Radiology
Vascular and Interventional Radiology Riccardo A. Lencioni, MD Hans-Peter Allgaier, MD Dania Cioni, MD Manfred Olschewski, MSc Peter Deibert, MD Laura Crocetti, MD Holger Frings, MD Joerg Laubenberger, MD Ina Zuber, MD Hubert E. Blum, MD Carlo Bartolozzi, MD
Index terms: Alcohol ablation, 761.1266 Liver, cirrhosis, 76.794 Liver, interventional procedures, 761.1266, 761.1269 Liver neoplasms, 76.323 Radiofrequency (RF) ablation, 761.1269 Published online before print 10.1148/radiol.2281020718 Radiology 2003; 228:235–240 Abbreviations: HCC ⫽ hepatocellular carcinoma PEI ⫽ percutaneous ethanol injection RF ⫽ radio-frequency RR ⫽ relative risk 1
From the Division of Diagnostic and Interventional Radiology, Department of Oncology, Transplants and Advanced Technologies in Medicine, University of Pisa, Via Roma 67, I-56125 Pisa, Italy (R.A.L., D.C., L.C., C.B.); and the Departments of Medicine II (H.P.A., P.D., H.F., I.Z., H.E.B.), Medical Biometry (M.O.), and Radiology (J.L.), University Hospital Freiburg, Germany. From the 2001 RSNA scientific assembly. Received June 17, 2002; revision requested August 8; final revision received November 26; accepted December 19. Address correspondence to R.A.L. (email:
[email protected]). Author contributions: Guarantor of integrity of entire study, R.A.L.; study concepts and design, all authors; literature research, R.A.L., D.C.; clinical studies, R.A.L., D.C., H.P.A.; data acquisition, R.A.L., H.P.A.; data analysis/ interpretation, all authors; statistical analysis, M.O.; manuscript preparation, definition of intellectual content, editing, revision/review, and final version approval, all authors.
©
RSNA, 2003
Small Hepatocellular Carcinoma in Cirrhosis: Randomized Comparison of Radio-frequency Thermal Ablation versus Percutaneous Ethanol Injection1 PURPOSE: To compare the effectiveness of radio-frequency (RF) thermal ablation with that of percutaneous ethanol injection (PEI) for the treatment of small hepatocellular carcinoma (HCC) in patients with cirrhosis. MATERIALS AND METHODS: A series of 102 patients with hepatic cirrhosis and either single HCC 5 cm in diameter or smaller or as many as three HCCs each 3 cm or smaller (overall number of lesions, 142) randomly received either RF ablation (n ⫽ 52) or PEI (n ⫽ 50) as the sole first-line anticancer treatment. Mean follow-up was 22.9 months ⫾ 9.4 (SD) in the RF group and 22.4 months ⫾ 8.6 in the PEI group. Prognostic value of treatment techniques was assessed with univariate and multivariate Cox proportional hazards regression models. RESULTS: One- and 2-year survival rates were 100% and 98% in the RF group and 96% and 88% in the PEI group, respectively (univariate relative risk [RR] ⫽ 0.20; 95% CI: 0.02, 1.69; P ⫽ .138). One- and 2-year local recurrence-free survival rates were 98% and 96% in the RF group and 83% and 62% in the PEI group, respectively (univariate RR ⫽ 0.17; 95% CI: 0.06, 0.51; P ⫽ .002). One- and 2-year event-free survival rates were 86% and 64% for the RF group and 77% and 43% for the PEI group, respectively (univariate RR ⫽ 0.48; 95% CI: 0.27, 0.85; P ⫽ .012). RF treatment was confirmed as an independent prognostic factor for local recurrencefree survival rates with multivariate analysis (adjusted RR ⫽ 0.20; 95% CI: 0.05, 0.73; P ⫽ .015). CONCLUSION: RF ablation is superior to PEI with respect to local recurrence-free survival rates. ©
RSNA, 2003
Patients with hepatic cirrhosis or chronic hepatitis are followed up with imaging techniques in combination with serum ␣1-fetoprotein assays in efforts to identify asymptomatic small hepatocellular carcinomas (HCCs) (1). Early detection programs have increased the number of patients suitable for surgical treatment. However, patients with HCC are often poor candidates for resection because of the lack of hepatic reserve as a result of coexisting cirrhosis or the presence of multiple tumors (2,3). Also, the shortage of donors limits the number of patients who can undergo orthotopic hepatic transplantation. Image-guided percutaneous ethanol injection (PEI) is an accepted treatment for patients with unresectable small HCC (1). Results in several articles (4 –7) show that PEI can cause complete necrosis of small HCCs and that the long-term survival rates in patients treated with PEI are similar to those in matched patients who underwent resection. Nevertheless, PEI has limitations, including uncertainty of tumor ablation, long treatment times because 235
Radiology
of the need of multiple sessions to achieve tumor destruction, and a local recurrence rate of 10%–30% (1). Recently, other local ablation methods to achieve more effective tumor necrosis were developed and clinically tested (8). Radio-frequency (RF) thermal ablation is a safe and effective treatment for hepatic malignancies (9 –12). In a pilot clinical study (13), the rate of complete tumor response with RF ablation was higher than that with PEI. One of the current recommendations for research in the field of HCC treatment is to compare PEI and RF ablation through randomized trials to assess not only initial tumor response but also long-term survival outcomes (1). Thus, the purpose of our study was to compare the effectiveness of RF thermal ablation with that of PEI for the treatment of small HCC in patients with cirrhosis.
