Radiofrequency Ablation of the Porcine LiverIn Vivo: Increased Coagulation with an Internally

Cooled Perfusion Electrode1

Jeong Min Lee, Joon Koo Han, Jung Min Chang, Se Young Chung, Se Hyung Kim, Jae Young LeeMin Woo Lee, Byung Ihn Choi

Rationale and Objectives. A major limitation of radiofrequency (RF) ablation is its inability to produce a large enoughdiameter of coagulation necrosis to encompass hepatic tumors with an appropriate ablative margin at a single RF applica-tion. We evaluated the in vivo efficiency of RF ablation (RFA) using an internally cooled perfusion (ICP) electrode withhypertonic saline infusion to induce coagulation necrosis compared with that of RFA using single needle electrode types.

Materials and Methods. RF was applied to a porcine liver in monopolar mode using a 200 W generator and an internallycooled electrode (group A) or an ICP electrode (group B) at 200 W for 12 minutes or using a 60 W generator with a per-fusion electrode at 40 W for 20 minutes (group C). In total, 36 (3 � 12) ablation zones were created using the three dif-ferent regimens. In group B, 14.6% NaCl solution was infused at 1 mL/minute and in group C, 0.9% NaCl solution wasinfused at 1.5 mL/minute. The three groups were compared in terms of amount of delivered RF energy and dimensionsand the coefficients of variation of the ablation zones.

Results. The mean energies applied in the three groups were 52.3 � 10.3 kJ for group A, 115.4 � 10.5 kJ for group B,and 38.5 � 11.5 kJ for group C, respectively (P � .05). The mean ablation volumes in groups A, B and C were 13.1 �4.7 cm3 in group A, 43.7 � 17.5 cm3 in group B, and 26.3 � 20.2 cm3 in group C, respectively (P � .05). In addition,the coefficients of variation of the volumes of the ablation zones in groups A, B, and C were 0.36, 0.4, and 0.78, respec-tively.

Conclusions. RFA using the ICP electrode showed better performance in terms of creating a larger ablation zone thanRFA using an internally cooled or a perfusion electrode.

Key Words. Experimental study; interventional procedures; liver; radiofrequency ablation.© AUR, 2006

In recent years, tumor ablation using radiofrequency (RF)energy has been increasingly used as an alternative formof minimally invasive therapy for the treatment of pri-

Acad Radiol 2006; 13:343–352

1 From the Department of Radiology, and Institute of Radiation Medicine,28 Yongon-dong, Chongno-gu, Seoul 110-744, Korea (J.M.L., J.K.H.,J.M.C., S.Y.C., S.H.K., J.Y.L., M.W.L., B.I.C.), and Clinical Research Insti-tute, Seoul National University Hospital, Seoul, Korea (J.M.L., J.K.H.,S.H.K., J.Y.L., M.W.L., B.I.C.). Received August 24, 2005; accepted Octo-ber 14, 2005. Address correspondence to: J.K.H. e-mail: hanjk@radcom.snu.ac.kr

©

AUR, 2006doi:10.1016/j.acra.2005.10.020

mary and secondary hepatic malignancies (1–4). Themost commonly used RF technique is the dry monopolarRF ablation (RFA) (5–8). However, one limitation ofRFA is its inability to produce a large enough diameter ofcoagulation necrosis to encompass hepatic tumors with anappropriate ablative margin at a single RF application(9–11). This limitation is attributed to too great a currentdensity at the metal electrode-tissue interface from thelow electrical conductivity of tissue, which leads to tissuedesiccation and carbonization. Consequently, a significantincrease in impedance occurs and further RF current is

delivered (7,12). As a strategy to avoid the carbonization

343

LEE ET AL Academic Radiology, Vol 13, No 3, March 2006

of tissues adjacent to the electrode, RF devices use salineperfusion electrode (Berchtold RF System; Tuttlingen,Germany) or some other form of electrode internal cool-ing (Cooled Tip Electrode; Valleylab RF System, Burl-ington, MA) (13–15). However, these techniques are lim-ited in terms of preventing overheating in the area aroundthe electrode. In clinical practice, the treatment of livertumors larger than 4 cm often requires multiple overlap-ping ablations (16,17). However, because this approach isboth time consuming and technically challenging, there isa need to increase coagulation dimensions using a singleRF application (18,19).

We recently designed and tested ex vivo a prototypeelectrode that simultaneously allows the interstitial infu-sion of saline and intra-electrode cooling (20,21). A pre-vious ex vivo study (21) demonstrated that wet RFA us-ing the prototype internally cooled perfusion (ICP) elec-trode created a larger ablation zone in excised bovineliver than did dry RFA using an internally cooled elec-trode. Therefore, the purpose of this study was to confirmthat our developed ICP electrode could administer in anin vivo pig liver model, a larger region of coagulationnecrosis than two other types of available single needleelectrodes.

