Intrahepatic radiofrequency ablation versus electrochemical treatment in vivo

Surgical Oncology 21 (2012) 79e86

Contents lists available at ScienceDirect

Surgical Oncology journal homepage: www.elsevier.com/locate/suronc

Review

Intrahepatic radiofrequency ablation versus electrochemical treatment in vivo Ralf Czymek a, *, Jan Nassrallah a, Maximilian Gebhard b, Andreas Schmidt a, Stefan Limmer a, Markus Kleemann a, Hans-Peter Bruch a, Philipp Hildebrand a a b

Department of Surgery, University of Luebeck Medical School, Ratzeburger Allee 160, D-23538 Luebeck, Germany Institute of Pathology, University of Luebeck Medical School, Ratzeburger Allee 160, Luebeck, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Accepted 27 October 2010

Background: Radiofrequency ablation (RFA) and electrochemical treatment (ECT) are two methods of local liver tumour ablation. The objective of this study was to compare these methods when applied in proximity to vessels in vivo. Methods: In a total of ten laparotomised pigs, we used ECT (Group A, four animals) and RFA (Group B, four animals) to create four areas of ablation per animal under ultrasound guidance within 10 mm of a vessel. Group C consisted of two control animals. Chemical laboratory tests were performed immediately before and after each procedure and on days 1, 3 and 7 after surgery. Following the last tests, the livers were harvested for morphological evaluation. Results: The mean duration of surgery was 5 h 40 min in Group A (ECT), 2 h 47 min in Group B (RFA), and 2 h 30 min in Group C (control animals). After ECT, the harvested livers showed a mean volume of necrosis of 1.84 cm3
0.88 at the anode and 2.59 cm3
1.06 at the cathode. The presence of vessels did not influence the formation of necrotic zones. Ablation time was 67 min when a charge of 200 coulombs was delivered. We measured pH values of 1.2 (range: 0.9e1.7) at the anode and 11.7 (range: 11.0e12.1) at the cathode. In one of the 16 RFA ablations (6%), the target temperature was not reached and the procedure was discontinued. After 14 of 16 RFA procedures (88%), morphological analysis showed incomplete ablation in perivascular sites. Both ECT and RFA were associated with a reversible increase in monocyte, C-reactive protein (CRP) and aspartate aminotransferase (AST) levels. There was no significant increase in interleukin-1b (IL-1b), tumour necrosis factor-a (TNF-a) and IL-6. Conclusion: In the majority of cases, intrahepatic RFA in vivo leads to incomplete necrosis in proximity to vessels and the presence of histologically intact perivascular cells. Without a reduction in liver perfusion, the central application of RFA should be considered problematic. ECT is a safe alternative. It is not associated with a heat sink effect but has the disadvantage of long treatment times. Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: Electrochemical treatment Radiofrequency ablation Liver Cytokine In vivo Heat sink effect

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Electrochemical treatment in close proximity to vessels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Radiofrequency ablation in close proximity to vessels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Conflict of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Authorship statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

* Corresponding author. Tel.: þ49 (0) 451 500 2046; fax: þ49 (0) 451 500 6353. E-mail address: [email protected] (R. Czymek). 0960-7404/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.suronc.2010.10.007

