time therapeutic effect immediately after balloon-occluded RFA.

In the current study, the rate of minor complications (localized warmth and pain and transient fever) was higher with balloon-occluded RFA compared with the standard RFA. However, there were no cases of interrupted treatment because of the complications. There were two major complications (intraperitoneal hemorrhage and a skin burn) with balloon-occluded RFA, which also occurred with RFA. There were no significant differences between the two groups in the serum AST concentration before and after treatment. One week after balloon-occluded RFA, liver function returned to almost the same level that was observed before treatment.

The main goal of percutaneous ablation is complete eradiation of the tumor. However, patients with moderately or poorly differentiated HCC have a high incidence of microscopic metastases adjacent to the primary tumor.31 Therefore, it is necessary to produce extensive necrosis not only of the tumor lesion but also of the surrounding parenchyma. The criteria for the safety margin, which means the ablated distance beyond the tumor margins, were not defined. In our study, the frequency of detection of residual foci of unablated tumors was high with a safety margin less than 5 mm. Although there were no significant differences between the two groups for detection of residual foci of unablated tumors, the incidence of residual foci of unablated tumors in balloon-occluded RFA decreased (P = 0.13). In addition, perfusion-mediated “heat sink” effect limits the heating of perivascular tumors,32 resulting in subsequent local tumor recurrence. Balloon-occluded RFA may also better ablate perivascular tumors, leading to reduced local control. The limits of the current study are due to patient population size. However, further study with greater patient accrual is likely to demonstrate meaningful statistical differences in local control and possibly even patient survival. Although the invasiveness of the angiographic procedure is a demerit of balloon-occluded RFA, larger areas of coagulation and the real-time therapeutic effect are merits of this procedure.

In the current study, we investigated the efficacy of balloon-occluded RFA using an RITA 500PA and an expandable electrode (Model 30). Other RF systems using a powerful RF generator with an expandable electrode,11 a cooled RF single electrode,10 or a cluster electrode9, 12, 29 have increased the coagulation area beyond that achieved in our study. However, our procedure can be performed equally well using other RF systems. We have also performed balloon-occluded RFA using a cool RF single electrode, which achieved a much larger necrotic area. It is also possible to perform balloon-occluded RFA for large HCC nodules (> 3 cm in diameter; data not shown). In addition, this report was a retrospective comparative study. Therefore, it is necessary to investigate this technique in a prospective randomized study. We have already started such a study using the cool RF single electrode.

In conclusion, balloon-occluded RFA is superior to standard RFA for treating many hepatocellular lesions, especially when larger coagulation areas are required. In addition, combined balloon-occluded RFA and angio-CT assistance15 techniques can achieve one-session treatment for patients with small HCC. Combined techniques have the advantage of a short stay in the hospital for HCC patients.

Ancillary