MATERIALS AND METHODS Study Design and Enrollment Criteria The primary end point of the study was overall survival. Secondary end points were local recurrence-free survival and event-free survival (ie, survival free from local recurrence, new HCCs, and extrahepatic metastases). Requirements for entering the study were as follows: (a) adult patient with hepatic cirrhosis and either a single HCC 5 cm in diameter or smaller or as many as three HCCs each 3 cm in diameter or smaller, (b) HCCs located at least 1 cm away from the hepatic hilum or the gallbladder, (c) absence of vascular invasion or extrahepatic metastases, (d) hepatic cirrhosis classified as ChildPugh class A or B, (e) prothrombin time ratio (ie, normal time divided by patient’s time) greater than 50%, (f) platelet count higher than 50,000 per cubic millimeter (50 ⫻ 109/L), (g) no previous treatment for HCC, and (h) ineligibility for surgical resection or transplantation. The study was performed with the approval of the institutional ethics committees. Written informed consent was obtained from all patients after the nature of the procedures was fully explained.
Patients A series of 104 consecutive patients (67 men, 37 women; age range, 40 – 82 years; mean age, 68 years) with 144 HCCs who met the criteria were enrolled. The patients were assigned randomly to un236
䡠
Radiology
䡠
July 2003
dergo either RF ablation or PEI according to country by using a computer-generated randomization list that was not available to the treating physician. Fiftyfour patients with 71 HCCs formed the RF group, while 50 patients with 73 HCCs formed the PEI group. Two patients with solitary HCC who were assigned to the RF group had to be excluded from the study shortly after randomization, because in one patient the tumor diameter was larger than 5 cm and in the other patient extrahepatic tumor was identified retrospectively. Thus, 102 patients formed the study group. The number of HCCs and absence of vascular invasion were established on the basis of ultrasonographic (US), color Doppler US, and dual-phase spiral computed tomographic (CT) images. Each imaging study was performed at one of two authors’ institutions and interpreted by three authors (R.A.L., D.C., or H.P.A., J.L., H.E.B.) with consensus. On US scans, every discrete solid lesion was considered possible HCC. If the lesion detected at US clearly matched with a lesion showing characteristic features of HCC, including arterial hypervascularization, at dual-phase spiral CT, biopsy confirmation was considered unnecessary. According to current noninvasive criteria devised by the European Association for the Study of the Liver (1), HCC can be diagnosed in patients with hepatic cirrhosis by means of coincident findings with at least two modalities—US, CT, or magnetic resonance imaging—that show characteristic features in a focal lesion larger than 2 cm (1). Percutaneous US-guided biopsy was performed to determine the nature of lesions that could not be diagnosed with noninvasive criteria. Histologic confirmation of HCC was obtained in 45 of 102 patients, while a noninvasive diagnosis of HCC was accepted in the remaining 57, in whom coincident findings from US and dual-phase spiral CT consistently indicated HCC. The maximum dimension of HCCs was estimated with US. The absence of extrahepatic metastases was ascertained by means of full clinical and laboratory assessment, chest radiography, spiral CT of the abdomen, and bone scintigraphy performed by one of several authors (R.A.L., H.P.A., D.C., P.D., H.F., J.L., I.Z., H.E.B., C.B.) according to standard diagnostic criteria.
