Radiological findings of porcine liver after electrochemotherapy with bleomycin
Article Category: Research Article
Published Online: Oct 10, 2019
Page range: 415 - 426
Received: Aug 15, 2019
Accepted: Sep 12, 2019
DOI: https://doi.org/10.2478/raon-2019-0049
© 2019 Maja Brloznik, Nina Boc, Gregor Sersa, Jan Zmuc, Gorana Gasljevic, Alenka Seliskar, Rok Dezman, Ibrahim Edhemovic, Nina Milevoj, Tanja Plavec, Vladimira Erjavec, Darja Pavlin, Masa Bosnjak, Erik Brecelj, Ursa Lampreht Tratar, Bor Kos, Jani Izlakar, Marina Stukelj, Damijan Miklavcic, Maja Cemazar, published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.
Surgical resection is the gold standard for the treatment of hepatic neoplasia. However, the majority of hepatic tumors are unresectable at the time of diagnosis. In these patients, alongside systemic chemotherapy, different local ablative techniques, such as radiofrequency and microwave ablation, are used.1, 2, 3 However, these two techniques are not optimal for tumors located near major bile ducts and larger hepatic vessels due to the heat sink effect or potential damage to vital structures.2, 3, 4 Hence, novel techniques, such as irreversible electroporation (IRE) ablation and electrochemotherapy (ECT), are advantageous for patients with unresectable hepatic neoplasia adjacent to major bile ducts and vessels.1, 2, 3, 5 ECT is a method for the delivery of chemotherapeutic drugs by reversible electroporation, which enables the entry of otherwise poorly or nonpermeant exogenous molecules into cells by the application of short high-intensity electric pulses that induce reversible cell membrane permeabilization.6, 7, 8 ECT is nowadays widely used in European centers for the treatment of cutaneous tumors.9, 10
Clinical studies have shown that ECT is safe and effective also for the treatment of liver metastases and hepatocellular carcinoma (HCC)3, 4, 6, 7, 11, 12, 13, and also other deep-seated tumors like pancreatic carcinoma.14, 15, 16 Potentially hazardous electrode insertion into the lumen of major hepatic vessels has been reported in a study with percutaneous ECT of portal vein tumor thrombosis in patients with HCC.11 From ECT3, 4, 6, 7, 11, 12, 12, and IRE17, 18, 19, 20, 21, 22 studies, it is assumed that ECT of large hepatic vessels is safe, but this assumption has not yet been demonstrated with early radiologic examinations, which could reveal possible intra-abdominal hemorrhage or thrombus formation that could prove fatal for the patient.
Radiologic findings after IRE of hepatic tissue have been thoroughly described and used to determine the area that was irreversibly electroporated22, 23, 24, 25, 26, while there has only been one report of radiologic findings after ECT of liver, where ultrasonographic (US) changes in hepatic tumors were described as indicators of adequate electric field tumor coverage for effective ECT.27 The effect of ECT on healthy hepatic parenchyma and large hepatic vessels has not been previously studied by diagnostic imaging methods. The aim of this study was to characterize radiologic findings after ECT of large hepatic vessels and hepatic parenchyma in a porcine model, and hence to confirm the safety of the procedure.
In this prospective animal model study with regulatory approval issued by the National Ethics Committee at The Administration of the Republic of Slovenia for Food Safety, Veterinary, and Plant Protection (Approval number: U34401-1/2017/4, Approval date: 17.03.2017), nine female grower pigs, purchased from an authorized swine breeder (Globocnik, Sencur, Slovenia) 3–17 days before experiment, were included.28 Experimental animals were reared according to the European Council directive for minimum standards for the protection of pigs (2008/120/EC). All procedures complied with relevant national and European legislation (2010/63/EU).
Animals were 12 weeks old, weighed 31 ± 2.5 kg, and their liver volume estimated from CT images was 865 ± 85 cm3.
During open surgery, four US-guided electropo-rated regions were created.