MATERIALS AND METHODS

Animals, Anesthetics, and Surgical TechniqueThe experimental protocol was approved by the Ani-

mal Use and Care Administrative Advisory Committee ofour institution. All experiments were performed accordingto a protocol approved by the local institutional animalcare committee and were in accord with the generalguidelines issued by the National Institute of Health forthe care of laboratory animals. Six female farm pigs wereused in this study (weight range 30–40 kg). The animalswere fasted overnight but had free access to water beforethe experiments.

Each of the six pigs was anesthetized using an intra-muscular injection of 50 mg/kg of ketamine hydrochloride(Ketamine; Yuhan, Seoul, Korea) and 5 mg/kg of xyla-zine (Rumpun; Bayer Korea, Ansan, Korea) and preparedfor RF ablation. Booster injections of up to one half ofthe initial dose were administered as needed. Ringer lac-tate solution was continuously infused during the experi-ment (300 mL/hour). Endotracheal intubation was per-formed and anesthesia was maintained using inhaled en-

flurane (Gerolan; Choongwae Pharma Corporation, Seoul,

344

Korea). Mechanical ventilation was used throughout theprocedure, and cardiac and respiratory parameters weremonitored throughout the entire procedure period. Eachanimal’s lateral hindquarters was shaved bilaterally, andtwo 8 � 12 cm wire-mesh grounding pads coated withconductive gel were placed on each hind limb.

The surgical procedure involved placing a pig in thesupine position. Via a midline incision, the liver was dis-sected free to expose the lower two thirds. Because ofdifficult access to a computed tomography machine at ourinstitute, we performed the RFA procedure using a lapa-rotomy instead of under computed tomography guidance.After the RFA procedure, the incision was closed usingnonabsorbable sutures.

RF Devices, Ablation Protocols, and Procedures

Design of the ICP electrode.—Based on previous stud-ies (20,21), we developed a coaxial electrode system(Taewoong, Seoul, Korea.) consisting of a 17-guage inter-nally cooled electrode with a 14-gauge outer insulatingsheath (Fig 1). The outer metal sheath was insulated withTeflon except for the distal 3 cm, and covered the innerelectrode. The space between the 14-gauge sheath and theinternally cooled electrode tip permitted saline infusionalong the electrode. A side hole for saline infusion intothe coaxial electrode region was positioned proximally inthe sheath. In addition, 20 � 0.5-mm diameter side holes

Figure 1. Photograph of a prototype “internally cooled perfu-sion” electrode. The electrode has a coaxial design consisting ofa 17-guage internally cooled electrode and a 14-gauge outer in-sulated sheath with a 3-cm long distal active portion containing20 holes. A side hole for saline infusion into the coaxial electroderegion was positioned proximally in the sheath (arrow).

were positioned in the distal active tip.

Academic Radiology, Vol 13, No 3, March 2006 RF ABLATION USING AN ICP ELECTRODE

Study design.—To minimize potential variations in theRF ablation procedures, all procedures were performed byconsensus by two radiologists with extensive experienceof all three systems on a routine basis. RF ablation wasperformed using one of the three RF systems: group A: a200-W generator (Valleylab) and an internally cooledelectrode (Cool tip; Valleylab); group B: a 200-W genera-tor (Valleylab) and an ICP electrode; and group C: a60-W generator (HiTT 106; Berchtold) and a perfusionelectrode. In these three groups, an electrode was insertedinto the liver to a depth of 35 mm under intraoperativeultrasound guidance (Sonoace 5500; Medison, SeoulKorea). The RF system and the liver lobe, as well as theposition and order in which each ablation procedure wasperformed, were randomly assigned. A new electrode wasused for each ablation session. All settings used were inline with the manufacturer’s recommendations. The auto-mated control mechanisms of the RF systems are basedon measuring total impedance between electrodes. Theapplied current, power output, and impedance were con-tinuously monitored by the generator system during RFAand were recorded. To allow the continuous measurementof temperature in the liver during the procedure, a ther-mocouple was inserted into the liver at a position 15 mmfrom the electrode. To ensure that the thermocouple was15 mm from the electrode, the electrode and the thermo-couple were placed in the liver though an acryl plate withmultiple holes at 5-mm intervals. In addition, to avoid theeffect from blood flow in the measurement of tissue tem-perature, both the thermocouple and electrode wereplaced under ultrasound guidance where portal vein orhepatic vein larger than 5 mm was absent. To decreasethe risk of bleeding related to the placement of multiplethermocouples, only one thermocouple was used for tem-perature monitoring. The technical aspects of the RFAincluding impedance and wattage changes and the dimen-sions of the RF-coagulated area for each system werecompared.