80

R. Czymek et al. / Surgical Oncology 21 (2012) 79e86

Introduction Local ablative therapies such as radiofrequency ablation (RFA) or electrochemical treatment (ECT) are possible management options for patients with malignant liver lesions that cannot be cured by resection. RFA is a widely used thermal procedure that is reported to lead in many cases to incomplete tumour ablation [1,2], which is the result of a cooling or heat sink effect of nearby vessels. ECT involves the placement of two or more electrodes into tissue and the continuous delivery of a direct current between the electrodes. Pioneer work in ECT was carried out from 1978 to 1983 by Bjoern Nordenstrom, a Swedish radiologist, who used this procedure to treat malignant tumours of the lung [3,4]. Major clinical studies were conducted in Russia and especially in China in the 1990s [5,6]. Electrochemical treatment is reported to be associated with few complications and to effectively destroy tumour tissue. Although ECT has been widely used in Asia, scientific data in the literature is limited and heterogeneous. To our knowledge, there is only one study that compared RFA and ECT and investigated areas of ablation in peripheral liver tissue [7]. Since a heat sink effect is most likely to occur when RFA is performed in close proximity to a vessel, the effects of RFA and ECT in perivascular sites are of particular interest. The objective of this study was therefore to investigate and compare the effectiveness of RFA and ECT in terms of perivascular ablation in vivo with a special focus on chemical laboratory findings. Methods The study was approved by the Ministry of Agriculture, Environment and Rural Areas of the State of Schleswig-Holstein in Kiel (Ref. V 312-72241.122-7 e 89-9/08). Ten domestic pigs with a median weight of 40 kg (range: 26e45 kg) underwent a midline laparotomy and a transverse right upper abdominal incision under general anaesthesia and intubation. Depth of anaesthesia was controlled on the basis of a target heart rate of 80e100 min1 using a combination of 30 ml of Propofol-Lipuro 2% (20 mg/ml), 2 ml of ketamine 10% (100 mg/ml), 3 ml of midazolam (5 mg/ml) and 0.5 ml of xylazine 2% (20 mg/ml), which was administered via a syringe pump. After intubation, the animals received pressure-controlled mechanical ventilation (Fabius Tiro ventilator, Draeger, Luebeck, Germany) with a mixture of 50% oxygen and 50% nitrous oxide at a rate of 14 min1, an inspiration/expiration rate of 1:2, a positive end-expiratory pressure (PEEP) of 2 mbar and a ventilation pressure of 20 mbar. The 10 animals were randomised into three groups (Group A: 4 animals, Group B: 4 animals, Group C: 2 animals). Electrochemical treatment was used to create four intrahepatic central areas of ablation in each of the Group A animals. Two platinum electrodes were used and separated by 15 mm. An electric charge of 200 coulombs was delivered using an electric current of 50 mA at a voltage of 25 volts. Treatment time was 67 min. Radiofrequency ablation was performed to create four ablation areas in each of the Group B animals. Treatment time was 10 min. Platinum electrodes were placed into the livers of Group C animals (control group) after laparotomy and left in situ for a period of 120 min. No electric current was delivered. Under ultrasound guidance, all electrodes were placed in central locations at a distance of 5e10 mm from a vessel with a diameter of at least 5 mm. In order to investigate electrochemical treatment, we used a direct current generator (ECU 300, Soering Medizintechnik, Quickborn, Germany) and platinum electrodes with a diameter of 1 mm. A specially designed guiding template helped us place the ECT electrodes at exact locations. We performed our RFA experiments using a monopolar radiofrequency ablation system (RITA, Medical Systems, Manchester, United States, Model 1500X

Generator, and a StarBurst Talon electrode array with a length of 25 cm and a diameter of 4 cm). In our study, the system delivered a maximum of 150 watts. Blood samples were collected under analgesic sedation induced by an intramuscular injection of xylazine (RompunÒ 2%, 1 mg per kg body weight), ketamine (KetaminÒ 10%, 10 mg per kg body weight), midazolam (0.5 mg per kg body weight) and atropine (0.03 mg per kg body weight) at five different time points (before surgery, after surgery, and on days 1, 3 and 7). Chemical laboratory analyses were conducted by the Central Laboratory of the University of Luebeck Medical School. Tumour necrosis factor-a (TNF-a), interleukin-1b (IL-1b) and interleukin-6 (IL-6) were analysed using commercially available ELISA kits (R & D Systems, Minneapolis, United States). On day 7 after surgery, blood samples were collected for the fifth time and a second laparotomy was performed under general anaesthesia. Following open liver ultrasound, the pigs were sacrificed and the livers were harvested. After the specimens had been prepared and stained according to a standard haematoxylin and eosin protocol, they were analysed histologically by independent experts at the Institute of Pathology of the University of Luebeck Medical School. The volumes of the ECT-induced cylindrically shaped foci of necrosis were calculated using the formula: V ¼ p r2  h. In addition, the maximum widths of the RFA-induced areas of necrosis were measured. pH levels were measured with an Orion 3-Star Plus Portable pH Meter (Thermo Fisher Scientific, Waltham, United States). The sensor had a diameter of 2 mm. The Chi-square test, Fisher’s exact test and the ManneWhitney U test were used for statistical analysis. Results Electrochemical treatment in close proximity to vessels The mean weight of the animals was 38.5 kg (range: 26.0e43.5 kg) in Group A (ECT), 42.5 kg (36.0e45.0 kg) in Group B (RFA) and 37.5 kg (35.0e40.0 kg) in Group C (control animals). The differences were not significant. The livers, which were harvested on day 7 after surgery, had a mean weight of 1114 g
176 g. There were no significant differences between the groups. The mean duration of the procedure was 5 h 40 min in Group A (ECT), 2 h 47 min in Group B (RFA) and 2 h 30 min in Group C. During surgery, body temperatures dropped by 1.9  C in Group A (ECT), increased by 1.9  C in Group B (RFA), and dropped by 2.2  C in Group C (controls). All animals had an uneventful postoperative course. Current and voltage parameters remained constant during all phases of electrochemical treatment. A newly developed guiding template allowed us to place the electrodes at an exact distance of 15 mm between anode and cathode. The electrodes were positioned on both sides of an ultrasonographically identified vessel with a diameter of at least 5 mm. After a mean treatment time of 130 s
31 s, gas bubbles were seen at both electrode placement sites at the level of the liver capsule. In all procedures, the electric current generator switched off after 67 min. ECT-induced liver lesions were imaged by ultrasonography, which demonstrated sharply demarcated areas treated by ECT. The images showed necrotic zones of low echogenicity at the cathodes and necrotic zones of higher echogenicity with well-defined areas surrounding the electrode track at the anodes (Fig. 1). Thrombosis of a major vessel occurred in one of the 16 ECT procedures (Fig. 1). Before the livers were harvested on day 7 after surgery, the lesions were once again evaluated by ultrasonography. A direct comparison of the findings at the two different time points revealed that it was hardly possible one week after surgery to distinguish the