The treated area electric field was calculated by the finite element method using the software package Comsol Multiphysics (COMSOL AB, Stockholm, Sweden) with MATLAB (Mathworks, Natick, MA, USA).29, 30 The electrical parameters and bleomycin dosage were consistent with the European standard operating procedures for the ECT protocol and ECT for colorectal liver metastases clinical trials.3, 31, 32
The animals were euthanized at two or seven days after ECT/EP using 3 ml/10 kg i.v. T61 euthanasia solution (Intervet, Boxmeer, Netherlands); the liver was explanted for histologic analyses, which were reported in a separate paper.28
Ultrasonographic examinations (US) were performed immediately and up to 20 minutes after the procedure. Different US machines were used, as available: Resona 7 and/or M9 (Mindray, Shenzhen, China) and/or Logiq S7 Pro (GE, Milwaukee, WI, USA). US examinations included B-mode to provide information about echogenicity and echotexture changes, Doppler (color and pulse-wave) to examine vessel patency and contrast enhanced ultrasound (CEUS) to evaluate the perfusion of hepatic parenchyma. The region of CEUS investigation was chosen according to the available US machine. In the case of the Mindray machine, a linear probe was used and the near-field area was preferred, whereas with the GE machine, the far-field was chosen since contrast works only with a convex probe. CEUS was performed before and after EP/ECT. Low mechanical index (<0.1) was applied, and a transpulmonary contrast agent Sonovue (Bracco, Milan, Italy), based on sulfur hexafluoride microbubbles, was administered into the cephalic vein (2.4 ml), and then flushed with saline (2 ml). From the time of contrast application, a 90-second cine-clip was made for further image analysis.
CT of the liver was performed before and at 60 to 90 minutes after EP/ECT with a Somatom Scope CT scanner (Siemens, Erlangen, Germany) in seven pigs. In two pigs treated with ECT with variable linear geometry electrodes, CT was also performed at 1 week after ECT. The following CT parameters were used: referenced 170 mA, 110 kVp, 3.0 mm slice reconstruction thickness, 2.0 mm reconstruction increment, and beam pitch of 1.4. The dynamic study with the contrast medium iopromide (Ultravist; Bayer, Leverkusen, Germany, 0.5 ml/ kg), which was administered intravenously with a Missouri dual chamber injector pump (Ulrich medical, Ulm, Germany) at a velocity of 3 ml/s, consisted of 5 phases: pre-contrast, arterial with bolus tracking in abdominal aorta, and 3 subsequent phases in 30-second intervals (at 30, 60 and 90 seconds after arterial phase).
All radiologic findings were interpreted in consensus by three radiologists (M.B., N.B., R.D.) with more than 10-years of experience in liver imaging.
CEUS perfusion curve, presenting the signal intensity, was analyzed with the US machine built-in software.
CT images were evaluated with a free and open source software program Horos (
Histologic examinations were performed by an experienced pathologist (G.G.) with more than 10-years of experience in surgical pathology. To provide a comparison of radiologic and histologic data, immediate core biopsies of ECT-treated and untreated areas were performed in two pigs treated with ECT using linear geometry electrodes. In the treated parenchyma, two biopsies located between and oriented parallel to the electrodes were collected: one adjacent to the electrode and one 1 cm away from the electrode to the middle of the treated area. Histologic samples were fixed overnight in 10% buffered formalin, embedded in paraffin, cut into 3- to 4-μm-thick sections and stained with hematoxylin and eosin (H&E).
Statistical package computer program SPSS (SPSS Inc., Chicago, Illinois, USA) was used. For analysis of hepatic parenchyma contrast enhancement, the Shapiro-Wilk test was utilized to test the normality. Since the variables were not normally distributed, the Mann-Whitney U test and the Kruskal Wallis H test were used. Statistical significance was defined as P ≤ 0.05. For other data in this study, descriptive statistics were applied.
