About TCRT

TCRT has been on continuous publication since 2002. It is covered currently by all the major data systems such as Medline, PubMed, Web of Science, Thomson's ISI and SCI and Scopus.

The 2012 Impact factor for TCRT is 1.943

TCRT Open Access
TCRT seeks original articles
Cancer Watch
Open Access

Design of an Irreversible Electroporation System for Clinical Use (p. 313-320)

Irreversible electroporation is an ablation modality in which microseconds, high-voltage electrical pulses are applied to induce cell necrosis in a target tissue. To perform irreversible electroporation it is necessary to use a medical device specifically designed for this use.

The design of an irreversible electroporation system is a complex task in which the effective delivery of high energy pulses and the safety of the patient and operator are equally important. Pulses of up to 3000 V of amplitude and 50 A of current need to be generated to irreversibly electroporate a target volume of approximately 50 to 70 cm3 with as many as six separate electrodes; therefore, a traditional approach based on high voltage amplifiers becomes hard to implement.

In this paper, we present the process that led to the first irreversible electroporator capable of such performances approved for clinical use. The main design choices and its architecture are outlined. Safety issues are also explained along with the solutions adopted.

Key words Electroporation; Irreversible Electroporation; Ablation Modality; High Voltage Pulses; Tumor Treatment; Medical Device.

Introduction

Electroporation is a technique that uses micro to milliseconds electric pulses to create pores in the cell membrane, thus allowing molecules that, due to their physical and/or chemical properties, would normally not be able to cross the cell membrane, to enter the cell (1, 2, 3, 4, 5). Electroporation finds applications in many fields in particular for gene insertion in cells (electrogenetherapy) (6, 7) and for the treatment of cancer (electrochemotherapy). In electrochemotherapy, the combination of chemotherapy and electroporation of tumors, the effects of drugs that usually show little cytotoxicity are greatly increased (8). The opening of pores in the cell membrane allows the chemotherapeutic agent to enter the cell at greater, more effective concentration and exert its cytotoxic action killing the target cell (9, 10, 11).

However, if the applied electric field is above a certain threshold cells are unable to seal the pores formed, thus causing cell death due to the loss of homeostatic mechanisms (11). This phenomenon is termed ?irreversible electroporation?. The threshold for irreversible electroporation strongly depends on the type of tissue under treatment (13).

The main advantage of irreversible electroporation over electrochemotherapy is the possibility to avoid the use of drugs, as it relies only on the effect of the applied electric field to kill the cancer cells. However, to obtain irreversible electroporation of the target tissue, high voltage pulses need to be generated, the electric field is applied by means of electrodes inserted in the tissue to be treated. Since the amplitude and the gradient of the field depend on the applied voltage as well as on the distance between the electrodes, high values of the electric field can be also achieved by arranging the electrodes closer to each other. However, when treating deep seeded tumors, we desire to introduce the electrodes through the skin both to minimize the procedure?s impact and avoid open surgery; thus, introducing a large number of closely arranged electrodes to effectively porate the target tissue is not feasible. In this scenario the goal is to treat a volume of approximately 50 to 70 cm3 with up to six electrodes. This requires applied voltages in the order of 3000 V to obtain irreversible electroporation. The specifications of the device ensure that the electrical field gradient in a volume 40 cm3 is at least 800 V/cm, this is considered to be the threshold for irreversible electroporation in most cells (13). The electrodes used with the device are 15 cm long stainless steel needles, partially insulated, with a diameter of 1 mm. The conductive, non insulated, distal end of the electrodes can be up to 4 cm long. The electrodes are inserted under ultrasound guidance.

The high voltages and currents (up to 50 A) used represent the main problems to be taken into account when designing a device for irreversible electroporation. Safety issues arise because of the high energy involved and because of the variable working conditions present in the operating theatre. In fact, the delivered current depends on the ohmic characteristics of the tissue under treatment and can differ from point to point, particularly in the typical heterogeneous tissue that constitutes tumors. Moreover, the tissue can have its electric characteristics altered during the treatment, as a consequence of the deep modifications caused to the cells and the extra cellular environment when high currents are applied.

Consequently in the design of an irreversible electroporator, one has to keep under control the patient leakage current, to design a sturdy and fast high voltage pulse generator, with an high reactivity in case of failure and provide a user-friendly interface for the operators to minimize the chance of user error.

