You are on page 1of 6

Inducing Thermal Lesion on a Cholangiocarcinoma Considering a SalineEnhanced Radiofrequency Ablation

Department of Electrical Engineering and Computer Science University of Coimbra, Portugal (2) RIANDA Research Centro de Investigao em Energia, Sade e Ambiente Coimbra, Portugal lemos.antunes@apdee.org
Abstract-Cholangiocarcinoma is an adenocarcinoma of the bile ducts which drain bile from the liver into the small intestine. Unfortunately, most patients are diagnosed on an advanced stage of the disease with almost no chances for surgery, the only potentially curative treatment. As nitinol stents can be used to reduce stricture problems of the bile duct, these can be also considered as potential electrodes for hyperthermia treatments. Previous works show that in fact these metallic stents might be used as part of a feasible solution for delivering radiofrequency (RF) energy into a tumor located in a hollow organ to destroy the tumor tissue. However the tissue lesion induced is not completely uniform due to convective heat transfer associated to the blood flow in the nearby vessels. In this paper it is studied the use of saline solution for modifying the electrical conductivity of the tissue in order to obtain a more uniform lesion. A numerical analysis using finite element method on a simplified model of the porta hepatis is performed. Results show that it is possible to obtain a more regular volume, by this way the tumor tissue is preferentially heated, although there are still some risks on exceeding the dimension of the bile duct.

Carlos L. Antunes(1, 2), Tony R. Almeida(1) and Nlia Raposeiro(2)


(1)

a radiofrequency ablation using a stent-based electrode considering an infused saline solution in the tumor tissue used to modify its electrical properties so a more regular lesion can be achieved. II. MATERIAL AND METHODS A. The Bioheat Equation The radiofrequency ablation procedure consists of heating up the tissue in order to destroy it by converting electric energy into thermal energy. The current flows from the active electrode through the tissue to a return electrode pad usually placed on the back or the upper leg of the patient. The numerical simulation of the models consists of the analysis of a thermoelectrical coupled field problem. The temperature at each point of the tissue can be expressed by the bioheat equation (1):

I. INTRODUCTION Liver cancer has a very poor prognosis, being the number of deaths almost the same as the number of new cases. It is therefore the third most common cause of death from cancer [1, 2]. Cholangiocarcinoma is a malignant cancer arising from the neoplastic transformation of the epithelial cells lining the intra-hepatic and extra-hepatic bile ducts, and it is the second most common primary hepatic malignancy [3]. Because there are no early symptoms, the majority of patients are diagnosed at advanced stage, when surgical therapies are excluded [4]. As nitinol stents can be used to reduce stricture problems of the bile duct, these can be also considered as potential electrodes for hyperthermia treatments. Previous works [5-7] show that nitinol stents can be considered as a feasible solution for delivering radiofrequency (RF) energy in a tumor located in a hollow organ. However, the tissue lesion induced using this kind of electrode is not completely uniform due to the convective heat transfer associated to the blood perfusion on the portal vein and hepatic artery. Also it was verified the tissue next to the electrode ends are preferentially heated which also contributes to obtain a non-uniform lesion[8]. In order to overcome this situation, it was considered the possibility of modifying the properties of the biological tissue, particularly the electrical conductivity in the middle section of the tumor. It has been demonstrated that it is possible to increase RF tissue heating during a RF ablation procedure by injecting a saline solution thus modifying the electrical conductivity, energy deposition and heating of the tissue [911]. In this work is intended to perform a numerical analysis of

