# Investigation of safety for electrochemotherapy and irreversible electroporation ablation therapies in patients with cardiac pacemakers

### Experimental evaluation of the effect of electroporation pulses on the pacemaker

The measurements were performed at room temperature (21 °C) in a glass container filled with either physiological saline (0.9% NaCl solution) or the saline diluted with distilled water at 1:4 ratio, thus resulting in two media with conductivities 1.57 and 0.34 S/m, respectively (the lower value mimicked the conductivities encountered in tissue during ECT of liver metastases). The conductivity was measured at 21 °C with SevenCompact S230 conductometer (Mettler Toledo, Columbus OH, USA). The experimental setup is presented in Fig.  [38].

In the first stage of the study, the bipolar ventricular lead (type CapSure Z Novus 5054) was not connected to the pacemaker; the connector (not submerged) allowed us to measure the maximum possible voltage or current (with the contacts open or shorted, respectively) between the pacemaker electrodes due to application of electroporation pulses. In the second stage, the ventricular lead was connected to an Adapta pacemaker (ADDR01 model, Medtronic, Minneapolis, USA) which was programmed into asynchronous D00 pacing mode and submerged. The atrial bipolar lead was also connected. Atrial pacing pulses were converted into adjustably delayed TTL pulses (0 V and 5 V output values) to trigger the generation of electroporation pulses. Thus, we were able to observe the effects on the pacemaker function for electroporation pulses delivered at different times with respect to the charge-balanced ventricular pacing pulse. The pacemaker is assumed to be in its most vulnerable state during generation of the pacing pulse due to relatively low internal impedance that could allow harmful currents flowing into the device due to electroporation interference.

Electroporation pulses were generated by Cliniporator Vitae device (IGEA, Carpi, MO, Italy) and delivered via two needle electrodes for clinical ECT (type VG-1230M20; conductive length 3 cm, diameter 1.2 mm) submerged in parallel into the medium (Fig. ). The inter-electrode distance was fixed at 3 cm, the maximum distance limited by the hardware capacity and also recommended in the standard operating procedures for ECT [10]. Standard rectangular pulses (1000 and 3000 V amplitude, 100 µs duration) were generated individually or in sequences of four pulses (repetition rate 5 kHz, i.e., 100 µs on, 100 µs off). The voltages appearing between the electrodes of the ventricular lead were sensed at the ventricular electrodes in the medium. The measurement instrumentation included HDO6104A oscilloscope, two HVD3605 differential high-voltage and two CP031A high-current probes (Teledyne LeCroy, Chestnut Ridge, NY, USA) for monitoring of generated electroporation pulses and interferences on the ventricular lead.

### Numerical modeling

The impact of the presence of a metal-encased pacemaker on effectiveness and safety of electroporation-based therapies was further investigated by means of numerical computation. Two scenarios for treatment of a subcutaneous tumor were investigated: ECT and IRE. In both scenarios the influence of a metal-encased pacemaker was evaluated with the pacemaker in contact and without contact with one of the electrodes. A control scenario without the pacemaker was also evaluated. A previously designed numerical framework for planning of electroporation-based treatments was adapted for all computations [20, 28, 39].

All numerical computations were performed in COMSOL Multiphysics software (Comsol AB, Stockholm, Sweden), however the computations were set up and controlled in MATLAB (MathWorks, Natick, MA, USA) scripting environment through LiveLink. A simplified geometry including both the tumor and the pacemaker was used in this study. Placement of the pacemaker mimicked its position on the fascia of the pectoralis major muscle (Fig. ). The tissue model consisted of three isotropic and homogeneous components: the spherical tumor (12 mm diameter), the fat tissue, and the underlying muscle tissue. The skin was not included in the model due to subcutaneous location of both the tumor and the pacemaker. The electrical and thermal properties of tissues and electrodes were taken from literature and databases and are listed in Table along with the relevant references. The pacemaker model consisted of the titanium housing and the silicone-covered lead connectors. The built-in material properties from COMSOL were used (Titanium beta-21S and Silicon). An unstructured tetrahedral mesh was built in COMSOL and consisted of 27,643 elements for the ECT model and 21,623 elements for the IRE model. When compared to the finest possible mesh that was still manageable in the transient computation (197,119 elements for the ECT model and 138,480 elements for the IRE model), the use of a coarser mesh produced a < 1% error in calculated electric current and maximum temperature while greatly reducing the computation time.