EP were delivered only after US confirmation of electrode position (Figure 1A) .Immediately after
Figure 1
B-mode ultrasonography.
the delivery of EP and removal of the electrodes, hyperechogenic microbubbles were observed along the electrode tracks (Figure 1B and C) .In the next minutes, the treated parenchyma became hypoechogenic, and the hyperechogenic microbubbles started to diffuse (Figure 1D); the area of hypoechogenicity was larger than the area between the electrodes. In the next 5 to 10 minutes, the hyperechogenic microbubbles diffused through the treated area. There was no obvious difference between EP and ECT with variable linear geometry electrodes, while in the two cases of ECT with fixed hexagonal geometry electrodes, the hypoechogenicity was less evident compared to that with variable linear electrodes. After 10 minutes, the hypoechogenicity started to fade, and in the case of the treatment with hexagonal electrodes, it was no longer visible. The hypoechogenicity of the parenchyma was in contiguity to the treated vessel; however, the vessel wall appeared intact. There was no hemorrhage observed. Furthermore, the patency of the vessel was normal; there were no thrombi, stenotic lesions or extravasation noted, and the Doppler examination showed laminar flow (Figure 2A) .CEUS showed that the perfusion of the treated area was significantly decreased (Figure 2B, C and D) .The area of the hypoperfused parenchyma was larger than the area between the electrodes (Figure 2B),
Figure 2
Doppler and contrast-enhanced ultrasonography.
Dynamic contrast enhanced CT at 60 to 90 minutes after EP/ECT showed subtle hypoattenuating electrode tracks in the pre-contrast and arterial phases (Figure 3A and B), while in the 3 subsequent phases, hypoenhancing areas of treated hepatic parenchyma were noted (Figure 3C, D and E) .These areas were most clearly observed at 30 or 60 seconds after the arterial phase, and the phase with most evident hypoenhancing areas was used for further evaluation. As can be seen in Table 1, different number of hypoenhancing areas were measured in each animal. In the case of the treatment with linear geometry electrodes, the hypoenhancing regions were larger than those in the case of the treatment with hexagonal geometry electrodes (Figures 4A, B, 5A and 5B) .An example of the multiplanar reconstruction (MPR) of one of the areas treated with variable linear geometry electrodes is presented in Figure 4C .All treated vessels were patent with no evidence of thrombosis (Figure 6A) .Hypoenhancing areas were in contiguity with the treated vessels. There was no narrowing of the treated vessels, vessel wall and patency were not affected. No contrast extravasation was identified from large vessels in which the electrodes were inserted.
Figure 3
Dynamic CT study, area treated with hexagonal electrodes is circled; note that hypoenhancing areas are most clearly seen 30 seconds after arterial phase.
Area and attenuation of hypoenhancing regions for each animal
| Number of measured hypo- enhancing regions | Total area (in mm2) | Median area (in mm2), IQR | A Attenuation of total area (in HU) Σ ((area x attenuation) / total area) | B Attenuation of untreated parenchyma (in HU), ± SE 15 measurements in each animal | Difference in attenuation (B – A) | Corrected difference in attenuation (B–A)/B*100 | ||
|---|---|---|---|---|---|---|---|---|
| 48 | 8454 | 94.5, 46–168 | 72 | 118 ± 2 | 46 | 39 | ||
| ECT | 66 | 3178 | 31.5, 19–53 | 86 | 114 ± 2 | 28 | 25 | |
| Linear | 91 | 5146 | 39, 25–75 | 82 | 123 ± 1 | 41 | 33 | |
| electrodes | EP | 61 | 4538 | 51, 30–104 | 75 | 108 ± 1 | 33 | 30 |
| 55 | 3563 | 32, 18–92 | 122 | 165 ± 3 | 43 | 26 | ||
| 25 | 750* | 21, 11.5–32* | 98 | 124 ± 2 | 26 | 21 | ||
| Hexagonal electrodes | ECT | 46 | 2328* | 39.5, 25–72* | 80 | 101 ± 1 | 21* | 21* |
| 1 week after ECT | 11 | 99** | 8, 5–11** | 74 | 99 ± 1 | 24** | 25** | |
| with electrodes linear | 13 | 300** | 19, 12–27.5** | 76 | 101 ± 1 | 25** | 25** |
ECT = electrochemotherapy; EP = electroporation; HU = Hounsfield units; IQR = interquartile range; STD = standard deviation; *P < 0.01 compared to groups with variable linear geometry electrodes; **P < 0.01 compared to groups immediately after EP/EC
Figure 4
After careful delineation of the hypoenhancing areas in each of the transverse CT images (Figure 6B), a statistically significant difference between attenuation of untreated parenchyma and attenuation of treated areas was observed in all animals. For each animal, CT measurements and calculations are presented in the Table 1. In the case of the treatment with hexagonal geometry electrodes, the total area of the hypoenhancing regions was smaller than that in the case of the treatment with linear electrodes (P < 0.001), which is in accordance with computer simulation. The results of the computer simulations are shown in Figure 4A and B, where the difference in the shape of the local electric field strength produced by the different electrodes is observed. Figure 4A shows that the treated volume produced by the variable linear geometry electrodes measures 30 mm in width and 20 mm in height (corresponding to the distance between the electrodes and the applied voltage) and
Figure 5
Computed tomography. Hypoenhancing areas after electrochemotherapy
Figure 6
Computed tomography.