This paper describes the solutions adopted to solve all these challenging issues in the design and implementation of a device for irreversible electroporation to be used in the clinical practice.

General Safety Remarks

In addition to general safety requirements for medical devices stated by regulation and standards (14, 15, 16), specific safety issues characterize electroporation devices. The principal hazards derive from the high energy that is accumulated on capacitors and from the delivery of high electrical current to the patient, which involves the risk of electrocution for both the patient and the operator.

Energy release to the patient must be reliably controlled and limited: unintended or incorrect release has to be avoided. Strictly related to this issue is device ruggedness: in the absence of adequate protective measures, a failure of some critical component may lead to a complete, unwanted discharge of the accumulated energy.

Reliable energy delivery control characterize the normal device operation and can be ensured by specific safety measures implemented in the software/firmware. This goal is achieved by carrying out risk analysis in the earliest phases of the device architecture design, in order to identify those hardware parts that must support software safety features; ensuring that critical device control is carried out by reliable programmable systems, and implementing best software/firmware development practices (17).

Energy delivery limitation refers to fault conditions. It is a strategic choice: risk control may be obtained either by limiting the probability of a failure, in particular for those critical components whose failure may lead to uncontrolled energy delivery to the patient, or by implementing an independent system that prevents energy delivery above the maximum normal-condition value; both solutions have pros and cons. In any case the system must be sufficiently rugged to confine such failure to a remotely likely event.

Ruggedness is intended in particular against short circuits and sparks that are likely to occur between electrodes due to the unpredictability of the resistive load of biological tissues, the presence of conductive solutions and human error.

The likelihood of an electroporation treatment to lead to electrocution depends on several factors: the applied pulse voltage, the length of the pulses, the number of pulses, the pulse repetition rate and, of course, the distance between the hearth and the electrodes (18). Generally speaking, synchronization of treatment delivery with the refractory period of the cardiac cycle is always advisable when there is not enough confidence that the electroporation treatment cannot determine a current density lower than fibrillation threshold of the myocardium.

General Device Structure

The electroporation device is composed of two main parts: the user interface (UI), that calculates the treatment parameters based on data inserted by the operator, shows and elaborates data and signals measured during the treatment, and a power unit (PU), which actually generates the pulses by using the parameters provided by the interface and acquires signals.

The UI is a medical grade PC suitable for use in a operating room and it guarantees compliance with standards for medical electrical devices. A standard Operating System (OS) is used to provide an intuitive graphical user interface, allowing the operator to easily set treatment parameters and to process measured treatment data. Once the parameters have been set, the PU should be able to perform the treatment independently from the non real-time OS and the unpredictable delays associated with it. This approach guarantees the execution of the operations involved in the treatment delivery with an exact timing sequence and to react immediately to the events being monitored. The described configuration allows a fine control over the charge on the capacitors, the pulse length, the pause between pulses, and to immediately react to safety-related alarms.

Strict control of timing delays and signals can be better provided by an FPGA (Field Programmable Gate Array) based platform than a microcontroller. The FPGA processes the treatment parameters, controls if the capacitors store enough energy to deliver a high-voltage, high-current pulse, then handles the generation of pulses and the measurement of the delivered treatment.

A schematic representation of an electroporator device is shown in Figure 1. The PU is composed of high energy parts (highlighted in gray in Figure 1) and a digital control part. The control block mainly includes the FPGA and an interface for communicating with the UI that directly supervises and handles the operations of the other blocks. The high voltage capacitors store all the energy for a sequence of pulses with the desired voltage, as commanded by the FPGA. Once the capacitors have charged, the FPGA waits for a signal to proceed with the treatment. The signal is provided directly by the operator by means of a double pedal with IPX-8 degree of protection from liquids (as imposed by the standards for the use in operating theatres). The first pedal is used by the operator to enable the second pedal that actually starts the treatment.


Figure 1: Chart of the main blocks composing the device. The power unit is composed of high energy parts, highlighted in gray, and a digital control part.

Once triggered, the FPGA is in charge of two main tasks: it controls the output pulse block and acquires the measure of voltage and current actually delivered to the patient through the electrodes. The pulse length can be suitably controlled by using two high voltage IGBT (Insulated Gate Bipolar Transistor): one to control the pulse length and the second dedicated to avoid the delivery of uncontrolled pulses under failure conditions.