RFA Protocols.—In groups A and B, which used thesame 480-kHz generator capable of producing a maxi-mum power of 200 W (Valleylab), the RF power wasmanually increased to a maximum of 200 W and held fora total of 12 minutes. The 17-gauge, internally cooled,and 14-gauge ICP electrodes have a 3-cm long active dis-tal part, respectively. The circuitry incorporated into thegenerator allows continuous monitoring of impedancebetween the active portion of the electrode and thegrounding pads. Electrode cooling was achieved by the

perfusion of chilled saline driven by a peristaltic pump

(PE-PM; Valleylab), which allowed the electrode to main-tain tip temperatures below 25°C during RF delivery. Ingroup B, RFA was performed using 14.6% hypertonicsaline (HS) infusion at a rate of 1 mL/minute. Accordingto our preliminary study results (unpublished data) regard-ing the optimal infusion rates and HS concentrations forHS-enhanced RFA, which demonstrated that 14.6% HS at1 mL/minute created a larger coagulation area comparedwith 0.9%, 5%, and 36% HS infusion at 1–2 mL/minute,14.6% HS was infused at a rate of 1 mL/minute throughan ICP electrode using an infusion pump (Pilotec IS;Fresenius Medical Care, Alzenau, Germany). The reasonwe did not use infusion of 36% NaCl was that infusing36% NaCl solution during RFA (12 minutes) induced agradual decrease of tissue impedance lower than 50 �.This was not helpful to increase RF-induced coagulationnecrosis because too much decease of tissue impedanceinduces a remarkable increase of current needed to createheat. Furthermore, infusion of 36% NaCl solution at alow infusion rate (0.5 mL/minute) caused frequent block-age of the saline perfusion holes because of evaporationof liquid in the active tip portion of the electrode. RFcurrent was emitted for 12 minutes with the generator setto deliver the 200 W in the impedance control mode. Thismode allows maximum power to be delivered until theimpedance increases to 10 � above the baseline value. Atthis point, the current is switched off for 15 seconds auto-matically to avoid further local temperature increase;thereafter, it is switched on again (22).

In group C, a 375-kHz generator (HiTT 106; Berch-told) capable of producing a maximum power of 60 Wthrough a 1.7-mm diameter monopolar electrode with anactive needle tip length of 1.5 cm was used. The elec-trode for this saline-enhanced technique was doublewalled at its distal part, whereas the inner wall had smallholes. A 0.9% saline solution was used as a perfusionliquid; this flowed through the hollow shaft of the elec-trode and permeated through the holes into the liver tis-sue. Continuous interstitial perfusion of saline was started30 seconds at 90 mL/hour before RF application andmaintained during RFA using a digitally controlled sy-ringe pump (Pilot C; Frensenius). RF current was appliedfor 20 minutes at 40 W, in accordance with the manufac-turer’s recommendations. The control mechanism stabi-lizes the RF power within moderate impedance changesbetween 100 � and 350 �. If the impedance exceeded900 � an additional saline bolus was given (23).

In all three groups, grounding for the RF procedure

was done via two externally (dorsally) attached grounding

345

LEE ET AL Academic Radiology, Vol 13, No 3, March 2006

pads, such that the RF currents were distributed evenlythrough tissue in the direction of the grounding pads.Tissue impedance was monitored using the circuitry in-corporated into the generator. Applied current, outputpower, and impedance were continuously monitored dur-ing RFA. In groups A and B, the data were recorded au-tomatically using a computer program (Real Time Graph-ics Software V 2.0; Valleylab).

Evaluation of the Ablation ZoneThe pigs were followed and euthanized 4 days after

the procedure. Livers containing lesions were sliced in thetransverse plane at 5-mm intervals. Specimens werestained for mitochondrial enzyme activity by incubatingthem for 30 minutes in 2% 2, 3, 5,-triphenyl tetrazoliumchloride, or TTC (Sigma, St Louis, MO), at 20–25°C.This test is used to determine irreversible cellular injuryduring the early stages of RF-induced necrosis (24). Twoobservers using a caliper measured the two transversediameters (Dtr1, Dtr2) of the central, white region of theRF-induced ablation zones in slices showing maximalcoagulation diameter. The transverse diameters were de-fined as the distances between the two opposite edges ofthe coagulation perpendicular to the electrode shaft at theequator of the coagulation. The axial diameter (Dax) alongthe electrode insertion axis was measured by the determin-ing the number of slices showing an unstained area.

In addition, the slices were photographed using a digi-tal camera (Canon EOS 300D; Canon Inc, Tokyo, Japan),and images were saved to image management software(PhotoShop; Adobe, San Jose, CA). Area analysis wasperformed on a computer equipped with NIH Image Jsoftware (National Institutes of Health; http://rsb.info.nih.gov/ij/) (25). The area of coagulation, on each slicewas calculated using this computer program, and volumeswere calculated by multiplying areas by slice thicknessand summed to obtain total lesion volumes.

Ablation zonal shapes were evaluated using a roughestimate of lesion “roundness” in two dimensions bycomputing the isoperimetric ratio for each lesion in themost representative slice (26). This ratio is computed us-ing the following formula: R � 4�A/l2, where R is theisoperimetric ratio, A is the area of the measured zone,and l the perimeter of the lesion; the closer this value isto 1, the more circular the lesion shape (26,27). Valuesfor A and l were obtained using the computer programNIH image J (25). In addition, the shape of the ablationareas was defined by dividing Dax by the average of Dtr1

and Dtr2 (Dtrm). Thus a ratio near 1 implies a more

346

spherical ablated region shape. Volume variations in theindividual group were determine by calculating variationcoefficients, as follows: standard deviation of the ablationvolume/mean value of the ablation volume.