T = k T + J E hb (T Tb ) Qm t

(1)

where is the density [kg/m3], c is the specific heat [J/kgK], T is the temperature [K], Tb is the blood temperature [K], k is the thermal conductivity [W/mK], J is the current density [A/m], E is the electric field intensity [V/m], Qm is the energy due to metabolic process [W/ m3] and hb is the blood perfusion convective heat transfer coefficient. The energy generated by the metabolic process can be neglected since it is very small. Also, the term hb(T Tb) which refers to blood perfusion is neglected due to the presence of the porta vein and the hepatic artery. The blood temperature in these large blood vessels is considered unaffected by the thermal field in the surrounding tissue [12] and the blood flow is considered as a moving heat sink which adds the following contribution to the right hand of (1):

b Cb ub T

(2)

where b is the blood density [kg/m3], Cb is the blood specific heat [J/kgK] and ub is the velocity of the blood [m/s]. Most commercial generators of radiofrequency ablation work between 375 to 480 kHz. At this frequency range most part of the energy dissipated by the electric probe is through electrical conduction and so quasi-static approximation is valid [13]. The electrode energy deposition in (1) due to Joule loss can be calculated considering a RF voltage is applied between the stent and the return pad. The resulting voltage through the domain obeys Laplaces equation:

V = 0

(3)

where corresponds to the electrical conductivity [S/m] and V to the electric potential [V]. At each iteration, equation (3) is evaluated in order to calculate the distributed heat source JE to be used in (1) plus the contribution (2). Then temperature distribution is calculated and the tissues electrical conductivity, which is temperature-dependent, is recalculated. The steps during the solution of the finite element model started at 0.01s and they were subsequently and automatically controlled by the solver software. B. Model Geometry In this work it was considered a simplified 3D model of the porta hepatis. The porta hepatis is a transverse fissure of the liver where the portal vein and the hepatic artery enter the liver and the bile duct leaves. Cholangiocarcinoma can occur anywhere along the intrahepatic or extrahepatic biliary tree, and approximately 60% to 80% of cholangiocarcinomas encountered are located in the perihilar region [14]. The 3D models were created considering an external cylinder (liver) with 200 mm diameter and 100 mm height. The bile duct and the portal vein are cylinders of radius 5 mm and the hepatic artery is a cylinder of radius 2 mm [15]. The portal vein and the bile duct are positioned on a circumference of 6 mm radius separated by an angle of 120. The hepatic artery is located so the distance between the three ducts is the same. Fig. 1 shows the position of the blood vessels and the bile duct considering a circumference of r = 6 mm.

TABLE I MATERIAL PROPERTIES USED IN SIMULATION [7, 16, 17] Element Material [kg/m3] c[J/kgK] k[W/mK] Electrode Nitinol 6450 840 18 Hole Air 1.202 1 0.025 Tissue Liver 1060 3600 0.512 Tumor Tumor 1060 3600 0.512 tissue Blood Blood 1000 4180 0.543 vessels

[S/m] 1108 0 l(T) t(T) 0.667

TABLE II BLOOD VESSELS PROPERTIES USED IN SIMULATION [15] Blood Vessel Diameter [mm] Blood Perfusion [ml/min] Vena Cava 10 327.55 Hepatic Artery 4 20.5

The electrode is made up of 24 nitinol wires with 0.25 mm diameter. Each wire is a helix of radius 2 mm with a pitch of 25 mm. The whole electrode is 40 mm long placed inside the tumor. The whole model using across all simulations is presented in Fig. 2. C. Material Properties The material properties required for solving the models considered in this work were obtained from the literature [7, 15-17] and are summarized in Table I and II. For the electrical conductivity of the outer sections of the tumor it was considered a value of 0.269S/m [17]. The middle section corresponds to the tumor volume with a saline solution, so its electrical conductivity is increased and it is considered a multiple of the electrical conductivity of the outer sections, i. e.,

tc = kste

(4)

Fig. 1 Location of the blood vessels and the bile duct.

where tc is the electrical conductivity of the middle section of the tumor, te is the electrical conductivity of the tumor ends and ks is a proportional factor. In the present work it was considered ks varying from 1 (no saline solution) to 5. Finally, the electrical conductivity of liver tissue is 0.13 S/m [7]. The electrical conductivities of the healthy and tumor tissues were considered temperature-dependent, increasing 2% per degree Celcius, dropping to 0.01S/m above 100C, allowing this way to simulate the electrical insulation verified when gas forms at this temperature value [18]. D. Model Conditions Numerical simulations were performed in all models considering constant source voltages of 20, 22, 24, 26, 28 and 30 volts. The external boundary of the model was considered at ground potential (zero volts). The temperature at the external surfaces of the model was set to 37C. The initial temperature of the tissue was set also to 37C. Also the blood temperature was set to 37C.