The electroporator device is able to use up to six electrodes to apply an electric field with the most appropriate shape and intensity to homogenously cover the target tissue. A switching block is needed to route the treatment to the electrodes in sequenced fashion. In order to implement such capability, high voltage SPDT (Single Pole Double Throw) relays are used. For each electrode two relays are involved: one to select the polarity of the pulse and the second to connect the electrode. The second relay also has the additional purpose of performing an initial test of the device efficiency conducted using an internal resistive load.

Finally, the FPGA directly controls the switching activity of the relays of the switching block by connecting each electrode in turn to the output pulse generation block.

In order to reduce design complexity, we chose to consider the whole PU as the patient circuit, defined as any electrical circuit containing conductive parts, which are not insulated from the patient (16). This means that for sake of safety not only the high voltage part is completely insulated from ground but the whole PU. Based on the standards we found that the insulation required is 7500 V; therefore, we provided the PU with a transformer having an internal insulation of more than 8000 V.

Since the PC-based user interface is referred to ground and communicates with the power unit via USB (Universal Standard Bus), it is necessary to isolate the USB connection. To this end, we adopted an optically insulated USB cable.

Design Issues

Based on the general structure shown, we implemented an irreversible electroporation device capable of charging the capacitors up to 3500 V and to sustain a current drawn by the tissues of up to 50 A. Many issues have raised during the design process mainly because of the high energy involved and because of the need for safety associated with a medical device to be used in the operating room. The following paragraphs provide a hint on the solutions we adopted.

Pulse Generation

Ideally, the delivered pulses should be square ones. Unfortunately, the generation of square pulses at 3000 V is very expensive and difficult to achieve, as it requires the design of a very-high-voltage amplifier. We investigated two different approaches for the generation of pulses that approximate the ideal square shape, without requiring the use of an amplifier: the pulse transformer and the high voltage generator.

Pulse Transformer: When using a pulse transformer, pulses at lower voltage are applied to the primary coil of a step-up pulse transformer, which generates the required high voltage pulses on the secondary coil. The energy required is supplied by the charge stored on capacitors. The output energy to be delivered is stored on capacitors connected in parallel to reach the desired capacity. The value of the required maximum charged voltage on the capacitors is dependent on the transformer ratio.

Since the pulse transformer is less efficient at low frequencies than at higher frequencies, a bigger transformer is required to generate longer pulses. Another issue with the use of a transformer is the major difference between the achieved pulse shape and the ideal square one, especially in presence of low-resistance loads.

Tests have been carried out using a toroidal pulse transformer having 148 mm outer diameter; 56 mm height; 3.8 kg weight. The outcome was that the pulse transformer did not allow pulses longer than 30 μs and pulse repetition frequency had to be lower than 81 Hz (1/12300 μs). Figure 2 shows the effect of transformer?s core saturation that occurs at higher pulse repetition frequencies.



Figure 2: Two Pulses at 3000 V, 30 μs long, 30 μs pause, with a 500 Ω load.

The load effect became significant with loads lower than 250 Ω, as shown in Figure 3: as load impedance decreases, the voltage pulse amplitude on the secondary coil of the transformer becomes significantly lower, the pulse shape is also affected.


Figure 3: Pulse voltage (green) and current (blue) dependence from load with a 3000 V nominal amplitude, 30 μs long pulse: (a) 500 Ω, (b) 250 Ω, (c) 60 Ω.

High Voltage Generator: When using a high voltage generator the capacitors are directly connected to the load by means of switches in order to generate the pulses. With this approach the ideal square pulses are approximated to a time-controlled free discharge of a capacitor (see Figure 4).



Figure 4: Truncated slow-decay exponential. The square wave pulse is produced by a partial discharge of a large capacitor, which requires the interruption of high currents against high voltages.

The quality of this approximation depends on the load. This means that such kind of pulse could be well approximated to a square pulse only if the current drawn by the load is low. This is because the current flowing on the load discharges the capacitors, thus causing the amplitude loss shown in Figure 4. A small discharge current causes a little drop and the amplitude at the end of the pulse is only slightly lower than the initial capacitor voltage. On the other hand if the load has a particularly low resistance it will draw high current, and the pulse shape will have a considerable drop, although starting from the required voltage. The capacitor should be dimensioned so that the time constant of the voltage decay is much longer than pulse duration also in the least favorable load conditions.