Treatment Methods RF thermal ablation and PEI were performed with real-time US guidance with
3.5-MHz convex probes (AU5, Esaote Biomedica, Genoa, Italy; or EUB 525, Hitachi Medical Systems, Wiesbaden, Germany). A dedicated probe for US-guided intervention or a guidance device incorporated into the US probe was used to place the RF electrode needle or the PEI needle. All procedures were performed by one of two authors (R.A.L., H.P.A.). RF thermal ablation was performed by using a 50-W RF generator (model 500 L; Rita Medical Systems, Mountain View, Calif) in monopolar mode coupled with an active 15-gauge expandable electrode needle with four retractable lateral exit curved electrodes on the tip and a large dispersive electrode. Power output, electrode temperature, tissue impedance, and procedure time were displayed continuously by the RF generator, which was connected to a portable computer. Dedicated software in this computer was used to record continuously the temperature curves obtained during treatment. Patients were placed in either the supine or left lateral decubitus position, depending on lesion site and planned needle track, and they were prepared and draped in the usual sterile fashion. A local anesthetic (Xylocaine 2%; Astra Zeneca, Milan, Italy) was injected from the insertion point on the skin to the peritoneum along the planned puncture track. The skin was pricked with a small lancet, and the needle was then advanced precisely to the chosen area of the HCC. After the electrodes had been deployed inside the HCC, the RF generator was activated. The power was set in the automatic mode to keep an average temperature of 95°C. The target temperature was usually reached within 2–3 minutes. In HCCs 3 cm or smaller in the greatest dimension, a single 8-minute time-at-temperature RF thermal ablation was performed. In 3.1–5.0-cm HCCs, multiple overlapping ablations—as many as six— were performed through one to three insertions by using a pullback technique. In these cases, the tip of the needle was placed initially in the deepest margin of the lesion, and the RF generator was activated. Then the electrodes were retracted, the needle was withdrawn about 1.5 cm, and the generator was reactivated. During RF treatment, a hyperechoic area was observed around the electrode tips; this area increased progressively to cover a larger area or all of the HCC. At the end of the procedure, after the automatic cooldown of the RF system, the generator was reactivated to ablate the needle tract to prevent tumor dissemination. PEI was performed after the patient reLencioni et al
Radiology
ceived the local anesthetic. For each patient, the treatment schedule involved four to eight PEI sessions performed once or twice weekly. The number of PEI sessions for each patient was decided on the basis of the number and size of the HCCs. During each PEI session, one or two ethanol injections were administered in each HCC; 2–10 mL of sterile 95% ethyl alcohol was injected slowly, depending on the size of the tumor and the distribution of the injected ethanol within the nodule. Alcohol distribution was evaluated with real-time US by observing the spread of the typical hyperechoic image. Injection was stopped when homogeneous alcohol perfusion of the lesion was accomplished or when ethanol leaks outside the nodule were observed repeatedly despite changing the site of alcohol injection. To prevent reflux of ethanol along the needle tract and leakage into the peritoneum, the needle was left in place for 60–120 seconds after completion of injection and then was withdrawn slowly. The needles used for PEI were either 22-gauge noncutting needles (spinal needle; Becton-Dickinson, Rutherford, NJ) or 21-gauge needles with a closed conical tip and multiple terminal side holes (Pflugbeil, Ottobrunn, Germany; and Ethanoject, TSK, Tokyo, Japan). One of several authors (D.C., P.D., L.C., H.F., I.Z.) carefully monitored the patient during and after each RF ablation or PEI procedure. Any adverse event was evaluated. Pain was graded as absent or as slight discomfort that did not require therapy, mild to moderate pain that required administration of analgesics, or severe pain that required sedation.
