Determination of the temperature-dependent electric conductivity of liver tissue ex vivo and in vivo: Importance for therapy planning for the radiofrequency ablation of liver tumours

Abstract

Introduction: Knowledge about the changes in the electric conductivity during the coagulation process of radiofrequency ablation of the liver is a prerequisite for the predictability of produceable thermonecrosis in the liver.

Materials and methods: Continuous measurements of the electric conductivity σ in ex vivo porcine liver (n = 25) were done during the coagulation and cooling process at the temperature range of the radiofrequency ablation at a frequency of 470 kHz relevant for the radiofrequency ablation. Measurements of the electric conductivity were performed in both perfused porcine liver (n = 3) and a human surgical specimen from a colorectal liver metastasis.

Results: At a body temperature of 37°C, conductance σ was 0.41 siemens per metre (0.32 S/m; 0.52 S/m). Conductance σ increased continuously and uniformly at a temperature of 77°C. Maximum conductance σ with 0.79 S/m (0.7 S/m; 0.87 S/m) was reached at 80°C. A continuous reduction of conductance was observed during the cooling phase. At 37°C, the specific conductance σ in the healthy perfused porcine liver was 0.52 S/m, 0.55 S/m and 0.57 S/m (mean 0.55 S/m). The electric conductivity of the human colorectal liver metastasis was clearly higher.

Conclusion: Changes in the specific conductivity during the coagulation and the cooling phase play an important role for the produceable size of a coagulation necrosis and necessitates an adaptation of the therapy parameters during radiofrequency ablation.

• impedance
• radio-frequency-induced thermotherapy
• liver metastases

Introduction

In recent years, local in situ ablation procedures have become a permanent feature in the treatment of malignant liver tumours. Especially radiofrequency-induced thermoablation (RFA) is a frequently applied system in these patients 1–6. Besides the treatment of malignant liver tumours, radiofrequency ablation has been established as a commonly applied system for tumour destruction including kidney cell cancer 7, bronchial cancer 8, 9, bone cancer 10, and breast cancer 11. The problem of local ablation is the induction of sufficiently large necrotic zones with safe destruction of the tumour tissue with an adequate security margin. However, there is still no monitoring for intra-interventional assessment of the necrotic zone, which inevitably increases the risk of incomplete tumour destruction. Especially in the treatment for colorectal liver metastases this lack of monitoring during ablation exists. Despite the very good results obtained with radiofrequency ablation in the treatment of small tumours (<3 cm) 12–14, it is necessary to establish radiofrequency ablation as a safe procedure in the treatment of larger tumour in order to expand the clinical range of applications.

The principles of the mechanisms responsible for tissue destruction with the radio waves of radio frequency ablation are well known 15–17. The size of the thermonecrosis is significantly determined by the properties of the treated tissue. Besides the blood flow in the tissue 18, the specific electric conductivity has a special influence on the depth of the heat dispersion in the tissue 19. Some studies have been performed on the tissue conductivity of liver tissue at different frequencies 20, 21. These studies are, however, restricted to a low temperature range. The specific conductivity changes during the coagulation process of radiofrequency ablation. In order to be able to assess the depth of the heat dispersion in the tissue, it is necessary to know the changes in the tissue conductivity in the entire temperature spectrum of the radiofrequency ablation.

Studies on the specific conductivity during the coagulation and cooling process of a thermonecrosis in biological tissue have not yet been performed.

The aim of this study was to examine for the first time the changes in the specific conductance σ during coagulation in the temperature range of the radiofrequency ablation and during the cooling phase in porcine liver tissue. To compare the clinical situation of the radiofrequency therapy of patients with liver tumours, investigations were also done with regard to the specific conductance σ in the perfused porcine liver and tissue of a human liver metastasis.