Figure 5: Voltage and Current wave forms at 3000 V (at the end of ten 100 μs, 20 μs spaced pulses) on a 68 Ω load (approx. 50 A peak current), obtained using the high voltage generator approach. A single pulse sequence in (a). The repetition of 8 sequences in (b). The dwell between sequences is not shown.

The high voltage generator approach requires a high voltage power supply that charges the capacitor at the desired initial pulse voltage and high voltage capacitor of sufficient capacitance. The required high voltage may be obtained connecting capacitors in series, in spite of the total capacitance, which results in a fraction of the capacitance of each single capacitor.

In order to generate the pulses a high voltage Mosfet switch or IGBT can be used to connect the load with the capacitors only for the required time of the pulse length.

After some evaluation about dimensions, costs and considering the specifications, the high voltage generator approach generally has shown to be the best solution.

Speed Issue

Another main issue is the device responsiveness. The pulses delivered can have a length between 20 μs and 1 ms and require a fast control mechanism. Due to the frequency of the pulses and the high currents involved, we found that a safety protection using fuses is not reliable enough, since it is hardly possible to have fuses with a reaction time of few microseconds. For this reason we decided to implement a dedicated fast redundant hardware solutions to protect both the device and the patient.

Variable Load

A peculiarity of the electroporation process is that the resistive load of a biological tissue varies greatly and is, therefore, unknown at the time of treatment. Moreover, the resistance depends on the physical properties of the electrodes used and there is a decrease of the resistance during the treatment due to the changes induced by the electrical field in the target tissue.

As a consequence, it is almost impossible to make an accurate prediction of the load without a preliminary test and it is safer to consider the least favorable scenario. For this reason we designed the device considering the event of sparks and short circuits and introduced a current limitation to avoid injuries to the patient and failure of the device. Simulations allowed to estimate that in the worse non-short-circuit condition, the load should not draw more than 50 A; therefore, it is assumed that if during the treatment the drawn current exceeds this threshold, it is probably due to a short circuit or spark occurring at the electrodes. In this case the device interrupts the pulse sequence.

Size Issue

The size of the device can vary depending on the choices made during the design phase. We chose to build a device provided with a trolley (Figure 6), allowing us to have larger space for the electronics. The advantages of this choice are: the required electrical insulation is achieved more easily and it is possible to increase the robustness to noise induced by the high voltage parts. We decided to put the PC for the UI into a separate part close to the LCD display and the keyboard. The PU is located at the bottom of the trolley.


Figure 6: Trolley containing the high voltage power unit at the bottom and the PC with the user interface (keyboard, touchpad, and display) on top.

Software Architecture

As shown in the device structure, a double level of control is desirable. The software consists of two main parts that interact with each other but are logically and functionally independent: a high level part implements the software graphical user interface and a low level control, implemented in hardware and in the FPGA firmware, that is responsible for all functions that directly control the power part of the device.

The Graphical User Interface

The Graphical User Interface (GUI) was developed thinking of ease of use and enable also an inexperienced user to focus on the therapy rather than on the device usage.

The main structure of the GUI is a typical wizard application with the ?next? and ?back? buttons that allow navigation between pages.

The inner task of the GUI is to monitor the status of the PU at fixed time intervals and display the main information regarding the capacitors charge and the treatment delivery process.

The GUI is based on the following structure:
    ? COMM_LAYER: a layer that performs the basic communication with the power device, supervises the USB communication, and performs basic read/write operations;

    ? CHECK_LAYER: checks the input data inserted by the operator to prevent loading of inconsistent parameters on the power unit;

    ? TRANSLATION_LAYER: elaborates all data received from the power unit in a format such that the user can quickly analyze the information on the treatment delivered.
The COMM_LAYER and TRANSLATION_LAYER run in the background at all times during the treatment delivery and ensure a strict timing synchronization of the different tasks and between the GUI and PU control firmware. This feature is implemented by means of a handshake protocol that enables the operator to have full monitoring of the ongoing operations. In this way a treatment can be aborted by the operator at any time via software and both the PU and the GUI return in a safe state, waiting for new inputs. On the other hand, if an error occurs during an operation, the GUI immediately notifies it to the user who can decide how to proceed.