Posttreatment Assessment and Follow-up Short-term therapeutic effectiveness was assessed with dual-phase spiral CT (HiSpeed, GE Medical Systems, Milwaukee, Wis; or PQ6000, Picker International, Highland Heights, Ohio) performed 1 week after treatment. Dualphase spiral CT studies were optimized for arterial and portal venous phases of contrast material enhancement and were interpreted by three authors with consensus. Hypoattenuating nonenhancing areas in both the arterial and portal venous phases were considered to represent complete tumor necrosis. In contrast, any portion of the treated HCC that showed persistent intranodular enhancement was considered to represent residual viable neoplastic tissue. In cases of residual viable neoplastic Volume 228
䡠
Number 1
tissue, additional treatment with RF thermal ablation or PEI was scheduled. However, if residual viable neoplastic tissue persisted after repeated RF ablation or PEI, the HCC was considered unresponsive to percutaneous treatment, and further treatment with segmental transcatheter arterial chemoembolization was scheduled. Follow-up was computed as starting from the beginning of treatment for all patients. The follow-up protocol included measurement of ␣1-fetoprotein levels and US performed at 3-month intervals and dual-phase spiral CT performed at 6-month intervals. Each imaging study was interpreted by three authors with consensus. Patients were observed for recurrence within or around the treated tumor (ie, local recurrence) and for the emergence of new tumors remote from the treated lesion. All new HCCs that emerged during follow-up were treated with the same percutaneous technique (RF thermal ablation or PEI) if the patient still met the requirements for these therapies. If multicentric HCC developed, patients underwent transcatheter arterial chemoembolization or other medical treatment, depending on the kind of tumor involvement, the degree of hepatic function, and the overall clinical condition. Mean follow-up was 22.9 months ⫾ 9.4 (SD) in the RF group and 22.4 months ⫾ 8.6 in the PEI group.
Statistical Analysis Baseline data were presented as mean values plus or minus SD for quantitative variables, except for the skewed ␣1-fetoprotein levels for which the median and interquartile range were used, and as absolute and relative frequencies for qualitative variables. Differences between the RF and PEI groups in baseline characteristics were analyzed by means of the Wilcoxon rank sum test for quantitative variables and the Fisher exact test for qualitative variables. All survival probabilities were estimated by means of the Kaplan-Meier method. For overall survival rates, the time from the beginning of RF ablation or PEI treatment to last follow-up US, dual-phase spiral CT, or death was used. For local recurrence-free survival rates, the time to local recurrence or death was used. For event-free survival rates, the time to local recurrence, extrahepatic metastases, new HCCs, or death was used. Prognostic value of treatment techniques and of baseline characteristics was assessed by means of both univariate and
multivariate Cox proportional hazards regression models. Variables with a priori selected cutoff values—a P value less than or equal to .1 in a univariate comparison in at least one of the three end points— were entered into the multivariate analysis. Both univariate and multivariate results were presented as relative risks (RRs) with corresponding 95% CIs with P values from the Wald test. All significance tests were two-sided, and differences with a P value less than .05 were considered statistically significant. Data processing and analysis were performed with commercially available software (Statistical Analysis System; SAS Institute, Cary, NC).
RESULTS No statistically significant differences between the RF and PEI groups were observed with respect to baseline characteristics, except for patient age and albumin concentration (Table 1). One PEI cycle led to complete tumor response in 60 (82%) of 73 HCCs. During the first treatment cycle, an average of 5.4 PEI sessions ⫾ 1.6 was performed, with a mean total amount of alcohol injected per session of 20.6 mL ⫾ 14 (range, 4.5–75.0 mL) per lesion. In 13 lesions, a second PEI cycle was performed. After the second PEI cycle, tumor still persisted in four lesions. The median preinterventional ␣1-fetoprotein level was 54 g/L (interquartile range, 8 –143). One month after intervention, the median ␣1-fetoprotein level was 36 g/L (interquartile range, 7–102), and after 12 months it was 42 g/L (interquartile range, 10 –100). A single RF ablation session led to complete tumor response in 63 (91%) of 69 HCCs. An average of 1.1 sessions ⫾ 0.5 was performed, with a mean coagulation time of 14.1 minutes ⫾ 8.2 (range, 8 – 48 minutes) per lesion. In six lesions, a second RF ablation session was performed. After that second intervention, tumor still persisted in one lesion. The median preinterventional ␣1-fetoprotein level was 27 g/L (interquartile range, 11–140 g/ L). After 1 month, the median ␣1-fetoprotein level was 20 g/L (interquartile range, 11–103 g/L), and after 12 months it was 17 g/L (interquartile range, 8 – 88 g/L). Thirteen (26%) of 50 patients experienced mild to moderate pain during and/or immediately after PEI and required administration of analgesics. Fever (temperature higher than 38.5°C) was observed in five patients after PEI.