Figure 7
Histology of hepatic parenchyma immediately after electrochemotherapy (ECT). Fibrin thrombus in the lumen of a small venule (arrow).
is larger than the treated volume produced by the hexagonal electrodes, which is 20 mm by 20 mm (Figure 4B).
Furthermore, the difference and the corrected difference in attenuation were smaller after treatment with fixed hexagonal geometry electrodes than after treatment with variable linear electrodes (P < 0.001). There was no significant difference in the total area between the EP and ECT with linear electrodes (P = 0.908). In the CT findings at one week after ECT (Figure 6C), a regression of the ECT-induced changes was observed with only small hypoenhancing areas with diameters of up to 6 mm identified, which corresponds to the volume of irreversibly electroporated liver tissue determined by computer simulations, as shown in Figure 4A and B .The total area of the hypoenhancing regions was smaller than that observed immediately after the EP/ECT (P < 0.001). The difference and corrected difference in attenuation were both smaller after one week compared to those at 60 to 90 minutes after EP/ECT (P < 0.001, P = 0.007).
Histologic findings of treated hepatic parenchyma immediately after ECT showed fibrin thrombi in small venules (Figure 7), with no other histologic changes of other vessels and bile ducts or the hepatic parenchyma.
Radiologic findings after EP/ECT of large hepatic vessels and hepatic parenchyma were characterized in a porcine model, which was selected due to anatomical and physiological similarities with the human liver.1, 33 The results showed decreased perfusion in the treated area. This finding was an anticipated result since EP/ECT induces a local blood flow modifying effect or ‘vascular lock’ characterized by the vasoconstriction and increased wall permeabilization of small blood vessels. The effect on perfusion is shorter in EP compared to that in ECT and shorter in healthy compared to tumor tissue, which is known as the ‘vascular disrupting effect’. Chemotherapeutic drugs are cytotoxic to endothelial cells, especially neoplastic endothelial cells, and this effect prolongs decreased perfusion.34, 35, 36, 37, 38, 39, 40, 41 In our case, there was no difference between EP and ECT, and no vascular disrupting effect was observed in healthy hepatic parenchyma28, which confirms that bleomycin at the doses used has a negligible effect on healthy tissue.34, 35, 36
All radiologic modalities showed healthy vessel walls and patency, despite direct electrode insertion into the lumen of major hepatic vessels. These findings were consistent with previously published histologic results: no thrombosis was identified at two and seven days after EP/ECT in healthy liver.28 This result was an expected finding considering ECT3, 4, 8, 11, 12, 13 and IRE17, 18, 19, 20, 21, 22 clinical studies, where electrodes were inserted in the vicinity of3, 4, 8, 11, 12, 13, 17, 18, 19, 20, 21, 22 or into major hepatic vessels.13 The absence of bleeding, even if needles are inserted deep into the hepatic parenchyma, is an important safety aspect of ECT due to the transient local hypoperfusion and possible electrocoagulation related to the high current density at the surface of needle electrodes.29, 35
The US findings were dynamic. Hyperechogenic bubbles, which initially form around the electrodes, are a consequence of electrochemical reactions on the electrodes and electrocoagulation of the tissue.27, 29 The liberated gases are chlorine at the anode and hydrogen at the cathode.42, 43 Gas bubbles are formed in RFA44 and IRE18, 45 ablations. Hypoechogenicity of the treated parenchyma indicates a structural change, which presumably occurs due to decreased perfusion caused by the vasoconstriction and increased wall permeability (edema) of small vessels.34, 35, 36, 37 The histologic findings of the treated hepatic parenchyma immediately after ECT showed fibrin thrombi in small venules, which is consistent with decreased perfusion due to vessel spasms.