Fig. 2. Model considered for numerical simulation

The tumor is represented by a tube of 40 mm length, 5 mm radius and 3 mm thickness, placed in the middle of the bile duct. The tumor volume was cross-section divided into three parts for simulation volume regions with different electrical conductivities. The center section of the tumor is 15 mm long and the tumor ends are 12.5 mm long.

E. Software The stent structure was created in AutoCAD and exported in 3D ACIS format to COMSOL Multiphysics 4.1 (COMSOL, Inc. Burlington, MA, USA). The remain model was created within Comsol, which was also used for 3D finite element analysis. All models were solved with PARDISO solver considering

Fig. 3. Volume of lesion obtained for an applied voltage of 20V considering isothermal surfaces of 50C.

Fig. 6. Volume of lesion obtained for several voltage values considering isothermal surfaces of 60C (ks=1).

Fig. 4. Volume of lesion obtained for an applied voltage of 20V considering isothermal surfaces of 60C.

Fig. 7. Volume of lesion obtained for several voltage values considering isothermal surfaces of 60C (ks=3).

Fig. 5. Volume of lesion obtained for an applied voltage of 30V considering isothermal surfaces of 60C.

Fig. 8. Volume of lesion obtained for several voltage values considering isothermal surfaces of 60C (ks=5).

a RF ablation procedure of 180 seconds. Each model took an average time of 12.5 hours to solve using a computer with a Intel Core 2 Quad CPU @ 2.34Ghz, with 8Gb of RAM, on a 64 bits platform (Windows Vista).
III.

RESULTS AND DISCUSSION

Considering that cellular cytotoxicity is induced in 4 to 6 minutes for temperatures from 46C up to 50-52C, and that there is near instantaneous irreversible cellular damage above 60C [19, 20], isothermal surfaces of 50C and 60C were considered for analysis of the volume of lesion induced by the

radiofrequency thermoablation procedure. Taking into account the time interval simulated 180s it is expected that the volume included in the 60C isothermal surface represents a higher probability of inducing a volume of damaged tissue. After each simulation, 50C and 60C isothermal surfaces were obtained and the volume contained in each of these surfaces was calculated using Comsol. In Fig. 3 and Fig. 4 are presented the volumes obtained for an applied voltage of 20V, considering both volumes defined by 50C and 60C isothermal volumes, respectively. In both graphs it is shown that the volume obtained is larger as the electrical conductivity of the middle section of the tumor tc

15s

30s 60s 120s Fig. 9. Temperature distribution for an applied voltage of 20V (ks = 3).