Materials and methods

Measurements with the 4-needle measurement probe

A 4-needle measuring probe (HIOKI 3532–50) was used to measure the specific conductance. This enabled measurements in vivo at body temperatures and ex vivo during heating of the examined tissue. This method which minimises the effects of polarisation impedance has been commonly used in prior ex vivo and in vivo tissue studies 21, 22. The conductivity measurement system used in this study is described in detail by Tsai et al. 23. In Figure 1A the 4-needle measurement probe connected to the LCR measurement bridge (HIOKI 3532–50 LCR HiTester, Nagano, Japan) is shown, Figure 1B represents the 4-needle measurement probe in detail. In all experiments, calibrating was performed with physiological saline solution at a temperature of 25°C.

Figure 1. (A) Measurement set-up with 4-electrode measurement probe and LCR measurement bridge; (B) 4-electrode measurement probe in detail.

After calibrating the 4-needle measurement probe, the probe was inserted into the tissue to be examined. All measurements were performed at a frequency of 470 kHz relevant for the radiofrequency ablation.

Examinations in porcine liver

In order to examine changes of the specific electric conductivity σ during coagulation of the tissue and cooling phase of the coagulation necrosis, measurement series were done in the ex vivo porcine liver and in vivo measurements in the perfused porcine liver of an anaesthetised pig.

Fresh liver from sacrificed pigs was used for the ex vivo examinations (n = 25). In all examinations, slaughtering and removal of the porcine liver was performed 6–8 hours earlier. From these livers, 10 × 8 × 5 cm sections were cut to provide enough tissue for inserting the 4-needle measurement probe.

The measurements of the 4-needle measurement probe described above would have been disturbed and the measurement values falsified on account of the emission of electromagnetic waves through the application of high-frequency electricity during radiofrequency-induced ablation. Thus it was necessary to use an alternative heating method instead of RFA to achieve tissue coagulation.

Figure 2 shows the measurement set-up. The liver tissue to be examined was placed in a glass vessel with physiological saline solution to insure uniform heating of the liver tissue during the continuous measurements. This glass vessel was set in a water bath that was slowly heated by a gas heater. Attention was paid that the glass vessel containing the liver tissue did not sit directly on the floor of the water bath, since this would have resulted in uneven heating of the liver tissue. Both the water bath and the glass vessel with the liver tissue were tightly sealed with lids. These lids were suitably adapted to allow uniform and stable insertion of the 4-needle measurement probe into the tissue. In addition, there were two adaptations in the lids to place two temperature sensors (Thermocoax, typ LKI 05/50, Nagano, Japan) in the tissue at a defined distance of 10 mm to the measurement probe.

Figure 2. Experimental set-up and measurement order of ex vivo examination in porcine liver.

The LCR measurement bridge and the temperature sensor were attached to a laptop and the generated data were continuously recorded. Measurements and recordings of the specific electrical conductance σ and the temperature were done every 10 sec. Slow but continuous warming of the water bath was done to produce the coagulation. At 27°C, the conductance measurements were started. Heating was discontinued when the tissue temperature reached more than 90°C.

During the cooling phase, the hot water from the water bath was carefully removed by suction and cooling of the tissue was awaited. Twenty-five individual measurements in fresh liver tissue during the coagulation and the cooling process were then performed.

In vivo examinations were done in the perfused liver of a 30-kg pig. Permission for carrying out this trial was obtained from the Berlin State Office for Health and Social Affairs (G0015/02). Following intubation anaesthesia, arcuate transverse upper abdominal laparotomy was performed. After insertion of surgical hooks, the liver surface was freely exposed. The 4-needle measurement probe was placed into a sterile bag and subsequently inserted into the liver tissue perpendicular to the liver surface. The experimental set-up for the in vivo examinations is shown in Figure 3. A total of three measurements of the specific electric conductance σ were performed at 470 kHz and at a body temperature of 37°C in both liver flaps in different localisations of the liver.

Figure 3. Experimental set-up of in vivo examinations in perfused porcine liver.