R. Czymek et al. / Surgical Oncology 21 (2012) 79e86

81

Figure 1. The ultrasound scan in the left image shows clearly visible zones of necrosis around the cathode and anode immediately after ECT with intraoperative ultrasound. The image on the right shows the specimen (same as on the left) after removal on the 7th postoperative day and after fixation in formalin for 24 h. The thrombus that can be seen in the image was friable in consistency. The pathologist believed that the thrombus had been caused by ECT and not by an artefact during organ harvesting.

foci of necrosis from surrounding liver tissue. Likewise, the differences in the types of lesions at the anode and cathode were minimal, if at all present. A morphological assessment of the lesions revealed that the mean volume of necrosis was 1.84 cm3
0.8 at the anode and 2.59 cm3
1.06 at the cathode. The pH level was 1.2 (range: 0.9e1.7) at the anode and 11.7 (range: 11.0e12.1) at the cathode. Basophilic cell nuclei, which were weakly stained and surrounded by a halo, were found at the cathode (Fig. 2). A morphological comparison of the necrotic areas at the anode and at the cathode showed that the cells at the cathode were less confluent. The 16 necrotic areas at the anode and the 16 necrotic areas at the cathode contained 45 further vessels with a diameter of more than 3 mm. The total number of vessels was thus 61. No morphologically intact hepatocytes were detected in the anodic and cathodic zones where the presence of necrosis was suspected macroscopically. Both types of necrotic zones were surrounded by a 2- to 3- mmwide rim of fibrous tissue that contained spindle-shaped fibroblasts

Figure 2. Histological specimen on day 7 after electrochemical treatment. The necrotic zone at the anode (top) is separated from the area of colliquative necrosis at the cathode (bottom) by an eosinophilic zone (haematoxylin and eosin staining, magnification: 100).

and biliary canaliculi similar to those seen in cirrhotic patients. Morphologically intact hepatocytes were seen beyond this narrow area of fibrosis. Only a few lymphocytes and individual monocytes were detected in the area surrounding the necrotic zones at the anode and cathode. As in ex vivo experiments, vessels in the areas treated by ECT were destroyed through lysis without macroscopic or microscopic evidence of extravasation or the formation of haematoma. Histology revealed the presence of infarcted and avital tissue and the absence of morphologically intact hepatocytes in the macroscopically intact zone between the areas of colliquative and coagulative necrosis (Fig. 1). Radiofrequency ablation in close proximity to vessels The RFA probe was inserted centrally into the liver at a distance of 5e10 mm from an ultrasonographically identified vessel. Four ablations were performed per animal, amounting to a total of 16 procedures. Despite a treatment time of 10 min, the target temperature was not reached in one of the 16 ablations (6%). This was indicated by the RFA generator. In the other 15 procedures, the target temperature was not reached at some electrode tips. Since this was compensated by the other components of the RFA electrode array, the generator indicated a normal ablation procedure. The mean duration of the heating phase was 143 s
27 s, after which the target temperature of 105  C was reached. The subsequent treatment time was 10 min. At the end of the treatment episode, the electrode array was retracted and needle track ablation was performed for a further minute. An ultrasound scan that was obtained immediately after the intervention demonstrated a normal needle track. It also showed a sharp demarcation between ablated and macroscopically normal liver tissue (Fig. 3). When the livers were harvested on day 7 after surgery, the liver lesions were once again evaluated by ultrasonography after laparotomy. Unlike the ECT-induced liver lesions, necrotic and surrounding liver tissues were still sharply demarcated on day 7. A morphological evaluation of the 16 areas of ablation in close proximity to a vessel showed areas of incomplete ablation or deformities in the shape of the necrotic zones in 14 cases (88%). The location of these areas in relation to the vessel lumen suggested that these findings were attributable to a heat sink effect. The 16 foci of necrosis contained a total of 52 vessels with a diameter of

82

R. Czymek et al. / Surgical Oncology 21 (2012) 79e86

Figure 3. The ultrasound scan in the left image shows a clearly visible zone of necrosis immediately after RFA with intraoperative ultrasound. The image on the right shows the specimen (same as on the left) after removal on the 7th postoperative day and after fixation in formalin for 24 h. The right image shows that the zone of ablation extends across a transverse gap between two liver lobes.