180s

increases. At some instant, it is possible to observe that the lesion volume does not increase significantly with time: the lower the value of ks the sooner the volume stops increasing considerably. This can be explained with the sudden decrease of the electrical temperature of the tissue as soon as the temperature reaches 100C. From this point, the electrical current decreases and so the tissue is no longer significantly heated, which leads to a steady volume lesion. As the value of ks decreases, the tissue surrounding the electrode heats up more quickly. This leads to an earlier electrical isolation of the electrode and so a smaller volume of damaged tissue is obtained. This was verified for both volumes delimited either by a 50C isothermal surface or by a 60C isothermal surface at every value of voltage considered. This can be observed in Fig. 5 which depicts the volumes obtained for an applied voltage of 30V considering a 60C isothermal surface. From Fig. 3 to Fig. 5 it can be also observed that, at a first stage, the volumes obtained are very similar regardless of the value of ks. Later, these values diverge with time, obtaining larger volumes as the value of ks increases. This can be verified for volumes obtained from isothermal surfaces of 50C and 60C. Also, the higher the value of the voltage applied the sooner the values of damaged tissue volume diverge. For example, for an applied voltage of 20V, considering a induced lesion volume delimited by a 60C isothermal (Fig. 4), it can be observed that up to 80 seconds from the beginning of the simulation the induced lesion volume obtained is almost identical for every value of ks. After that, the volumes obtained diverge. Increasing the voltage to 30V (Fig. 5) these values of volume begin to diverge after 20 seconds. In Fig. 6 to Fig. 8 it is depicted the different values of volume delimited by a 60C isothermal surface obtained for a constant value of ks considering different applied voltages. These graphs clearly show that the amount of damaged tissue is smaller for larger values of voltage. On the other hand, as ks increases, the volume of damaged tissue also increases, as it was stated before. It is therefore necessary to set a compromise between the applied voltage and the enhanced electrical conductivity of the tissue. At this point not only the size of the volume obtained is important but also its shape is clearly an important factor to take into account. It is important to achieve a regular volume so the tumor tissue is preferably damaged. As it was mentioned before, the values of volume obtained are very close at an initial time interval. Although these are

15s

30s

60s

120s

180s

Fig. 10. Volume of lesion considering an isothermal surface of 60C at 20V (ks = 1).

15s

30s

60s

120s

180s

Fig. 11. Volume of lesion considering an isothermal surface of 60C at 20V (ks = 2).

15s

30s

60s

120s

180s

Fig. 12. Volume of lesion considering an isothermal surface of 60C at 20V (ks = 5).

15s

30s

60s

120s

180s

Fig. 13. Volume of lesion considering an isothermal surface of 60C at 30V (ks = 2).

1s 15s 30s 45s Fig. 15. Current density (norm) obtained for an applied voltage of 30V with ks = 3.

15s

30s

60s

120s

180s

Fig. 14. Volume of lesion considering an isothermal surface of 60C at 30V (ks = 5).

initially very alike, the shape of the volumes obtained differs. In Fig. 10 to Fig. 12 are depicted the volume obtained considering a 60C isothermal surface at 20V, for several values of ks. In these images it is possible to observe that at time instant of 15 and 30 seconds, the volumes obtained are almost identical, which agrees with the information attained from the graph in Fig. 4. However, at time instant of 60 seconds the shape of the volumes obtained are different. The main difference is in the middle portion of the volume. As ks increases, the volume grows thicker on the opposite side to the blood vessels. Yet, it does not develop in the same manner on the side next to the blood vessels. The tumor tissue on this side takes much longer to heat due to convective heat transfer in the vicinity of the blood vessels. As time elapses, the volume keeps growing on the opposite side of the blood vessels, becoming bigger for larger values of ks. Same results can be observed for larger values of voltage. For example, Fig. 13 and Fig. 14 show the volume shapes obtained for an applied voltage of 30V. In this case, at time instant of 30 seconds, the shape of the volumes obtained for ks = 2 and ks = 5 are almost identical but they differ at time instant of 60 seconds. From Fig. 9 to Fig. 14 two things become evident: 1) the tumor tissue ends are preferably heated at first; and 2) later, the middle part of the volume is heated more intensely, obtaining in this region a thicker volume lesion as the applied voltage increases. As soon as the voltage is applied, a large current density is attained, especially at both ends of the electrode which