The ablated regions in the six animals were fixed in10% formalin for routine histologic processing and finallyprocessed by paraffin sectioning and hematoxylin-eosinstaining for light microscopic study.

Statistical AnalysisOne-way analysis of variance with the Turkey test

(P � .05, two-tailed test) was performed to compare themean diameters of the thermal ablation areas, and themean values of R (isoperimetric ratio) of the threegroups. The data represent means � SD. To comparetemperatures at 15 mm from electrodes, the repeated mea-sures analysis of variance was performed. For all statisti-cal analyses, a P value of �.05 was considered signifi-cant. Statistical analysis was performed using the Instatprogram (GraphPad Software, Inc.; San Diego, CA).

RESULTS

Electrical MeasurementsBefore RF energy was applied mean initial tissue im-

pedance in the four groups was 65 � 11 �. In group A,the impedance rose intermittently and fluctuated between80 and 150 �. The pulsed technique was activated morethan 20 times (mean) during RFA. In group B, theimpedance was slightly lower than initial value for thefirst 3–5 minutes after starting RFA, but then rose inter-mittently and activated the pulsed algorithm (mean 5.8times) (Fig 2). In group C, the impedance gradually in-creased to 85 � without an abrupt increase. The meanenergy levels applied in the three groups were 52.3 �10.3 kJ for group A, 115.4 � 10.5 kJ for group B, and38.5 � 11.5 kJ for group C. Differences in mean appliedenergy between groups A and B and groups B and Cwere statistically significant (P � .05) (Table 1).

Temperatures and Ablation Region DimensionsThe final mean temperatures measured at 15 mm from

electrodes in each group were 79 � 15°C in group A, 88 �21°C in group B, and 67 � 21°C in group C, which wasnot significantly different among the groups (P � .05).

RFA regions created in treated livers exhibited a char-acteristic central white zone surrounded by a red hemor-

rhagic zone. After staining with 2% 2,3,5-triphenyl tetra-

inter

Academic Radiology, Vol 13, No 3, March 2006 RF ABLATION USING AN ICP ELECTRODE

zolium chloride, normal liver parenchyma and peripheralhemorrhagic zones appeared pink, but the central whitezones were unstained. The Dtr1s of the central whitezones in the groups A, B, and C were 3.0 � 0.9 cm,4.9 � 0.5 cm, and 4.0 � 1.5 cm, respectively, and thedifference between groups A and B was significant (P �

Figure 2. Graphic depiction of tissue im(middle row), and power changes (upper rinternally cooled electrode (a) or using the

.05). The Dtr2s of the central white zones in groups A, B,

and C were 2.6 � 0.4 cm, 3.4 � 0.6 cm, and 2.8 � 0.9cm, respectively (P � .09). The Dax values of the centralwhite zones in groups A, B, and C were 3.3 � 0.4 cm,4.8 � 0.8 cm, and 3.8 � 1.5 cm, respectively, and thedifference between groups A and B was significant(P � .05). The Dax/Dtrm ratios in groups A, B, and C

nce (lower row), radiofrequency currenturing radiofrequency ablation using thenally cooled perfusion electrode (b).

pedaow) d

were 1.2 � 0.4, 1.2 � 0.2, and 1.1 � 0.3, respectively

347

t (P �

LEE ET AL Academic Radiology, Vol 13, No 3, March 2006

(P � .05), and mean coagulation necrosis volumes ingroups A, B, and C were 13.1 � 4.7 cm3 in group A,43.7 � 17.5 cm3 in group B, and 26.3 � 20.2 cm3 ingroup C (Fig 3), and the difference between groups Aand B was significant (P � .05) (Table 1).

The isoperimetric ratios (R) of groups A, B, and Cwere 0.91 � 0.03, 0.78 � 0.1, and 0.75 � 0.8, respec-tively (P � .05). In all three groups, deformations ofthe coagulation shape were observed when a large vessel(�3 mm) was close to the RF delivery area, regardlessof which RF electrode system was used. In two group Canimals treated with a perfusion electrode, rather largeirregularities were observed in the coagulated area. Inaddition, in two group C animals treated with the perfu-sion electrode and in one group B animal treated with theICP electrode, a circular area of necrosis was observed atthe periphery of large vessels distant from the thermallyinduced coagulation area (Fig 4). In addition, the coeffi-cient of variation of the volumes of the RF-induced abla-tion zones produced using ICP and perfusion electrodeswere 0.36, 0.40, and 0.78, respectively. Therefore, RFAusing a perfusion electrode showed a larger variation inRFA volume than RFA using either an internally cooledelectrode or an ICP electrode.