more than 3 mm. As a result of the heat sink effect, incomplete areas of necrosis, which reflect incomplete ablation, were observed in 38 of these 52 vessels (73%). The lesions that were created in close proximity to a vessel had a mean width of 2.91 cm
0.81 cm. On day 7 after RFA, histology revealed the presence of confluent connective tissue trabeculae of the liver lobules and grey-brown cytoplasm with an acidophilic component in the necrotic zone (Fig. 4). After RFA, too, there was a fibrotic rim with phagocytosis of necrotic hepatocytes. As in the case with ECT, only a few lymphocytes or other immunocompetent cells were detected beyond the zone of fibrosis. Histology showed the presence of morphologically intact hepatocytes and lobules in macroscopically visible areas of incomplete ablation in perivascular sites. Chemical laboratory findings are presented in Fig. 5 through [7].

popular in Europe and the United States [8]. As early as 1948, Pennes [9] introduced the arterial-tissue heat transfer term to describe a factor that is believed by other authors to limit the size of the area of ablation [10]. Goldberg et al. [11] referred to this local effect, which depends on individual vascular anatomy, by the term “heat sink effect.” Not only does this effect limit the total size of ablation but it is also discussed as a possible cause of the presence of intact tumour cells in close proximity to vessels [12].

Discussion Among the available methods of local liver tumour ablation, radiofrequency ablation in particular has become increasingly

1200% 1000%

CRP [mg/l] 800% 600%

Mean (RFA) Mean (ECT)

400%

Mean (Control)

200% 0% 1

2

3

4

5

Time point of blood collection Figure 4. Histological specimen on day 7 after radiofrequency ablation. The thermal effect is absent in close proximity to the vessel lumen (left). As a result, there is a zone of intact liver tissue and an intact vessel wall (haematoxylin and eosin staining, magnification: 25).

Figure 5. Time points for blood collection: (1) ¼ before surgery, (2) ¼ immediately after surgery, (3) ¼ day 1 after surgery, (4) ¼ day 3 after surgery, and (5) ¼ day 7 after surgery. The figure shows the changes in percent from baseline (100%).

R. Czymek et al. / Surgical Oncology 21 (2012) 79e86

For a quantitative assessment of RFA-induced ablations, we did not use the formula that Mulier et al. [13] suggested for calculating the volume of necrotic areas. We attempted to place the electrodes in central locations and found in the majority of cases that the presence of several vessels in the area of ablation led to highly irregular geometric zones and many non-necrotic tissue portions. In our opinion, volume calculations using mathematical formulas would have falsified the results in an unacceptable manner. Such formulas may, however, be sufficiently accurate for ablations that are performed in peripheral locations. On the basis of our findings, it is doubtful that, without a reduction in liver perfusion, RFA can induce adequate and complete perivascular necrosis at a distance of 10 mm from a vessel. In our model, ECT e unlike RFA e caused complete and sharply demarcated cell destruction even in immediately perivascular areas without a heat sink effect. Although the pathophysiological mechanisms underlying electrochemical treatment are still not fully understood and clinical standards for this treatment method are far from being available, in vivo ablation procedures in the lung [14] and liver [5,15,16] have been performed worldwide and especially in China. ECT was reported by Chinese workers to be particularly effective in the treatment of superficial cutaneous and subcutaneous tumours [6,17]. In these cases, ECT was more successful than it was in the management of visceral neoplasms. Since the liver is well supplied with blood, it can more easily compensate for a shift in pH levels. According to current understanding, the extreme pH shift, which was demonstrated in multiple studies, [18e20] plays the decisive role in the induction of necrosis by electrochemical treatment. pH values between 1 at the anode [21] and 12.9 at the cathode [18] were reported. We were able to confirm these extreme values for porcine livers. Electrochemical treatment first leads to an acute pH change in the interstitial space and a pH decrease around the anode. When Hþ ions then enter intracellular space, pH also decreases in the cells. This leads to a perturbation of cell proliferation [22], the destruction of DNA, and irreversible cell damage [19]. The electric field and the pH increase at the cathode directly affect enzyme-regulated reactions [3] and destroy structure proteins, which results in inactivation or denaturation [19]. Intracellular alkalinisation leads to an increase in intracellular Ca2þ ions [23]. The further influx of extracellular ions and water into oedematous cells near the cathode, which we and other authors [18] observed, results in the swelling and rupture of cell organelles, especially mitochondria [19]. Whereas large amounts of data are available on the effects of electrochemical treatment in the management of subcutaneous lesions in small animals, there are only a few experimental studies [7,24] that address the use of ECT in the peripheral liver parenchyma of large animals. The effect of ECT in the immediate vicinity of vascular structures has been even more overlooked. WemyssHolden et al. [25] addressed this issue in an in vivo study involving six pigs. They placed ECT electrodes at either perivascular or intravascular sites in the liver and proved the safety of ECT. Surprisingly, Wemyss-Holden et al. reported that the vessel walls were histologically intact in both groups of animals despite the placement of electrodes in the immediate vicinity of vessels. This finding is in contrast to our analyses, which clearly demonstrated the histological destruction of vessel walls although there were no lumen discontinuities with bleeding complications. A possible explanation for this contradiction is that Wemyss-Holden and colleagues considered a vessel histologically intact if the continuity of the vessel lumen was maintained. In our histological examinations, the morphological alteration of endothelial cells was highly reproducible. The intrahepatic placement of electrodes in close proximity to a vessel in vivo e as in this study e or into a vessel [25] simulates a positioning of the electrodes that is probably associated