In order to improve safety in the use of this device, tests are performed at start-up before the GUI actually starts operating. These tests check the main functionality of the device such as a live communication between UI and PU, the correct firmware version on the power unit, complete check of the RAM chips of the PU control hardware, a complete charge and discharge cycle of the capacitors, a test on internal test loads, et cetera.

Power Unit Control

A FPGA has been chosen to control the whole PU. By using a dedicated board we allow the PC-based UI to represent the FPGA as a series of addresses where the treatment parameters are stored. The FPGA firmware has the structure shown in Figure 7 and it is divided into several parts, each one with a well defined task. The main parts are: the interface block used to communicate with the PC; the charge control part to handle the charge of the capacitors; the generation part to control the pulse generation; the measurement part to allow the measure of voltage and current during the treatment, and the switching part to control the relay configurations.



Figure 7: FPGA firmware architecture.

The system controls the pulses by means of two high voltage IGBTs. These two components have different purposes: the first is directly driven by the FPGA to determine the desired pulse length, and the second IGBT is used to limit the maximum length of each pulse to 1ms, for safety reasons in the event of failure. Since the specifications allow a maximum current of 50 A, we used the IGBTs to control the current as well. We improved a standard driver with a circuit capable of readily reacting to an over-current, by causing the pulse to stop within 10 to 20 μs.

Since both the capacitor voltage and the current and voltage of the pulses are not required to be measured at a high rate, the use of analogue-to-digital converters with a serial interface has been preferred. In this way, a compact SPI (Serial Peripheral Interface) communication system can be used between the measure blocks and the FPGA.

The charging of the capacitors is operated by a high voltage generator with a maximum voltage over 3500 V. The charge control block enables the generator when a charge phase is required and disables it both during the pulse delivery and during any discharge process. Furthermore, the charge control block checks if the generator is working properly: when the expected voltage cannot be reached the FPGA sends an alarm message to the UI and puts the system into a safe condition by immediately closing a relay that discharges the capacitors on an internal load. For safety reasons the system has always the internal load connected and only when the treatment is in progress it re-connects the output section.

To improve the efficiency of the pulse and generation control we chose to use two different clock domains: one reserved only for the communication section that uses a 30 MHz clock from the microprocessor used to implement the USB communication; the second clock domain, that controls the charge, the pulse generation, and measurement sections uses a separate 20 MHz clock. With this solution we ensure that even if the microprocessor fails, the system will complete the treatment correctly and afterwards will reach a safe state.

In consideration of a possible application of the treatment close to the heart we improved the safety of the device by introducing the capability to synchronize pulse delivery with an external ECG device so that the treatment is delivered only during the refractory period.

Regulation and Standards

Compliance with regulation and standards guarantees that the device meets electrical safety requirements.

Since the irreversible electroporation system is a medical device, we followed European standards EN60601-1 (14) and EN60601-2 (15), and the harmonized UL standard UL60601 (16) during the whole design process. Due to the high voltages involved in the use of the device, we decided to also follow other standards, like the one for defibrillators (19), and the one for high frequency surgical equipment (20).

Conclusions

The development of an irreversible electroporation system approved for the use in the clinical practice, allows a novel approach to tissue ablation. Tumor ablation based on irreversible electroporation relies solely on the application of an intense enough electric field to the target tissue to cause cell death. The device has been developed to allow the use of up to six independent electrodes. The electrodes are not restricted to a fixed geometry, rather they can be independently positioned based on the tumor size, shape, and position, to ensure that the target is entirely enclosed within the applied electric field, thus ensuring complete tissue ablation. The result of the development process is a reliable, safe, and high performing medical device for irreversible electroporation based tissue ablation.