Ex vivo examinations in human colorectal liver metastases/tumour-free human liver tissue

These examinations were done in a surgical specimen from a left hemihepatectomy in a patient with a large metachronous solitary liver metastasis of colorectal cancer. The patient's informed consent was obtained for carrying out this measurement in his surgical specimen. The ethical votum of the Ethics Commission of the Charité was also obtained. This solitary metastasis had a longitudinal diameter of 7.8 cm and an axial diameter of 6.4 cm. Left hemihepatectomy was done by Pringle's manoeuvre. Immediately after obtaining the surgical specimen measurements of the temperature of the liver samples were done, as well as two measurements of the specific electric conductance σ by inserting the 4-needle measurement probe into the tumour tissue. Measurements were also performed in the region of the healthy liver tissue.

Data analysis

For the ex vivo examinations several measurements during the coagulation and the cooling process were performed. Since the aim was a temperature rise of 1°C/min, several conductance measurements were generated during this temperature rise. The median of the generated conductance σ per °C for each individual measurement was then established for the graphic depiction. Following all ex vivo measurements in porcine liver, the median of the conductance σ per °C was established from the 25 single experiments and is shown in Figure 4.

Figure 4. Measurement curve of medians from 25 individual measurements from ex vivo porcine liver (grey points, changes during the heating phase; white points, changes during the cooling phase).

For the in vivo measurement in perfused porcine liver a total of three single measurements were performed. The mean (and the three single results) are given in the text. For the examination in the human surgical specimen multiple single measurements were performed and the results are given in the text.

Results

Examinations in porcine liver

For the ex vivo measurements, 25 continuous conductance measurements were performed during the heating and the cooling phase.

Figure 4 shows the course of the curve of conductance σ in relation to the temperature. For the temperature measurement we used two temperature sensors described above, which were placed in the liver tissue at a defined distance to the measurement probe. The congruent temperature of these both sensors supported that uniform tissue heating was achieved. Since several conductance measurements were determined for each temperature increase, it was necessary to first determine the median of the measured conductance per temperature increase for each individual measurement. This was followed by determining the medians of each temperature increase and depicting them graphically (Figure 4). The grey points demonstrate the changes in conductance during the heating phase, the white points show the changes in conduction during the cooling phase.

At a body temperature of 37°C, the median of the conductance σ was 0.41 S/m (minimum 0.32 S/m; maximum 0.52 S/m). A continuous uniform increase of conductance σ was observed up to a temperature of 77°C. The maximal conductance σ 0.79 S/m (min 0.70 S/m; max 0.87 S/m) was reached at a temperature of 80°C.

Under further heating, conductance σ remained at a plateau and there was no further increase. Starting at a temperature of 86°C, conductance σ decreased during further heating. Although the heat supply was interrupted at 90°C, a temperature increase to 98°C was observed. During the cooling phase, the conductance decreased continuously. In this context, conductance σ always remained under the measured values during the heating phase. The same conductance values were only observed when the initial temperature of 27°C was reached. Table I shows the median of the conductance (n = 25, median with min./max. is given) and the corresponding temperature during the coagulation process. To simplify data were listed in steps of 5°C in temperature range from 30 till 90°C. The data clarifies the continuous uniform increase of the conductance σ during coagulation process. At a temperature of 80°C a maximal conductance σ was reached. Analogically the median of the conductance during the cooling process is presented in Table II (n = 25, median and min./max. is given). Figure 4 is a plot of Table I and Table II.

Table I.  Specific conductivity during the coagulation process (median with min/max is given in steps of 5°C).

Temperature°C	Specific conductivity (median of 25 measurements)
30	0.36 S/m (0.28; 0.47)
35	0.40 S/m (0.31; 0.51)
40	0.43 S/m (0.36; 0.57)
45	0.47 S/m (0.35; 0.61)
50	0.52 S/m (0.40; 0.66)
55	0.56 S/m (0.42; 0.69)
60	0.61 S/m (0.46; 0.73)
65	0.66 S/m (0.50; 0.70)
70	0.71 S/m (0.56; 0.84)
75	0.77 S/m (0.61; 0.89)
80	0.79 S/m (0.70; 0.87)
85	0.78 S/m (0.58; 0.87)
90	0.75 S/m (0.54; 0.86)

Table II.  Specific conductivity during the cooling process (median with min/max is given in steps of 5°C).