83

with the highest risks or the most adverse effect of improper application, i.e. the inadvertent injury to a vessel. Even in such an extreme situation, which is unlikely to occur under ultrasound guidance, the procedure remains safe and the occurrence of hepatic vein thrombosis is improbable. In our studies, in which one vein thrombosed, peripheral liver tissue was macroscopically and microscopically viable and showed no evidence of ischaemia one week after ECT although the electrodes were placed in immediately perivascular locations. The absence of major complications such as bleeding, cholestasis or pulmonary embolism in the 16 necrotic areas has only limited statistical value. Clinical data from controlled studies are not yet available. Several authors, however, reported the occurrence of different types of complications in association with RFA. In our opinion, a clinical comparison of the two methods (RFA versus ECT) would therefore be desirable. RFA-related complications occur in 0e14% of the cases and include bleeding, necrosis with bile duct stenosis, hepatic abscess formation, and septic complications [2,26,27]. Jiao et al. [26] reported that two of 35 patients who were treated with RFA died (mortality rate: 6%). In addition, RFA leads to incomplete tumour ablation in 0e100% of treatments [1,2,26,27]. The decisive factor determining ECT-induced ablation is the electric charge that is applied to the tissue. Electric charge is the product of electric current (in amperes) and time (in seconds) and is expressed in coulombs (C) (1 C ¼ 1 A  1 s). When we delivered an electric charge of 200 C at 2 V and 50 mA in our study, treatment time was more than 1 h in vivo. The application of an electric current of 50 mA is in the region of the upper threshold for this procedure since tissue resistance makes it almost impossible for the generator to deliver this level of current to the target organ. For this reason, treatment time can be reduced only to a certain extent by increasing the level of electric current. The platinum electrodes that were appropriately cleaned and sterilised approximately a hundred times did not show any signs of erosion even after longterm use. This was confirmed by other authors [28] although a local release of platinum salts into the tissue is theoretically possible [28,29]. Whereas RFA has the advantage of shorter treatment times, ECT is cost saving since the electrodes are reusable. Fosh et al. [30] conducted a pilot study in which nine patients with liver metastases (tumour radius: 5e30 mm) from colorectal carcinoma underwent electrochemical treatment (200e1000 coulombs). Within the electrochemically treated lesion, seven of the nine patients had no evidence of recurrence in a follow-up period of nine months. It must be assumed that, irrespective of whether radiofrequency ablation or ECT is used, the local recurrence rate increases and long-term survival decreases with malignant lesion size. Whereas standard RFA electrodes are reported to reliably manage metastases with a size of no more than 3 cm, we created areas of necrosis with a width of 20e35 mm in close proximity to a vessel. Tumours with a diameter of 3 cm must therefore be considered to be insufficiently treatable in the immediate vicinity of vessels. When we used ECT and delivered a dose of 200 coulombs, we were able to create a reproducible volume of necrosis of 4e5 cm3 (sum at the anode and cathode) that was uninfluenced by cooling effects. We were able to demonstrate in a perfusion model that larger electrode spacing, the use of four electrodes for a period of 132 min, longer treatment times and thus higher electric charges (in coulombs) allow the area of ablation to be extended to more than 20 cm3 despite a logarithmic dose-response relationship. It should, of course, be noted that we and other authors [7,31] investigated healthy liver tissue in large animals. In vivo tumour models that are used to evaluate ECT usually involve the subcutaneous injection of mammary tumour cells in small animals such as mice or rats [19,32]. There are only a few studies that address the effects of ECT in large animal models. Euler et al. demonstrated that