References
  1. Neumann, E. and Rosenheck, K. Permeability Changes Induced by Electric Impulses in Vescicular Membrane. J Membrane Biol 10, 279-290 (1972).
  2. Crowley, J. M. Electrical Breakdown of Biomolecular Lipid Membranes as an Electromechanical Instability. Biophysical Journal 13, 711-724 (1973).
  3. Zimmermann, U., Vienken, J., and Pilwat, G. Dielectric Breakdown of Cell Membranes. Biophysical Journal 14, 881-899 (1974).
  4. Weaver, J. C. Electroporation of Cells and Tissues. IEEE Transactions on Plasma Science 28, 24-33 (2000).
  5. Puc, M., Corovic, S., Flisar, K., Petkovsek, M., Nastran, J., Miklavcic, D. Techniques of Signal Generation Required for Electropermeabilization. Survey of Electropermeabilization Devices. Bioelectrochemistry 64, 113-124 (2004).
  6. Neumann, E., Schaefer-Ridder, M., Wang, Y., Hofschneider, P. H., et al. Gene Transfer into Mouse Lyoma Cells by Electroporation in High Electric Fields. J EMBO 1, 841-845 (1982).
  7. Gehl, J. Electroporation: Theory and Methods, Perspectives for Drug Delivery, Gene Therapy and Research. Acta Physiol Scand 177, 437-447 (2003).
  8. Mir, L. M., Belehradek, M., Domenge, C., Luboniski, B., Orlowski, S., Belehradek, J., Schwaab, B., Luboniski, B., Paoletti, C. Electrohemotherapy, A Novel Antitumor Treatment: First Clinical Trial. C R Acad Sci Ser III 313, 613-618 (1991).
  9. Mir, L. M. Therapeutic Perspective of In Vivo Cell Electropermeabilization. Bioelectrochemistry 53, 1-10 (2000).
  10. Marty, M., Sersa, G., Garbay, J. R., Gehl, J., Collins, C. G., Snoj, M., Billard, V., Geertsen, P. F., Larkin, J. O., Miklavcic, D., et al. Electrochemotherapy ? An Easy Highly effective and Safe Treatment of Cutaneous and Subcutaneous Metastases: Results of ESOPE (European Standard Operating Procedures of Electrochemotherapy) Study. Eur J Cancer Supp 4, 3-13 (2006).
  11. Mir, L. M., Gehl, J., Sersa, G., Collins, G. C., Garbay, J. R., Billard, V., Geersten P. F., Rudolf, Z., O?Sullivan, G. C., Marty, M. Standard Operating Procedures of the Electrochemotherapy: Instructions for the Use of Bleomycin or Cisplatin Administered Either Sistemically or Locally and Electric Pulses Delivered by the Cliniporator™ by Means of Invasive or Non-invasive Electrodes. Eur J Cancer Supp 4, 14-25 (2006).
  12. Rubinsky, B., Onik, G., and Mikus, P. Irreversible Electroporation: A New Ablation Modality ? Clinical Implications. Technology in Cancer Research and Treatment 6, 37-48 (2007).
  13. Davalos, R. V., Mir, L. M., and Rubinsky B., Tissue Ablation with Irreversible Electroporation. Annals of Biomedical Engineering 33, 223-231 (2005).
  14. European Standard EN60601-1 Medical electrical equipment - Part 1: General Requirements for safety; Second Edition (1998).
  15. European Collateral Standard EN60601-1-2; 2. Collateral Standard: Electromagnetic Compatibility Requirements and Tests.
  16. UL Standard for Safety for Medical Electrical Equipment, Part 1: General Requirements for Safety, UL; 60601-1; First Edition (2003).
  17. General Principles of Software Validation, Final Guidance for Industry and FDA Staff, CDRH (2002).
  18. Patrick, R. J. Applied Bioelectricity: From Electrical Stimulation to Electropathology. Springster-Verlag New York, Inc. (1998).
  19. IEC 601-2-4:1983-01; Medical electrical equipment - Particular Requirements for the safety of cardiac defibrillators and cardiac defibrillator-monitors.
  20. IEC 601-2-2:1998-09; Medical electrical equipment - Part 2: Particular Requirements for the safety of high frequency surgical equipment.




TCRT August 2007

category image
Volume 6
No. 4 (p 255-360)
August 2007
ISSN 1533-0338

Claudio Bertacchini, M.S.*
Pier Mauro Margotti, M.S.
Enrico Bergamini, M.S.
Andrea Lodi, Ph.D.
Mattia Ronchetti, M.S.
Ruggero Cadossi, M.D.

IGEA S.r.l.
Via Parmenide, 10/A
I-41012 Carpi (MO) Italy
*c.bertacchini@igeamedical.com

Open Access Article
The authors, the publisher, and the right holders grant the right to use, reproduce, and disseminate the work in digital form to all users.

Click here to download PDF