Histopathologic ResultsRegardless of RF equipment, ablated regions in repre-

sentative cases demonstrated a central necrotic zone sur-

Table 1Measured Values of Radiofrequency-Inducedthe Three Groups

Coagulation Necrosis Group A

Dtr1 (cm) 3.0 � 0.8Dtr2 (cm) 2.6 � 0.4Dax (cm) 3.3 � 0.4Volume (cm3) 13.1 � 4.7Ratio of Dax/mean Dtr 1.2 � 0.42Coefficient of variation 0.36Isometric ratio 0.91 � 0.03Mean applied energy (kJ) 52.3 � 10.3

NOTE.—Dtr1 and Dtr2 � two diameters of thablation zones in slices taken perpendicular to tthe ablation zone along the electrode insertion ainternally cooled electrode, Group B � radiofreqradiofrequency ablation using a perfusion electr

�Difference between groups A and B was sign†Difference between groups A and C was sig‡Differences among all groups were significan

rounded by a peripheral hemorrhagic zone consisting of

348

necrotic hepatocytes, interstitial hemorrhage, and poly-morphonuclear leukocyte infiltrates. No viable cells werefound within the central necrotic zone, but within hemor-rhagic lesions areas of sinusoidal congestion and hemor-rhage were accompanied by advanced necrotic changesand patches of living cells.

DISCUSSION

In this in vivo study, RFA using an ICP electrode(group B: 43.7 cm3) was found to create larger coagula-tion regions than RF ablation using an internally cooled(group A: 13.1 cm3) or a perfusion electrode (group C:26.3 cm3) (Fig 3). The larger RF-induced ablation zonecreated using the ICP electrode could be attributed to sev-eral factors. First, the infusion of 14.6% NaCl solutionduring RFA increased the electrical conductance of thetissue and consequently prevented a marked impedancerise, and facilitated greater energy delivery (12), and moreenergy input at the electrode allowed more heat to be de-livered into the tissue. In the present study, the mean ap-plied energy in group B was greater than in groups A andC—52.3 kJ for group A, 115.4 kJ for group B, and 38.5kJ for group C, respectively (P � .05). Second, the infu-sion of 14.6% NaCl solution into liver tissue through theICP electrode during RFA increases thermal conductance.Third, the local cooling effect and the hydration from the

gulation Necrosis and Delivered Energy in

Group B Group C P Value

4.9 � 0.5 4.0 � 1.5 �.05�

3.4 � 0.6 2.8 � 0.9 .094.8 � 0.8 3.8 � 1.5 �.05�

43.7 � 17.5 26.3 � 20.2 �.05�

1.2 � 0.2 1.1 � 0.3 �.050.4 0.78

0.78 � 0.1 0.75 � 0.8 �.05†

15.4 � 10.5 38.5 � 11.5 �.01‡

tral, white region of radiofrequency-inducedlectrode insertion axis. Dax � axial diameter ofGroup A � radiofrequency ablation using any ablation using an ICP electrode, Group C �

nt (P � .05).nt (P � .05).

.05).

Coa

1

e cenhe exis.uenc

ode.ifica

nifica

instillation of room temperature saline could help prevent

Figure 3. Photographs of radiofrequency-induced coagulation areas cre-ated by applying a radiofrequency current using an internally cooled elec-trode (a), an internally cooled perfusion electrode (b), or a perfusion elec-trode (c). (a) Photograph of specimen from group A (radiofrequency abla-tion [RFA] using an internally cooled electrode). (b) Photograph of specimenfrom group B (RFA using an internally cooled perfusion electrode). (c) Photo-

graph of specimen from group C (RFA using a perfusion electrode).

349

n zo

LEE ET AL Academic Radiology, Vol 13, No 3, March 2006

tissue desiccation. Fourth, infused saline could removeelectrically insulating gases from the electrode (28).

Several researchers (14,15,29) have shown that wetRFA using a perfusion electrode (Berchtold) creates asimilar or larger ablation region dimension than a conven-tional electrode at lower power. Our study results were inaccord to the previous studies (14,15). More importantly,in our study, RFA using the ICP electrode created alarger volume of coagulation than RFA with a perfusionelectrode. In a previous study (21), the combined use ofHS infusion and internal cooling produced a larger abla-tion area than the use of either facility alone by prevent-ing adjacent tissue boiling and allowing larger energydelivery. Indeed, our study demonstrates that during RFA,the ICP electrode allowed a higher energy delivery thanthe perfusion electrode or the internally cooled electrode.Furthermore, the perfusion electrode showed largest varia-tion in terms of RF-induced ablation zone volume. Thiscould be attributed to the effect of the combination ofthe inhomogeneous diffusion of infused NaCl solutionand low RF energy delivery. This large variation in RF-induced ablation zone volume makes it difficult for clini-cians to plan an RFA procedure and represents a limita-

Figure 4. Photograph of gross pathologi20-minute radiofrequency application usinoration surrounding the large vessel wall (crosis, distant from the primary coagulatio

tion of the perfusion RF system.