causes the high heating of the ends of the tumor tissue (Fig. 15). The ends of the tumor model confront the air volume located above and below it. This leads to a larger energy deposition at these points than in the middle portion of the tumor. As the tissue is heated up over 100C, the electric conductivity decreases as well as the electrical current. The electrical current becomes very low at the ends of the tumor at first and it finally drops significantly all over the tissue (time instants of 15, 30 and 45 seconds in Fig. 15), leading to a slow damaged tissue growth with time, as stated in Fig. 3 to Fig. 8. Another important observation is related to the voltage applied and the shape of the volume obtained. As it was already mentioned, the volume of damaged tissue decreases as the applied voltage increases. However, the volumes obtained for higher voltages are more regular. As the applied voltage rises the tissue is heated more rapidly, this way overcoming the convective heat transfer due to the blood vessels. Finally, one last remark about the volumes obtained when considering a 60C isothermal surface. In Fig. 6 to Fig. 8 it is possible to observe that the curves obtained for the values of damaged tissue present an overshoot: the volume grows with time reaching a peak, then the value slightly decreases during the rest of the simulation, except for the case where ks = 1 (no saline solution). For this case the value of the volume obtained decreases during a brief time interval. After this the volume resumes increasing very slowly. When the electric conductivity drops there is a reduction of energy deposition. The heat transfer due to RF procedure is less than the convective heat transfer due to the blood flow in the nearby vessels and so the tissue is cooled. This can be easily observed in the top view of Fig. 10 to Fig. 14. It is noticeable at first that the isothermal surface increases and then it begins to recede because the tissue next to the blood vessels is being cooled down. Taking into account that there is near instantaneous irreversible cellular damage above 60C, there is a high probability that the damaged tissue is higher than the value obtained after 180 seconds. IV. CONCLUSION The study presented is an overview of the modeling, simulation and analysis of a saline-enhanced radiofrequency tissue thermoablation of a cholangiocarcinoma considering a stent-based electrode. It was intended to obtain a more regular volume of damaged tissue in order to heat and preferentially destroy the tumor tissue, modifying its electric conductivity with a saline solution. Several cases with different electric

conductivities for the tumor tissue at different applied voltages were considered and the results were analyzed taking into account volumes delimited by isothermal surfaces of 50C and 60C. As expected, altering the electric conductivity in the middle section of tumor tissue led to different shapes of volumes of damaged tissue. Two important facts should be highlighted: 1. As the electric conductivity of the middle section of the tumor increases the size of the volume of induced damage also increases; 2. As the applied voltage increases the volume obtained decreases. Besides the dimension of the volume attained, also the shape of the volume is essential. According to previous work [8], the ends of the tumor are rather heated than the middle section of it. Increasing the electrical conductivity excessively might lead, in a first stage, to an irregular shape of volume. On the other hand, higher values of voltage produce more regular shapes. Because it is more important to heat the tumor, we are not really concerned with a large volume of damaged tissue. Instead, it is important to induce a well-located lesion so the bile duct is not damaged as well during the radiofrequency ablation procedure. Combining a relative high voltage while increasing slightly the electrical conductivity of the tumor might have this effect. Finally, the volume obtained is not completely regular as expected and the simulations performed point out that the induced lesion can still exceed the tumor itself, which might damage the bile duct. It should be noticed that, unlike the simplified model considered for the present numerical simulation, the porta hepatis is a more complex structure, with different pulsating blood flows and stroma which was not considered in the present work. Actual work is being performed in order to take into account some of these limitations so a more controlled volume shape can be obtained, preserving the healthy tissue.
V.

ACKNOWLEDGMENT

This work was financially supported by the Foundation for Science and Technology (FCT, Portugal) through the project number PTDC/EEA-ACR/72276/2006.
VI.

REFERENCES

[1] J. Ferlay, H. R. Shin, et al. (2010, June 5th). GLOBOCAN 2008, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 10 [Internet]. Available: http://globocan.iarc.fr/ [2] D. M. Parkin, F. Bray, et al., "Global Cancer Statistics, 2002," CA Cancer Journal for Clinicians, vol. 55, pp. 74-108, March 1, 2005.