Hepatocellular Carcinoma in Cirrhosis: Thermal Ablation versus Ethanol Injection
䡠
237
Radiology
In the RF group, 15 (29%) of 52 patients experienced mild to moderate pain during the procedure; the pain was relieved with administration of analgesics in all patients. No RF ablation had to be interrupted. Postinterventional fever was observed in 10 patients. RF ablation of four lesions localized subdiaphragmatically resulted in pleural effusions that resolved spontaneously. In three patients, spiral CT revealed clinically asymptomatic arteriovenous shunts after RF ablation. One patient developed chemical thrombosis of a portal venous branch after PEI. In both groups, no procedure-related death, hemorrhage, infection, needle-track seeding, or hepatic failure was observed. During follow-up, five patients in the PEI group died. Four patients died of hepatic failure—two because of tumor progression, one because of advanced cirrhosis, and one because of continuous alcohol abuse—and one patient died of variceal bleeding after malignant portal venous infiltration. In the RF group, one patient died of renal failure that developed after surgery for a humerus fracture. Total survival rates after 1 and 2 years were 96% (95% CI: 90%, 100%) and 88% (95% CI: 78%, 98%) in the PEI group and 100% (95% CI not applicable) and 98% (95% CI: 93%, 100%) in the RF group (RR ⫽ 0.20; 95% CI: 0.02, 1.69; P ⫽ .138). Thirteen patients in the PEI group developed a local recurrence compared with recurrence in three patients in the RF group. Local recurrence–free survival rates were 83% (95% CI: 73%, 94%) at 1 year and 62% (95% CI: 46%, 77%) at 2 years in the PEI group compared with 98% (95% CI: 94%, 100%) at 1 year and 96% (95% CI: 90%, 100%) at 2 years in the RF group (RR ⫽ 0.17; 95% CI: 0.06, 0.51; P ⫽ .002) (Fig 1). New tumors occurred in 11 patients in the PEI group and 13 patients in the RF group. Multicentric HCCs were observed in seven patients in the PEI group and in five in the RF group. Distant metastases were not observed in either treatment group. Portal venous thrombosis was observed in four patients—three with tumor infiltration and one with chemical thrombosis—in the PEI group and in none in the RF group. In summary, 1- and 2-year event-free survival rates were 77% (95% CI: 65%, 89%) and 43% (95% CI: 27%, 60%) in the PEI group and 86% (95% CI: 77%, 96%) and 64% (95% CI: 49%, 79%) in the RF group (RR ⫽ 0.48; 95% CI: 0.27, 0.85; P ⫽ .012) (Fig 2). To investigate whether the treatment 238
䡠
Radiology
䡠
July 2003
TABLE 1 Baseline Clinical Characteristics of Patients Characteristic Age (y) Mean Range Sex Male Female Cause of cirrhosis Hepatitis B Hepatitis C Hepatitis B and C Alcohol use Other Liver cirrhosis Child-Pugh classification A B C Albumin Mean level (g/L) Concentration ⬎ 35 g/L Bilirubin Mean level (mol/L) Concentration ⱕ 34.2 mol/L Prothrombin time ratio ␣1-Fetoprotein (g/L) Median level Interquartile range No. of tumor lesions One Two Three Tumor diameter* Mean (cm) Tumor diameter ⱕ 3 cm
PEI Group (n ⫽ 50)
RF Group (n ⫽ 52)
69 ⫾ 7.4 40–82
67 ⫾ 6.0 52–78
30 (60) 20 (40)
36 (69) 16 (31)
9 (18) 20 (40) 12 (24) 7 (14) 2 (4)
6 (12) 22 (42) 13 (25) 8 (15) 3 (6)
35 (70) 15 (30) 0 (0)
45 (87) 7 (13) 0 (0)
37 ⫾ 6 30 (60)
41 ⫾ 7 42 (81)
26 ⫾ 26 40 (80) 80% ⫾ 14.6
19 ⫾ 8 48 (92) 84% ⫾ 13.2
54 8–143
27 11–140
31 (62) 15 (30) 4 (8)
40 (77) 7 (13) 5 (10)
2.8 ⫾ 0.8 42 (84)
2.8 ⫾ 0.6 46 (88)
Note.—Data are the number of patients, unless indicated otherwise. Numbers in parentheses are percentages. Differences between the two groups were not significant, except for patient age and albumin concentration. * One lesion or the largest diameter of the main tumor in patients with two or three lesions.