Temperature°C	Specific conductivity (median of 25 measurements)
90	0.62 S/m (0.51; 0.80)
85	0.58 S/m (0.48; 0.80)
80	0.54 S/m (0.46; 0.72)
75	0.52 S/m (0.43; 0.69)
70	0.50 S/m (0.41; 0.68)
65	0.47 S/m (0.39; 0.65)
60	0.46 S/m (0.37; 0.63)
55	0.44 S/m (0.36; 0.61)
50	0.42 S/m (0.34; 0.59)
45	0.40 S/m (0.32; 0.56)
40	0.38 S/m (0.31; 0.52)
35	0.36 S/m (0.30; 0.49)
30	0.35 S/m (0.28; 0.47)

The in vivo examinations were done in perfused porcine liver. The specific conductance σ in healthy perfused porcine liver obtained from three measurements at 37°C was 0.52 S/m, 0.55 S/m, and 0.57 S/m (mean 0.55 S/m).

Ex vivo examinations in a human colorectal liver metastasis/human tumour-free liver tissue

Two consecutive measurements in human colorectal liver metastases were performed. The first measurement was done 24 minutes after obtaining the hemihepatectomy specimen. The temperature in the tumour tissue was 31.5°C. At 470 kHz, the specific conductance σ was 0.76 S/m. The second measurement was done 28 minutes after obtaining the surgical specimen at a temperature of 31.2°C in the tumour tissue. Here, the specific conductance value was 0.75 S/m.

The measurement in the healthy liver tissue of the surgical specimen was done 36 minutes after resection at a temperature of 29.3°C in the liver tissue. Here, the specific conductance σ was 0.25 S/m and thus clearly under the measured conductance values in the tumour tissue.

Discussion

In the treatment with radiofrequency ablation for colorectal liver metastases the dispersion of heat in tissue and thus the extension of the thermolesion are largely influenced by known factors such as applicator length, number of applicators, applicator geometry, initial power output and length of application 24. Despite these device-dependent parameters 25, 26, especially tissue-dependent parameters play a decisive role in the heat dispersion in tissue. Besides perfusion of the liver these include the so-called tissue-specific properties themselves with the specific electric conductivity decisively determining the depth of the heat dispersion in the tissue 18. Ideally, the specific electric conductivity of the tissue and its changes during the coagulation process should be known to enable individual planning of the length and intensity of ablation. Analogous to radiofrequency ablation, the optical liver parameters for applying laser-induced thermotherapy are known and their influence on the heat dispersion in tissue has been demonstrated 27–29. The aim of this experimental study was to determine for the first time the changes in the specific electric conductivity σ during the complete coagulation process in the liver tissue.

A number of authors have shown that markedly larger coagulation necroses can be created by changing the biological properties of the tissue to be treated 30–33. Saline (NaCI) injections into the tissue change the electrical conductivity and result in markedly larger coagulation necroses. Based on this knowledge, within the last years internal cooling and the perfusion of saline solution around radiofrequency applicators have established themselves in recent years in clinical practice 26, 34, 35. These findings also show the importance of electrical conductivity for the produceable size of coagulation necrosis of radiofrequency ablation and the influence on entire tumour destroying with a sufficient security margin.

In order to gain precise knowledge of the influence of the tissue-specific parameters, a number of experimental studies were performed in the last few years. The great interest in this type of studies is based on the fact that only detailed knowledge of the tissue-specific parameters and their changes during radiofrequency ablation will make it possible to predict the possible size of the produceable coagulation necrosis. With this knowledge it may be possible to rectify the problem of not being able to predict the lesion size in radiofrequency ablation. Knowledge of these parameters would enable the development of computer-supported simulation programs that would be helpful for planning individual and oncologically safe therapies.

In an animal experimental study, Ahmed et al. 33 were able to show great differences among tissue- and tumour-specific parameters which can thus influence the results of radiofrequency ablation. Lobo et al. 36 demonstrated that the heat conductivity of the tissue changes during radiofrequency ablation, and thus has an influence on the coagulation necrosis. They concluded that RFA-induced temperature changes correlate directly with changes in the specific conductivity of the tissue.