[3] M. Gatto, M. C. Bragazzi, et al., "Cholangiocarcinoma: Update and future perspectives," Digestive and Liver Disease, vol. 42, pp. 253-260, April 2010. [4] M. Gatto and D. Alvaro, "New insights on cholangiocarcinoma," World Journal of Gastrointestinal Oncology, vol. 2, pp. 136-145, March 15 2010. [5] C. F. L. Antunes, T. R. O. Almeida, et al., "Thermal Ablation in Biological Tissue Using Tubular Electrode," in 14th Biennial IEEE Conference on Electromagnetic Field Computation, Chicago, USA, 2010. [6] C. F. L. Antunes, T. R. O. Almeida, et al., "Effects of the Geometry of a Tubular Electrode on the Temperature Distribution in Biological Tissue," in 14th Biennial IEEE Conference on Electromagnetic Field Computation, Chicago, USA, 2010. [7] C. F. L. Antunes, T. R. O. Almeida, et al., "A Tubular Electrode for Radiofrequency Ablation Therapy," in ICBET 2010 - International Conference on Biomedical Engineering and Technology, Paris, France, 2010. [8] C. F. L. Antunes, T. R. Almeida, and N. Raposeiro, "Finite Element Modeling of Cholangiocarcinoma Radiofrequency Ablation," presented at the 10th International Conference of the European Bioelectromagnetic Association, Rome, Italy, 2011. [9] M. G. Curley and P. S. Hamilton, "Creation of large thermal lesions in liver using saline-enhanced RF ablation," in 19th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, Chicago, USA, 1997, pp. 2516-2519. [10] Y. Miao, Y. Ni, et al., "Ex Vivo Experiment on Radiofrequency Liver Ablation with Saline Infusion through a Screw-Tip Cannulated Electrode," The Journal of surgical research, vol. 71, pp. 19-24, 1997. [11] S. N. Goldberg, M. Ahmed, et al., "Radio-Frequency Thermal Ablation with NaCl Solution Injection: Effect of Electrical Conductivity on Tissue Heating and Coagulation - Phantom and Porcine Liver Study," Radiology, vol. 219, pp. 157-165, April 1, 2001 2001. [12] J. C. Chato, "Heat Transfer to Blood Vessels (abstract)," Journal of Biomechanical Engineering, vol. 102, 1980. [13] R. Plonsey and D. Heppner, "Considerations of quasi-stationarity in electrophysiological systems," Bulletin of Mathematical Biology, vol. 29, pp. 657-664, 1967. [14] A. Nakeeb, H. A. Pitt, et al., "Cholangiocarcinoma. A spectrum of intrahepatic, perihilar, and distal tumors," Annals of Surgery, vol. 224, pp. 463-475, October 1996. [15] C. Tziafalia, M. Vlychou, et al., "Echo-Doppler measurements of portal vein and hepatic artery in asymptomatic patients with hepatitis B virus and healthy adults," Journal of Gastrointestinal and Liver Diseases, vol. 15, pp. 343-346, 2006. [16] D. Haemmerich, T. Staelin, et al., "Hepatic bipolar radio-frequency ablation between separated multiprong electrodes," IEEE Transactions on Biomedical Engineering, vol. 48, pp. 1145-1152, 2001. [17] D. Haemmerich, S. T. Staelin, et al., "In vivo electrical conductivity of hepatic tumours," Physiological Measurement, vol. 24, pp. 251-260, 2003. [18] D. Haemmerich, L. Chachati, et al., "Hepatic radiofrequency ablation with internally cooled probes: effect of coolant temperature on lesion size," IEEE Transactions on Biomedical Engineering, vol. 50, pp. 493500, April 2003 2003. [19] S. N. Goldberg, "Radiofrequency tumor ablation: principles and techniques," European Journal of Ultrasound, vol. 13, pp. 129-147, 2001. [20] Y. Ni, S. Mulier, et al., "A review of the general aspects of radiofrequency ablation," Abdominal Imaging, vol. 30, pp. 381-400, August 2005 2005.