effects were independent and whether other variables were prognostically relevant, we performed multivariate analyses. In addition to treatment, clinical characteristics with a P value of less than or equal to .1 were entered in a univariate test of at least one of the three end points considered. No variable showed prognostic relevance with respect to total survival rates. With respect to local recurrence–free survival rates, the superiority of RF treatment was confirmed as independent (adjusted RR ⫽ 0.20; 95% CI: 0.05, 0.73; P ⫽ .015) (Table 2). The other independent prognostic factor was tumor diameter, with tumors larger than 3 cm having an increased risk of local recurrence and/or death (adjusted RR ⫽ 8.67; 95% CI: 1.90, 39.66; P ⫽ .005). With respect to event-free survival rates, RF ablation was of no independent prognostic value (adjusted RR ⫽ 0.58; 95% CI: 0.28, 1.22; P ⫽ .15). Here, however, tumor diameter larger than 3 cm
Figure 1. Graph shows the probability of local recurrence–free survival in patients with HCC treated with PEI (n ⫽ 50) or RF thermal ablation (RFTA, n ⫽ 52). The difference between the groups was statistically significant (P ⫽ .002). The number of patients followed up at 6, 12, 18, 24, and 30 months was 46, 37, 24, 16, and five, respectively, for the PEI group and 49, 46, 33, 24, and 10, respectively, for the RF group.
(adjusted RR ⫽ 7.21; 95% CI: 2.26, 22.97; P ⬍ .001) and a bilirubin level exceeding 2 mg/dL (34.2 mol/L) (adjusted RR ⫽ Lencioni et al
TABLE 2 Factors to Predict Local Recurrence-free Survival and Event-free Survival
Radiology
Local Recurrence-free Survival Univariate Analysis
Event-free Survival
Multivariate Analysis
P Value
Univariate Analysis
Variable
RR
RR
P Value
RR
RF ablation treatment Bilirubin ⬎ 34.2 mol/L Albumin ⱕ 35 g/L Age ⱖ 70 y More than one HCC lesion HCC diameter ⬎ 3 cm Alcohol use
0.17 (0.06, 0.51) 3.81 (1.42, 10.20) 2.11 (0.86, 5.17) 1.96 (0.82, 4.66)
.002 .008 .10 .13
0.20 (0.05, 0.73) 3.31 (0.94, 11.67) 0.59 (0.18, 1.91) 1.02 (0.35, 2.93)
.015 .06 .38 .97
0.48 (0.27, 0.85) 2.77 (1.35, 5.69) 1.82 (1.02, 3.27) 1.32 (0.76, 2.31)
2.53 (1.07, 5.96) 8.38 (2.90, 24.27) 2.72 (0.98, 7.55)
.03 ⬍.001 .06
2.52 (0.97, 6.51) 8.67 (1.90, 39.66) 1.58 (0.33, 7.59)
.06 .005 .57
1.71 (0.97, 3.00) 5.03 (2.05, 12.32) 1.22 (0.51, 2.90)
Multivariate Analysis
P Value .012 .005 .04 .32 .06 ⬍.001 .66
RR 0.58 (0.28, 1.22) 2.35 (1.02, 5.43) 0.87 (0.42, 1.79) 0.97 (0.49, 1.91) 1.54 (0.84, 2.84) 7.21 (2.26, 22.97) 0.70 (0.23, 2.09)
P Value .15 .046 .71 .93 .17 ⬍.001 .52
Note.—Data in parentheses are the 95% CI.
Figure 2. Graph shows the probability of event-free survival in patients with HCC treated with PEI (n ⫽ 50) or RF thermal ablation (RFTA, n ⫽ 52). The difference between the groups was statistically significant (P ⫽ .012). The number of patients followed up at 6, 12, 18, 24, and 30 months was 45, 34, 21, 12, and two, respectively, for the PEI group and 49, 42, 30, 20, and eight, respectively, for the RF group.
2.35; 95% CI: 1.02, 5.43; P ⫽ .046) (Table 2) correlated with an unfavorable clinical course.