84

R. Czymek et al. / Surgical Oncology 21 (2012) 79e86

the results obtained for small animals cannot be simply transferred to larger species [33]. When we decided to use a porcine model to investigate the two methods of tissue ablation in an organ similar in size to that of a human liver, we accepted that our study involved tumour-free tissue since a reproducible tumour model for large animals is not yet available. Unlike a number of other authors [7,34], we did not inject a collagen or agarose gel in order to simulate an intrahepatic tumour. Although the injection of a combination of agarose, cellulose, glycerol and methylene blue can mimic the presence of a tumour and can cause a flow of ions, we are of the opinion that the absence of cellular structures in such a mass does not make this approach superior to the use of ablative methods in intact tissue. Of particular relevance to our experimental design are the results reported by Li et al. [18], who showed that electrochemical treatment destroyed both normal and malignant cells in a similar fashion. In addition, the creation of what was termed tumour mimics, which alter local conditions, was not suited to the purposes of our study, which involved the histological investigation of cell groups and the determination of the size of necrotic areas. When Hinz et al. [7] investigated organs that were harvested 48 h after RFA or ECT, they observed complete cell lysis in the areas surrounding the artificial tumour masses. Whereas cell membranes and nuclei were completely destroyed after ECT, cell nuclei were still morphologically intact after RFA. In addition, the authors observed lymphocyte infiltration after ECT but not after RFA. They concluded that ECT might lead to a more effective immune response as a result of the presence of lymphocytes, which can be antigen-presenting cells [7]. This, however, has not yet been proven. A number of authors found that both methods are able to induce an activation of the body’s immune system [6,7,35]. Unlike Hinz et al, Wemyss-Holden and colleagues did not observe lymphocyte infiltration in a number of studies even when the experimental animals survived for 72 h [36e38]. Wemyss-Holden et al. [37] described a sharply demarcated zone of ECT-induced necrosis surrounded by a narrow rim of tissue in which the small sinusoids were thrombosed. They proposed that these occluded vessels formed a “cocoon” isolating the former tumour zone from intact tissue. Wemyss-Holden et al. hypothesised that the isolation of a tumour in such a cocoon might prevent an immediate systemic immune response known to occur with RFA [25,36,38]. This theory is, however, controversial. Seven days after ECT, we found no histological evidence of leukocyte infiltration at the sites where the anode and the cathode had been placed. The necrotic area was surrounded by a zone of fibrosis where the removal of necrotic material from the margin towards the centre of ablation had started. Apart from proliferating fibroblasts, no significant numbers of immunocompetent cells such as lymphocytes were found. Similarly, we detected only a small number of lymphocytes in the tissue surrounding the RFA-induced zone of necrosis. It was interesting to note that, on day 7 after ECT, the liver tissue between the necrotic zones around the anode and the cathode appeared macroscopically intact but was infarcted and avital on histological examination (Fig. 1). We did not make this observation when we used an isolated liver perfusion model. The specimens that were prepared and analysed immediately after ECT showed intact tissue between the necrotic zones. This secondary in vivo effect of tissue infarction between the areas of colliquative and coagulative necrosis is of particular relevance since it increases the total volume of necrosis and thus the effect of ECT. Von Euler et al. [33] reported that aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels tended to rise immediately after electrochemical treatment in dogs. Other authors found that the liver enzymes that were elevated after ECT returned to normal within one week [24]. In the literature, there are no other

Figure 6. Time points for blood collection: (1) ¼ before surgery, (2) ¼ immediately after surgery, (3) ¼ day 1 after surgery, (4) ¼ day 3 after surgery, and (5) ¼ day 7 after surgery. The figure shows the changes in percent from baseline (100%).

studies comparing RFA and ECT in the same model, as we did here. An analysis of the levels of AST and ALT, which are enzymes associated with liver parenchymal cells, suggests that ECT causes more marked local liver destruction than RFA. An analysis of alkaline phosphatase (AP) and gamma-glutamyl-transferase (GGT), which are parameters of cholestasis, did not show significant differences. Bilirubin levels, however, were found to be elevated immediately after RFA. Figs.6 and 7.

Figure 7. Time points for blood collection: (1) ¼ before surgery, (2) ¼ immediately after surgery, (3) ¼ day 1 after surgery, (4) ¼ day 3 after surgery, and (5) ¼ day 7 after surgery. The figure shows the changes in percent from baseline (100%).

R. Czymek et al. / Surgical Oncology 21 (2012) 79e86

The fact that we found post-procedural leukocytosis immediately and on day 1 after ECT when compared with RFA and the control group possibly indicates that electrochemical treatment is associated with a more marked inflammatory response, which in turn might induce an activation of the immune system. Our analyses showed a significant difference in C-reactive protein (CRP) levels between the control group and the treated groups but not between the ECT and RFA groups. Teague et al. [38], who performed ECT in pigs, reported that CRP levels were raised in ECT and control animals at 24 h after ECT. Whereas CRP levels remained increased at 72 h after ECT as they did in our analyses, they dropped in the control group at this time point. As expected, there were no significant differences between haemoglobin, erythrocyte and haematocrit levels since no procedure was associated with notable bleeding. In a study involving 16 pigs that underwent intrahepatic electrochemical treatment and 4 pigs that served as controls, Teague et al. [38] delivered electric charges between 400 and 1000 coulombs. When they measured cytokine levels (IL-1b and TNF-a), they found that these levels were raised between 24 and 72 h after surgery. There was, however, no significant difference between the control and ECT groups. The authors concluded from these measurement results that a systemic activation of the immune system did not occur. This is confirmed by our analyses of IL-1b, TNFa and IL-6, which did not suggest an activation of the immune system in response to ECT. None of the three cytokines was significantly elevated. The activation of leukocytes which we observed in our study was possibly induced by mediators that were not assessed (e.g. IL-8, intercellular adhesion molecule-1 (ICAM-1), etc.).