350

In clinical practice, when a large tumor (�3 cm indiameter) is treated by RFA, multiple sequential RF ap-plication is inevitable. However, this is technically quitechallenging to perform, and nonablated tissue left at anysurface may result in local recurrence. In our study, meancoagulation volumes were 43.7 cm3 for the ICP electrode.A previous study by Pereira and colleagues (30) describedmean coagulation volumes were 16.2 cm3 for a 12-tineelectrode, 20.5 cm3 for a cluster electrode, and 9.8 cm3

for nine-tine electrode. Therefore, our study results sug-gest that when treating large tumors, a single RF applica-tion with a proposed ICP electrode is more likely to killthe tumor and in a shorter time.

Despite the fact that wet RFA using the ICP electrodecreated much a larger coagulation volume than did othercommercially available electrodes, deformations of thecoagulation shape were observed when a large vessel(�3 mm) was close to the RF delivery area regardless ofwhich RF electrode system was used. This limitation ofablation areas adjacent to large vessels because of theheat-sink effect, may be a universal limitation of monopo-lar RFA. Perfusion-mediated RFA or RFA using a bipolaror multipolar RFA system with a higher current density

cimen in a group C animal treated witherfusion electrode. Note circular discol-), which corresponds to coagulation ne-

ne.

c speg a parrow

near the electrodes, may provide a possible solution to

Academic Radiology, Vol 13, No 3, March 2006 RF ABLATION USING AN ICP ELECTRODE

this problem (7,31). In addition, the ICP electrode hasanother potential problem (ie, the ICP electrode has 14gauges, which is rather larger than an internally cooledelectrode [17 gauge] in group A or a perfusion electrode[16 gauge]). Theoretically, a larger needle electrode mayhave a higher risk of bleeding, but the use of RF needletrack ablation during needle withdrawal may decrease therisk of track bleeding.

HS infusion into the tumor tissue during RFA may in-duce not only an increment in electrical and thermal conduc-tivities but also an additional antitumoral effect. Lin andcolleagues (32) demonstrated that hypertonic saline injectionhad the effect of controlling the growth of VX2 carcinomacells and extending the life of rabbits with VX2 tumors. Intheir evaluation of the effect of NaCl concentration on theviability of VX2 cancer cells, they demonstrated that at theseventh and twelfth hours, most cancer cells were dead inthe test tubes of 5%, 15%, 25%, and 36.5% NaCl solutions,whereas the control and 0.9% test tubes had the same num-ber of tumor cells as at the fifth hour.

In our study, we used a 14.6% NaCl solution throughthe ICP electrode during RFA to improve electrical andthermal conductivity according to our preliminary studyresults. A previous study by Goldberg and colleagues (33)showed that preinjection of 6 mL saturated saline beforeRFA created a larger coagulation volume than preinjec-tion of other NaCl solutions. However, we did not use36% NaCl as an infusate because continuous infusion of36% NaCl solution during RFA (12 minutes) was shownin our preliminary study to pose a risk of decreased tissueimpedance (below 50 �). This great decrease of tissueimpedance was not helpful in increasing RF-induced co-agulation necrosis because too much decease of tissueimpedance induces a remarkable increase of currentneeded to create the heat. This discrepancy in the effectof saturated saline infusion during RFA to create coagula-tion necrosis between wet RFA and RFA with HS prein-jection may be explained by the fact that wet RFA allowscontinuous replacement of diffused-out saline and contin-uous filling of the gap between the electrode and the adja-cent tissue with infused saline, and consequently providesa more effective increase of electrical conductance thanRFA with HS preinjection (20). Furthermore, infusion of36% NaCl solution at a low infusion rate (0.5 mL/minute)showed frequent blockage of the saline perfusion holesbecause of evaporation of liquid at the active tip portionof the electrode. In the present study, no cases showed amarked decrease of impedance during RFA in group B

with wet RFA. Although an increasingly negative linear

correlation was observed between the tumor coagulationdiameter and the overall baseline system impedance, toogreat a decrease of the impedance of surrounding tissueas well as decrease of tumoral impedance by injected HScould potentially hinder complete tumor destruction (34).From this perspective, if the RFA procedure time is in-creased, wet RFA using a 14.6% NaCl solution also posesa risk of too great a decrease of impedance.