DISCUSSION There is no universal algorithm implemented worldwide for the treatment of HCC in patients with cirrhosis. If HCC in cirrhosis is diagnosed at an early stage, patients should be considered for any of the available treatment options that may provide a high rate of complete response. These options include surgical resection, hepatic transplantation, and percutaneous tumor ablation techniques (1). According to the guidelines of the European Association for the Study of the Liver (1), PEI should be considered the standard technique for percutaneous treatment, and newer methods of tissue ablation, such as RF ablation, should be compared Volume 228
䡠
Number 1
with PEI in randomized trials to assess long-term survival rates. In the current prospective randomized study, we compared PEI with RF ablation for the treatment of small HCC in patients with hepatic cirrhosis. Complete tumor response was achieved in 91% of HCCs treated with RF thermal ablation, with an average of 1.1 treatment sessions ⫾ 0.5, and in 82% of HCCs treated with PEI, with an average of 5.4 treatment sessions ⫾ 1.6. Our data can be used to confirm findings in a previous pilot study (13), in which RF thermal ablation was shown to be superior to PEI in achieving complete tumor response and requiring fewer treatment sessions. No serious side effects or procedurerelated complications (eg, hemorrhage, infection, needle tract seeding, hepatic failure, or death) occurred after RF thermal ablation or PEI. In our series, pain and fever were the most common side effects in both groups, and these symptoms were treated; no intervention had to be stopped. In one patient undergoing PEI, a chemical portal venous thrombosis developed, a complication which was reported (14). After RF thermal ablation, four patients with subdiaphragmatic tumors developed pleural effusions that resolved spontaneously. Iatrogenic tumor dissemination after RF ablation of subcapsular HCCs or HCCs with invasive tumor pattern was reported (15). In our series, we did not observe any case of tumor seeding after RF ablation during long-term follow-up. Primarily because of the short observation period and low number of deaths in both treatment groups, we were not able to demonstrate a statistically significant difference between RF thermal ablation and PEI with respect to survival rates. Nevertheless, a trend toward increased
survival in the RF thermal ablation group is obvious. A trend in the extent of cirrhosis in favor of the RF thermal ablation group would have contributed to better survival; however, no clinical variable showed prognostic relevance with respect to survival rates in multivariate analyses. Our study results demonstrate that there was local recurrence of small HCCs treated with either method. With respect to local recurrence-free survival, however, RF thermal ablation was clearly superior to PEI according to results of multivariate analysis. This finding reflects the superior local ablation effect of RF thermal ablation compared with that of PEI. PEI is associated with a local recurrence rate as high as 33% if careful long-term follow-up is performed (16). Despite the absence of viable neoplastic tissue on posttreatment spiral CT images, residual microscopic nests of tumor led to local recurrence in some patients. In the multivariate analysis, lesions more than 3 cm in diameter correlated with a higher probability of local recurrence, which probably reflects a higher rate of extracapsular penetration and microscopic vascular invasion of larger tumors. With respect to event-free survival, RF thermal ablation was of no independent prognostic value in the multivariate analysis. Tumor diameter greater than 3 cm and a bilirubin level higher than 2 mg/dL (34.2 mol/L), however, could be used to predict a poor outcome. The development of new HCCs was similar in both treatment groups, as expected, because the local treatment strategies did not affect the underlying risks of HCC development. The RF generator used in this study has been replaced by newer devices that use higher power output and novel electrode
Hepatocellular Carcinoma in Cirrhosis: Thermal Ablation versus Ethanol Injection
䡠
239
Radiology
needles. Results of several studies (17–19) show that current RF ablation technology may achieve larger volumes of coagulation necrosis. Also, refined techniques for modifying tumor tissue response to RF ablation and combined treatment strategies result in better local tumor response rates (20 –23). Therefore, the results of RF ablation might be underestimated in our study. Nevertheless, patients treated with RF ablation had significantly higher local recurrence–free and event-free survival rates than did patients treated with PEI. Our study results show that RF thermal ablation is more effective than PEI in the treatment of small HCC in patients with cirrhosis. Therefore, RF ablation should be considered the percutaneous treatment of choice for patients who are not candidates for resection or transplantation. Further investigation is warranted to clarify whether current RF ablation technology could offer improved longterm results in patients with more advanced tumors. References 1. Bruix J, Sherman M, Llovet JM, et al, for the EASL Panel of Experts on HCC. Clinical management of hepatocellular carcinoma: conclusions of the Barcelona-2000 EASL conference. J Hepatol 2001; 35:421– 430. 2. Colombo M, Sangiovanni A. The European approach to hepatocellular carcinoma. Hepatogastroenterology 2002; 49: 12–16. 3. Bruix J, Llovet JM. Prognostic prediction and treatment strategy in hepatocellular carcinoma. Hepatology 2002; 35:519 –524. 4. Shiina S, Tagawa K, Niwa Y, et al. Percutaneous ethanol injection therapy for hepatocellular carcinoma: results in 146
240
䡠
Radiology
䡠
July 2003
5.