### Table 1

Electrical and thermal properties of modeled tissues taken from relevant literature (given in brackets)

Fat Tumor Muscle
Initial electrical conductivity σ0 (S/m) 0.080 [40, 41] 0.200 [39, 42] 0.135 [35, 39]
Final electrical conductivity σend (S/m) 0.240 [35, 39] 0.600 [35, 39] 0.405 [35, 39]
Threshold for reversible EP (V/cm) 100 [35, 39] 400 [35, 39] 200 [35, 39]
Threshold for irreversible EP (V/cm) 900 [35, 39] 900 [35, 39] 900 [35, 39]
Thermal conductivity k (W/m K) 0.21 [41] 0.52 [20, 40] 0.49 [41]
Specific heat capacity Cp (J/ kg K) 2348 [41] 3540 [20] 3421 [41]
Density ρ (kg/m3) 911 [41] 1079 [41] 1090 [41]
Perfusion rate ω (1/s) 0.00043 [40] 0.01798 [40] 0.00069 [40]
Thermal coefficient of conductivity αT (%/°C) 1.5 [20] 1.5 [20] 1.5 [20]

For the ECT model (Fig. a) a hexagonal-electrode configuration of seven electrodes was used with a standard ECT protocol: eight 100 μs pulses per each of 12 active electrode pairs delivered in two sequences of four pulses with reversed pulse polarities. The sequences were delivered at 1 Hz and the pulses within each sequence at 5 kHz repetition rate. The applied voltage was 730 V [6, 10].

For the IRE model (Fig. b) four needle electrodes were modeled, surrounding the tumor in a rectangular configuration. IRE delivery protocol from [20] was used in the simulation: 90 pulses of 90 μs duration per electrode pair with a 1500 V/cm voltage-to-distance ratio delivered at 1 Hz with a pause of 3 s after each set of 10 pulses. The pacemaker was positioned either 5 mm from the nearest electrode (Fig. a) or in direct contact with the nearest electrode (Fig. b).

Electric field distribution in tissue is determined through solving the stationary Laplace partial differential equation for electric potential. The outer boundaries of model domain are considered electrically insulated while the continuity equation is applied to the inner domain boundaries. Electroporation is implemented as a non-linear electric field dependent increase in tissue electrical conductivity [20]. Electric field distribution is calculated separately for each active electrode pair in the treatment. The computed electric field for the n-th electrode pair is compared to computed field from all previous pairs (1 to n − 1) and the maximum contributions from all pairs are combined into treatment equivalent field Eeq,n of n-th electrode pair as follows:

$Eeq,n=maxEeq,n-1,En;n>1En;n=1;1≤n≤N,$

where N is the total number of electrode pairs, Eeq,n is the treatment equivalent field after application of pulses to the n-th electrode pair, Eeq,n−1 is the treatment equivalent field from electrode pairs 1 to n − 1 and En is the actual computed electric field produced by the n-th electrode pair. The final electric field distribution in tissue is represented by the equivalent electric field after application of pulses to all N electrode pairs (Eeq,N). The percentage of tumor volume covered in target electric field strength, 400 V/cm for ECT and 650 V/cm for IRE ablation [39], was extracted from the final field distribution.

Computations of heat dissipation are performed separately with a transient model through solving the bioheat transfer equation [14, 28, 43]:

$ρCp∂T∂t+∇∙-k∇T=Q+ρCpωTblood-T+Qmet,$

Right side of the equation represents the heat sources in the model—the heat source Q approximated by a Joule heating term and source terms representing tissue perfusion and metabolism. Similarly to the computation of electric field distribution, the outer boundaries of the model domain are thermally insulated, in order to create the “worst case” conditions, while continuity condition is applied to the inner boundaries. All parameters descriptions and values are provided in Table . Maximum tissue temperature is calculated at the end of pulse delivery for each electrode pair (Fig. ).

a 3D model for the simulation of ECT treatment of a spherical subcutaneous tumor. The distance from the pacemaker casing to the nearest electrode is 5 mm. b 3D model for the simulation of IRE ablation of a spherical subcutaneous tumor. Pacemaker is in direct contact with rightmost electrode. Only one scenario (contact or no contact) is shown for each treatment