The problems associated with NaCl solution infusionduring RFA include the risk of unexpected thermal injuryrelated to hot saline diffusing through vital structures andof peritoneal or the track seeding of tumor cells withleaking saline. Theoretically, the risk of thermal injury isincreased when a high volume or a rapid infusion ofNaCl solution is used (33,35). Therefore, further experi-mental studies concerning the optimal of amount and con-centration of HS are warranted. Boehm and colleagues(36) reported higher complication rates using a perfusionelectrode than an internally cooled electrode for the treat-ment of breast cancer in an animal model. However, in arecent study (37) concerning complications after wet RFAusing a perfusion electrode, no unexpected thermal injuryto vital structures and no cutaneous or abdominal wallseeding were observed. In our study, two group C ani-mals treated with the perfusion electrode and one group Banimal treated with the ICP electrode showed a circulararea of necrosis at the periphery of the large vessels dis-tant from the thermally induced coagulation area (Fig 4).Even if those animals did not show any significant abnor-mal symptoms compared with the other animals duringthe 4-day observation period, it suggests that HS infusionduring RFA may cause unexpected thermal injury alongthe hepatic vasculature and may thereby result in seriousdamage to the bile duct near the portal vein.

Our study has limitations that should be mentioned. First,the extent to which the results of the present in vivo studyrelate to the human liver are limited. Moreover, in thepresent study, all ablations involved the normal liver paren-chyma and not tumor tissue. Despite these shortcomings, webelieve that our model provides a means of testing the effi-cacy and safety of RFA for creating coagulation necrosis.Second, to resolve the safety issues concerning the irregularshape of coagulation necrosis produced by wet RFA as re-ported previously (12,30), its therapeutic efficacy and safetyshould be examined in liver tumor model in large animalsbefore the developed perfused cooled electrode is used inhumans. Third, we used 14.6% HS solution as an infusate ata fixed flow rate of 1 mL/minute as indicated by previous

preliminary data. However, in view of the fact that marked

351

LEE ET AL Academic Radiology, Vol 13, No 3, March 2006

increases in tissue conductivity from excessive sodium chlo-ride have the undesirable effect of reducing tissue heating(34), an automatic-control HS infusion rate system based onthe continuous monitoring of impedance during RFA wouldbe desirable. Last, tissue temperature was measured at onlyone point. If we had measured the tissue temperature atmany points, a three-dimensional temperature distributionwould have been obtained.

In conclusion, the present study shows that wet RFAusing the developed ICP electrode produced ablationzones that were significantly larger those achieved usinga perfusion electrode or an internally cooled electrode.Based on our study results, we believe that the larger vol-ume of coagulation necrosis created using the developedICP electrode may increase the clinical utility of RFAtherapy by allowing the successful treatment of largerhepatic tumors or by reducing the number of sessionsneeded to treat a given tumor.

REFERENCES

1. McGahan JP, Dodd GD. Radiofrequency ablation of the liver: currentstatus. AJR Am J Roentgenol 2001; 176:3–16.

2. Gazelle GS, Goldberg SN, Solbiati L, et al. State of the art: tumor abla-tion with radio-frequency energy. Radiology 2000; 217:633–646.

3. Garcea G, Lloyd TD, Aylott C, Maddern G, et al. The emergent role offocal liver ablation techniques in the treatment of primary and second-ary liver tumors. Eur J Cancer 2003; 39:2150–2164.

4. Lau WY, Leung TWT, Yu SC, et al. Percutaneous local ablative therapyfor hepatocellular carcinoma: a review and look into the future. AnnSurg 2003; 237:171–179.

5. Goldberg SN, Ahmed M. Minimally invasive image-guided therapies forhepatocellular carcinoma. J Clin Gastroenterol 2002; 35:S115–S129.

6. Gillams AR. Thermal ablation of liver metastases. Abdom Imaging2001; 266:361–368.

7. Goldberg SN. Radiofrequency tumor ablation: principles and tech-niques. Eur J Ultrasound 2001; 13:129–147.

8. Goldberg SN, Dupuy DE. Image guided radiofrequency tumor ablation:challenges and opportunities-part I. J Vasc Interv Radiol 2001; 12:1021–1032.

9. Livraghi T, Goldberg SN, Lazzaroni S, et al. Hepatocellular carcinoma:radiofrequency ablation of medium and large lesions. Radiology 2000;214:761–768.

10. Komorizono Y, Oketani M, Sako K, et al. Risk factors for local recur-rence of small hepatocellular carcinoma tumors after a single session,single application of percutaneous radiofrequency ablation. Cancer2003; 97:1253–1262.

11. Dupuy DE, Goldberg SN. Image-guided radiofrequency tumor ablation:challenges and opportunities-part II. J Vasc Interv Radiol 2001; 12:1135–1148.

12. Leveillee RJ, Hoey MF. Radiofrequency interstitial tissue ablation: wetelectrode. J Endourol 2003; 17:563–577.

13. Goldberg SN, Gazelle GS, Solibati L, et al. Large volume radio-frequency tissue ablation: increased coagulation with cooled-tip elec-trodes. Radiology 1996; 201:387–343.

14. Kettenbach J, Kostler W, Rucklinger E, et al. Percutaneous saline-enhanced radiofrequency ablation of unresectable hepatic tumors: ini-tial experience in 26 patients. AJR Am J Reontgenol 2003; 180:1537–1545.

15. Hansler J, Frieser M, Schaber S, et al. Radiofrequency ablation of he-

patocellular carcinoma with a saline solution perfusion device: a pilotstudy. J Vasc Interv Radiol 2003; 14:575–580.