6.
7.
8.
9.
10. 11.
12.
13.
14.
patients. AJR Am J Roentgenol 1993; 160: 1023–1028. Kotoh K, Sakai H, Sakamoto S, et al. The effect of percutaneous ethanol injection therapy on small solitary hepatocellular carcinoma is comparable to that of hepatectomy. Am J Gastroenterol 1994; 89: 194 –198. Lencioni R, Bartolozzi C, Caramella D, et al. Treatment of small hepatocellular carcinoma with percutaneous ethanol injection: analysis of prognostic factors in 105 Western patients. Cancer 1995; 76:1737– 1746. Livraghi T, Giorgio A, Marin G, et al. Hepatocellular carcinoma and cirrhosis in 746 patients: long-term results of percutaneous ethanol injection. Radiology 1995; 197:101–108. Dodd GD III, Soulen MC, Kane RA, et al. Minimally invasive treatment of malignant hepatic tumors: at the threshold of a major breakthrough. RadioGraphics 2000; 20:9 –27. Goldberg SN, Gazelle GS, Mueller PR. Thermal ablation therapy for focal malignancies: a unfied approach to underlyng principles, techniques, and diagnostic imaging guidance. AJR Am J Roentgenol 2000; 174:323–331. McGahan JP, Dodd GD III. Radiofrequency ablation of the liver: current status. AJR Am J Roentgenol 2000; 176:3–16. Goldberg SN, Gazelle GS, Compton CC, Mueller PR, Tanabe KK. Treatment of intrahepatic malignancy with radiofrequency ablation. Cancer 2000; 88:2452– 2463. Lencioni R, Cioni D, Bartolozzi C. Percutaneous radiofrequency thermal ablation of liver malignancies: techniques, indications, imaging findings, and clinical results. Abdom Imaging 2001; 26:345–360. Livraghi T, Goldberg SN, Lazzaroni S, et al. Small hepatocellular carcinoma: treatment with radio-frequency ablation versus ethanol injection. Radiology 1999; 210:655– 661. Lencioni R, Caramella D, Sanguinetti F, Battolla L, Falaschi F, Bartolozzi C. Portal vein thrombosis after percutaneous etha-
15.
16.
17.
18.
19.
20.
21.
22.
23.
nol injection for hepatocellular carcinoma: value of color Doppler sonography in distinguishing chemical and tumor thrombi. AJR Am J Roentgenol 1995; 164: 1125–1130. Llovet JM, Vilana R, Bru C, et al. Increased risk of tumor seeding after percutaneous radiofrequency ablation for single hepatocellular carcinoma. Hepatology 2001; 33: 1124 –1129. Koda M, Murawaki Y, Mitsuda A, et al. Predictive factors for intrahepatic recurrence after percutaneous ethanol injection therapy for small hepatocellular carcinoma. Cancer 2000; 88:529 –537. Goldberg SN, Solbiati L, Hahn PF, et al. Large volume tissue ablation with radiofrequency by using a clustered, internally cooled electrode technique: laboratory and clinical experience in liver metastases. Radiology 1998; 209:371–379. Goldberg SN, Stein MC, Gazelle GS, Sheiman RG, Kruskai JB, Clouse ME. Percutaneous radiofrequency tissue ablation: optimization of pulsed-technique to increase coagulation necrosis. J Vasc Interv Radiol 1999; 10:907–916. Livraghi T, Goldberg SN, Lazzaroni S, et al. Hepatocellular carcinoma: radio-frequency ablation of medium and large lesions. Radiology 2000; 214:761–768. Goldberg SN, Hahn PF, Halpern EF, Fogle RM, Gazelle GS. Radio-frequency tissue ablation: effect of pharmacologic modulation of blood flow on coagulation diameter. Radiology 1998; 209:761–767. Goldberg SN, Hahn PF, Tanabe KK, et al. Percutaneous radio-frequency tissue ablation: does perfusion-mediated tissue cooling limit coagulation necrosis? J Vasc Interv Radiol 1998; 9:101–111. Rossi S, Garbagnati F, Lencioni R, et al. Percutaneous radio-frequency thermal ablation of nonresectable hepatocellular carcinoma after occlusion of tumor blood supply. Radiology 2000; 217: 119 –126. Goldberg SN. Comparison of techniques for image-guided ablation of focal liver tumors. Radiology 2002; 223:304 –307.
Lencioni et al