Conclusions In the majority of cases, intrahepatic RFA in vivo leads to incomplete necrosis in proximity to vessels and the presence of histologically intact perivascular cells even seven days after treatment. Without a reduction in liver perfusion, the central application of RFA should be considered problematic. ECT is a safe alternative. It is not associated with a heat sink effect but has the disadvantage of long treatment times. Whereas RFA has the advantage of shorter treatment times, ECT is cost saving since the electrodes are reusable. Conflict of interest Hereby we state that the final manuscript (Intrahepatic radiofrequency ablation versus electrochemical treatment in vivo) has been approved by all authors. We do not have any direct or indirect financial incentive associated with publishing the article nor any other conflict. The manuscript is not under consideration by another journal and has not been previously published.

Authorship statement Guarantor of the integrity of the study: Czymek Ralf. Study concepts: Bruch Hans-Peter. Study design: Hildebrand Philipp. Definition of intellectual content: Kleemann Markus. Literature research: Czymek Ralf. Experimental studies: Nassrallah Jan. Data acquisition: Schmidt Andreas. Data analysis: Limmer Stefan. Statistical analysis: Gebhard Maximilian. Manuscript preparation: Czymek Ralf, Kleemann Markus. Manuscript editing: Limmer Stefan, Bruch Hans-Peter. Manuscript review : Hildebrand Philipp, Czymek Ralf.

85

References [1] Mulier S, Ni Y, Jamart J, Ruers T, Marchal G, Michel L. Local recurrence after hepatic radiofrequency coagulation: multivariate meta-analysis and review of contributing factors. Ann Surg 2005;242:158e71. [2 Curley SA, Izzo F, Delrio P, Ellis LM, Granchi J, Vallone P, et al. Radiofrequency ablation of unresectable primary and metastatic hepatic malignancies e results in 123 patients. Ann Surg 1999;230:1e8. [3] Nordenström B. Biologically closed electric circuits: clinical, experimental and theoretical evidence for an additional circulatory system. Stockholm: Nordic Medical Publications; 1983. [4] Nordenström B. Survey of mechanisms in electrochemical treatment (ECT) of cancer. Eur J Surg 1994;(Suppl. 574):93e109. [5] Xin Y. Organization and spread of electrochemical therapy (EChT) in China. Eur J Surg 1994;(Suppl. 574):25e30. [6] Xin Y. The clinical advance in application of EChT within the past ten years. Preprints from the 2nd International Symposium on electrochemical treatment of Cancer. Beijing. 27e30 Sept. 1998: p. 81e92. [7] Hinz S, Egberts JH, Pauser U, Schafmayer C, Fändrich F, Tepel J. Electrolytic ablation is as effective as radiofrequency ablation in the treatment of artificial liver metastases in a pig model. J Surg Oncol 2008;98(2):135e8. [8] Lubienski A. Radiofrequency ablation in metastatic disease. Recent Results Cancer Res 2005;165:268e76. [9] Pennes HH. Analysis of tissue and arterial blood temperatures in the resting human forearm. J Appl Physiol 1948;1:93e122. [10] Lu DS, Raman SS, Vodopich DJ, Wang M, Sayre J, Lassman C. Effect of vessel size on creation of hepatic radiofrequency lesions in pigs: assessment of the “heat sink” effect. AJR Am J Roentgenol 2002;178:47e51. [11] Goldberg SN, Charboneau JW, Dodd 3rd GD, Dupuy DE, Gervais DA, Gillams AR, et al. International Working Group on Image-Guided Tumor Ablation. Image-guided tumor ablation: proposal for standardization of terms and reporting criteria. Radiology 2003;228:335e45. [12] Horigome H, Nomura T, Saso K, Fujino N, Murasaki G, Karo Y, et al. Percutaneous radiofrequency ablation therapy using a clustered electrode in the animal liver. Hepatogastroenterology 2001;48:163e5. [13] Mulier S, Ni Y, Frich L, Burdio F, Denys AL, De Wispelaere JF, et al. Experimental and clinical radiofrequency ablation: proposal for standardized description of coagulation size and geometry. Ann Surg Oncol 2007;14: 1381e96. [14] Xin Y, Xue F, Zhao F. Effectiveness of electrochemical therapy in the treatment of lung cancers of middle and late stage. Chin Med J 1997;110(5): 379e83. [15] Wang H-L. Electrochemical therapy of 74 cases of liver cancer. Eur J Surg 1994;(Suppl. 574):55e7. [16] Xin Y. The clinical application of electrochemical therapy of malignant tumours. Gen Clin J 1990;6:25e8. [17] Nilsson E, von Euler H, Berendson J, Thörne A, Wersäll P, Näslund I, et al. Electrochemical treatment of tumours. Bioelectrochemistry 2000;51:1e11. [18] Li K, Xin Y, Gu Y, Xu B, Fan D, Ni B. Effects of direct current on dog liver: possible mechanisms for tumour electrochemical treatment. Bioelectromagnetics 1997; 18(1):2e7. [19] von Euler H, Strahle K, Thörne A, Yongqing G. Cell proliferation and apoptosis in rat mammary cancer after electrochemical treatment (EChT). Bioelectrochemistry 2004;62:57e65. [20] Samuelsson L, Olin T, Berg N. Electrolytic destruction of lung tissue in the rabbit. Acta Radiol Diagn 1980;21:447e54. [21] Miklavcic D., Sersa G., Kryzanowski M., Novakovic S., Bobanovic F., Golouh R., Vodovnik L. Tumour treatment by direct electric current: tumour temperature and pH, electrode material and configuration. Bioelectrochemistry and Bioenergetics Book Chapter Vol. 52. 1993; 417e27. [22] Park HJ, Lyons JC, Ohtsubo T, Song CW. Acidic environment causes apoptosis by increasing caspase activity. Br J Cancer 1999;80:1892e7. [23] Smith RD, Eisner DA, Wray S. The effects of changing intracellular pH on calcium and potassium currents in smooth muscle cells from the guinea-pig ureter. Pflugers Arch 1998;435:518e22. [24] Wemyss-Holden SA, Robertson GS, Dennison AR, Vanderzon PS, Maddern GJ. A new treatment for unresectable liver tumours: long-term studies of electrolytic lesions in the pig liver. Clin Sci 2000;98:561e7. [25] Wemyss-Holden SA, Hall P, Robertson GS, Dennison AR, Vanderzon PS, Maddern GJ. The safety of electrolytically induced hepatic necrosis in a pig model. Aust N Z J Surg 2000;70:607e12. [26] Jiao LR, Hansen PD, Havlik R, Mitry RR, Pignatelli M, Habib N. Clinical shortterm results of radiofrequency ablation in primary and secondary liver tumours. Am J Surg 1999;177:303e6. [27] Rossi S, Di Stasi M, Buscarini E, Quaretti P, Garbagnati F, Squassante L, et al. Percutaneous RF interstitial thermal ablation in the treatment of hepatic cancer. AJR Am J Roentgenol 1996;167:759e68. [28] Samuelsson L, Jönsson L, Lamm IL, Linden CJ, Ewers SB. Electrolysis with different electrode materials and combined with irradiation for treatment of experimental rat tumours. Acta Radiol 1990;32:178e81. [29] Rosenberg B, VanCamp L, Trosko JE, Monsour VH. Platinum compounds: a new class of potent antitumour agents. Nature 1969;222:385e6. [30] Fosh BG, Finch JG, Lea M, Black C, Wong S, Wemyss-Holden S, et al. Use of electrolysis as an adjunct to liver resection. Br J Surg 2002;89:999e1002.