352

16. Dodd GD, Frank MS, Aribandi M, Chopra S. Chintapalli KN. Radiofre-quency thermal ablation: computer analysis created by overlapping ab-lations. AJR Am J Roentgenol 2002; 177:777–782.

17. Choi D, Lim HK, Kim MJ, et al. Overlapping ablation using a coaxialradiofrequency electrode and multiple cannulae system: experimentalstudy in ex-vivo bovine liver. Korean J Radiol 2003; 4:117–123.

18. Goldberg SN, Solbiati L, Hahn PF, et al. Large-volume tissue ablationwith radio frequency by using a clustered, internally cooled electrodetechnique: laboratory and clinical experience in liver metastases. Radi-ology 1998; 209:371–379.

19. Haemmerich D, Lee FT Jr, Schutt DJ, Sampson LA, Webster JG, FineJP, et al. Large volume Radiofrequency ablation of ex vivo bovine liverwith multiple cooled cluster electrode. Radiology 2005; 234:563–568.

20. Lee JM, Han JK, Kim SH, et al. Comparison of wet radiofrequency ab-lation with dry radiofrequency ablation and radiofrequency ablation us-ing hypertonic saline preinjection: ex vivo bovine liver. Korean J Radiol2004; 5:258–265.

21. Lee JM, Han JK, Kim SH, et al. Optimization of wet radiofrequency ab-lation using a perfused-cooled electrode: a comparative study in exvivo bovine livers. Korean J Radiol 2004; 5:250–257.

22. Goldberg SN, Stein M, Gazelle GS, Sheiman RG, Kruskal JB, ClouseME. Percutaneous radiofrequency tissue ablation: optimization ofpulsed-RF technique to increase coagulation necrosis. J Vasc IntervRadiol 1999; 10:901–916.

23. Schmidt D, Trubenbach J, Brieger J, et al. Automated saline-enhancedradiofrequency thermal ablation: initial results in ex vivo bovine livers.AJR Am J Reontgenol 2003; 180:163–165.

24. Goldlust EJ, Placzynski RP, he YY, Hsu CY, Coldberg MP. Automatedmeasurement of infarct size with scanned images of triphenyltetrazo-lium chloride-stained rat brains. Stroke 1996; 27:1657–1662.

25. Image J. Available online at http://rsb.info.nih.gov/ij/. AccessedNovember 20, 2004.

26. Chinn SB, Lee FT Jr, Kennedy GD, et al. Effect of vascular occlusionon radiofrequency ablation of the liver: results in a porcine model. AJRAm J Roentgenol 2001; 176:789–795.

27. Do Carmo MP. Differential geometry of curves and surfaces. Engle-wood, NJ: Prentice-Hall, 1976.

28. Miao Y, Ni Y, Yu J, Zhang H, et al. An ex vivo study on radiofrequencytissue ablation: increased lesion size by using an “expandable-wet”electrode. Eur Radiol 2001; 11:1841–1847.

29. Brieger J, Pereira PL, Trubenbach J, et al. In vivo efficiency of fourcommercial monopolar radiofrequency ablation systems: a comparativeexperimental study in pig liver. Invest Radiol 2003; 38:609–616.

30. Pereira PL, Trubenbach J, Schenk M, et al. Radiofrequency ablation: invivo comparison of four commercially available devices in pig livers.Radiology 2004; 232:482–49.

31. Haemmerich D, Tungjitkusolmun S, Staelin ST, et al. Finite-elementanalysis of hepatic multiple probe radio-frequency ablation. IEEE TransBiomed Eng 2002; 49:836–842.

32. Lin YC, Chen JH, Han KW, et al. Ablation of liver tumor by injection ofhypertonic saline. AJR Am J Roentgenol 2005; 184:212–219.

33. Goldberg SN, Ahmed M, Gazelle GS, et al. Radiofrequency thermal ab-lation with NaCl solution injection: effect of electrical conductivity ontissue heating, and coagulation-phantom and porcine liver study. Radi-ology 2001; 219:157–165.

34. Lobo SM, Afzal KS, Ahmed M, et al. Radiofrequency ablation: model-ing the enhanced temperature response to adjuvant NaCl pretreatment.Radiology 2004; 230:175–182.

35. Gillams AR, Lees WR. CT mapping of the distribution of saline duringradiofrequency ablation with perfusion electrodes. Cardiovasc InterventRadiol 2005; 28:476–480.

36. Boehm T, Malich A, Goldberg SN, et al. Radio frequency tumorablation: internally cooled electrode versus saline enhanced techniquein an aggressive rabbit tumor model. Radiology 2002; 222:805–813.

37. Giorgio A, Tarantino L, de Stefano G, et al. Complications after percu-taneous saline-enhanced radiofrequency ablation of liver tumors:3-year experience with 336 patients at a single center. AJR Am J

Roentgenol 2005; 184:207–211.