In order to examine these changes in the specific tissue conductivity during a temperature increase, a number of experimental phantom examinations have been carried out in the last few years 30, 37 that served as a basis for the development of simulation programs. Although studies on the conductivity of various biological tissues have been performed already in the first half of the twentieth century 38, they were, however, always done under constant temperatures. Studies on dynamic changes of the specific conductivity during temperature changes in the area of radiofrequency ablation in biological tissue have not been done thus far. One of the problems with continuous measurements of the specific conductivity is that the high-frequency electricity of radiofrequency ablation directly affects the conductivity measurement technique, since these measurements are based on different tensions.

In our study we used the established 4-needle measurement method to determine the electric conductivity 21–23, 39. In order to avoid any influence of the high-frequency electricity of radiofrequency ablation on this measurement technique in our ex vivo examinations in porcine liver, we preferred as an alternative, warming of the tissue by heating in a water bath. For the development of simulation programs it is also important to know the specific conductance value of the tissue during the cooling phase, since areas with different temperatures are simultaneously observed in the treated tissue during radiofrequency ablation depending on the distance to the applicator. Thus we performed these continuous conductance measurements also during the cooling phase.

As demonstrated by our results, the specific conductivity increases during warming. At a temperature of 80°C the highest specific conductivity in the liver parenchyma is found–further heating does not result in a higher conductance. At temperatures over 80°C the specific conductivity rather decreases. Therefore a RFA should be performed using an intended temperature of 80°C in terms of ‘slow cooking’ because at this temperature the optimal conductance is reached and the best possible destruction of the tumour could be achieved.

The extent to which the specific conductivity depends on the biological state of the examined tissue became evident through the exemplary measurement of the specific conductivity in the perfused porcine liver and the ex vivo measurements of a human liver metastasis with surrounding healthy liver tissue. In our ex vivo examinations in porcine liver, the median of our specific conductance value σ was 0.41 S/m and 0.55 S/m in the perfused porcine liver at a body temperature of 37°C, the specific conductance value in the human liver metastasis was 0.75 S/m and clearly higher at 31°C. In contrast, in the healthy human liver, this value was 0.25 S/m at 29°C and was thus markedly lower.

Haemmerich et al. 40 could also demonstrate a higher electrical conductivity in liver metastases. This higher electrical conductance in liver tumour tissue compared to the surrounding healthy liver parenchyma supposes a tumour selectivity in the application of RFA. Hence RFA could be applied to parenchyma saving.

These results clearly show that understanding about temperature-dependent changes in the specific conductance is important for RFA of malignant liver tumours. In the future both temperature and the biological condition of the tissue should be considered during planning of the therapeutic regimen. An impedance-controlled/adapted therapy taking into account the changes in the electrical conductance would be a reasonable approach due to total tumour destruction including a sufficient safety margin. Another possibility would be that prior to radiofrequency ablation, the specific conductance value of the tissue to be treated is directly determined and integrated into the individual planning and execution of therapy. In the radiofrequency ablation of metastases, it may also be very useful to know about the tissue properties of the primary tumour.

Conclusion

In radiofrequency ablation, biological properties such as specific electric conductivity play an important role in the produceable size of coagulation necroses. For an ideal oncological therapy plan it is necessary to know the biological properties and to especially recognise the changes involved during the coagulation and cooling phase in the temperature range of radiofrequency ablation. Using the 4-needle measurement technique, we were able to measure for the first time the changes in the specific electric conductivity during the coagulation and the cooling phase. At a temperature of 80°C the highest specific conductivity in the liver parenchyma is found–further heating does not result in a higher conductance. Therefore a RFA should be performed using an intended temperature of 80°C in terms of ‘slow cooking’ to achieve the best possible destruction of the tumour. Furthermore the increased conductivity in the tumour specimen compared to healthy liver parenchyma allows for tumour selectivity in RFA as a parenchyma saving approach.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.