Decreased perfusion of the treated areas was confirmed with CEUS and dynamic CT studies, the latter proving the decreased contrast enhancement of the treated areas at 30 to 90 seconds after arterial phase. The area of decreased perfusion was smaller after treatment with fixed hexagonal electrodes than after treatment with variable linear geometry electrodes. This effect is due to a larger distance between electrodes in the case of variable geometry electrodes where higher voltage must be applied to achieve same therapeutic effect, and is in accordance with computer simulation and histology.28 The difference between the two geometry electrodes can be ascribed to a higher local electric field strength adjacent to the electrodes in the case of linear electrodes46, as shown in Figure 4A and B .Despite the difference in size of hypoperfused hypoenhancing areas due to the higher local electric field, there is no difference in efficacy of electroporation between the variable linear- and fixed hexagonal-geometry electrodes.29, 46 In routine clinical practice hexagonal electrodes are used for smaller and superficial liver tumors, while linear electrodes are used for deep-seated and larger liver tumors.47 The subtleness of the radiological changes of liver after EP/ECT agrees with laboratory and histologic findings that the procedure is safe, while on the contrary, it could indicate that CT findings might not be a good indicator of procedure efficacy in healthy liver. The vascular structures of the portal spaces as well as branches of the hepatic and portal veins in the liver parenchyma display different changes depending on their size and position in the ablated area: those situated in the central parts of the ablated areas and close to the electrodes show complete necrosis. Smaller structures are more sensitive to electroporation than larger structures with arterioles and bile canaliculi more resistant than venules.12
The CT studies at one week after ECT showed only small hypoenhancing areas, which is in accordance with histologic studies showing the existence of scar tissue.12, 28 Where electrodes punctured the wall of the large vessels, the architecture of the vessel wall was effaced with missing endothelium and no thrombosis present.28
Our study has several limitations. Due to the time limit of open surgery, there was limited time for US examinations, and the CT studies were performed in a range of 60 to 90 minutes after the EP/ ECT. Performing radiologic examinations at different times was a major limitation because radiologic findings were dynamic and evolved in time. Another limitation of this study is the use of various US machines, which precluded numerical and statistical analyses of the US findings. Different heart rates and blood pressures of pigs influenced CEUS assessment of perfusion; therefore, comparisons among animals would be challenging. Another limitation of the study was that CT studies were only performed at 1 week after ECT with variable linear geometry electrodes. Furthermore, in CT studies, various treated areas could not always be differentiated from other areas. This limitation was overcome in the study with the percutaneous ECT of portal vein tumor thrombosis13 and in a study with IRE of porcine liver22 with a coaxial angiocatheter to define electrode orientation and position relative to the ablation zone. Only healthy liver was studied, further investigation of other tissues, particularly tumor tissue is required, because it is reasonable to expect that radiologic findings after ECT of hepatic neoplasia differ from radiologic findings after ECT of healthy liver due to differences in vascular and extracellular spaces. Furthermore, relatively small number of animals has been investigated, which prevented further statistical analyses and a better correlation between different groups.
Radiologic findings after EP/ECT of porcine liver did not show clinically significant damage to large liver vessels and parenchyma; intact vessel walls and patency were observed, the hepatic parenchymal changes indicated by US hypoechogenicity and CT hypoenhancement were subtle. Histologic changes immediately after, and 2 and 7 days after treatment were in accordance with radiologic findings, and these results confirm that ECT is safe for the treatment of tumors that are adjacent to large hepatic vessels. Notably, radiologic features after EP/ECT are dynamic, and further studies are required to thoroughly investigate these features to provide definite answers, which, if any, are useful as indicators of adequate electric field distribution and as possible predictive factors that could guide decisions regarding the course of further treatment.