86

R. Czymek et al. / Surgical Oncology 21 (2012) 79e86

[31] Bitsch RG, Düx M, Helmberger T, Lubienski A. Effects of vascular perfusion on coagulation size in radiofrequency ablation of ex vivo perfused bovine livers. Invest Radiol 2006;41(4):422e7. [32] Colombo L, Gonzalez G, Marshall G, Molina FV, Soba A, Suarez C, et al. Ion transport in tumors under electrochemical treatment: in vivo, in vitro and in silico modelling. Bioelectrochemistry 2007;71:223e32. [33] von Euler H, Olsson JM, Hultenby K, Thörne A, Lagerstedt A-S. Animal models for treatment of unresectable liver tumours: a histopathologic and ultrastructural study of cellular toxic changes after electrochemical treatment in rat and dog liver. Bioelectrochemistry 2003;59:89e98. [34] Preziosi L. Cancer modelling and simulation. London, UK: CHAPMAN & HALL/ CRC; 2003.

[35] Wissniowski TT, Hansler J, Neureiter D, Frieser M, Schaber S, Esslinger B, et al. Activation of tumor-specific T lymphocytes by radio-frequency ablation of the VX2 hepatoma in rabbits. Cancer Res 2003;63:6496e500. [36] Wemyss-Holden SA, Robertson GSM, Dennison AR, de la M, Hall P, Fothergill JC, et al. Electrochemical lesions in the rat liver support its potential for treatment of liver tumours. J Surg Res 2000;93:55e62. [37] Wemyss-Holden SA, Dennison AR, Hall P, Finch GJ, Maddern GJ. Electrolytic ablation as an adjunct to liver resection: experimental studies of predictability and safety. Br J Surg 2002;89:579e85. [38] Teague BD, Court FG, Morrison CP, Kho M, Wemyss-Holden SA, Maddern GJ. Electrolytic liver ablation is not associated with evidence of a systemic inflammatory response syndrome. Br J Surg 2004;91:178e83.