Combination of thermal and cavitation effects to generate deep

By Walter Lee,2014-05-19 01:35
19 views 0
Combination of thermal and cavitation effects to generate deep

    Combination of thermal and cavitation effects to generate deep lesions with an

    endocavitary applicator using a plane transducer: Ex vivo studies.

    David Melodelima, Jean Yves Chapelon, Yves Theillère and Dominique Cathignol.

    INSERM, Unité 556, 151 Cours Albert Thomas, 69424 Lyon, France

Corresponding author: David Melodelima

    Pho. : (33) 4-72-68-19-30

    Fax : (33) 4-72-68-19-31

    E-mail :

    Running title : Combination of thermal and cavitation effects


    Combination of thermal and cavitation effects to generate deep lesions with an

    endocavitary applicator using a plane transducer: Ex vivo studies.


    In the HIFU field it is well known that the cavitation effect can be used to induce lesions of larger volume. The principle is based on the increase in the equivalent attenuation coefficient of the tissue in the presence of the bubbles created by cavitation. The elementary lesions produced by combination of cavitation and thermal effect, using focused transducers, were spherical and developed upstream of the focal point. This paper presents a method that combines cavitation with a thermal effect to obtain deeper lesions using a plane transducer rather than a focused one. The cavitation effect was

    2produced by delivering intensities of 60 W/cm at the face of the transducer for 0.5 s.

    The applicator was then rotated through 90? at a constant speed of between 0.5 and 1.5?/s. During this rotation, ex vivo tissues were exposed continuously to an acoustic

    2intensity of 14 W/cm to combine the cavitation effect with a thermal effect. The necroses were on average twice as deep when the cavitation effect was used compared with those obtained with a thermal effect alone. Observed macroscopically, the lesions have a very well delimited geometry. Temperature measurements made at different angles of treatment have shown that they were coagulation necroses.

    Key words: Ultrasound, thermal ablation, cavitation, endocavitary, plane transducer, high intensity ultrasound



    High intensity ultrasound affects biological tissue by different mechanisms, the two most important of which are thermal effects and cavitation effects. The great surge in interest in HIFU (high intensity focused ultrasound) over the last ten years have led to numerous ex

    vivo, in vivo and clinical studies. Analysis of the damage suffered by the biological tissue has given a better understanding of the importance of the mechanisms activated in the formation of coagulation necroses (Fry et al. 1955, Hynynen et al. 1993, ter Haar et al. 1989, Chapelon et al. 1999, Chavrier et al. 2000). When tissue destruction is due to purely thermal effects, the shapes of the elementary lesions, which are very well defined and reproducible, depend on the geometry of the transducer employed. The lesions are ellipsoidal if HIFU is used (Fry et al. 1970), cylindrical for cylindrical transducers (Deardorff and Diederich 1998), or in the shape of a rectangular parallelepiped if a plane transducer is used for sonication (Lafon et al. 1999, Melodelima et al. 2003).

    Numerous authors have noted the appearance of a cavitation effect leading to a change in the geometry of the necroses when the intensity of the ultrasound wave is increased (Chapelon et al. 1990, Fry 1993, Lizzi 1993, Sanghvi et al. 1995, Watkin et al. 1996). The HIFU-induced lesion grows broader, moves closer to the transducer and becomes tadpole-shaped. There are a number of therapeutic applications where it is useful to generate the cavitation effect. For instance, studies in the field of shock-wave lithotripsy (Cathignol et al. 1995, Coleman et al. 1996) have clearly demonstrated experimentally that the action of the pressure release waves on the bubbles made it possible to increase the rate at which the kidney stones are destroyed. Coleman et al. (1987) has also shown that the condensation liquid accumulated in the bubbles during the first cycles of stable cavitation is expelled as


    "microjets" as the bubbles implode. These "microjets" play a decisive part in attacking the solid/liquid interface, leading to the destruction of the stone.

    When ultrasound is used for surgical applications, the effects of uncontrolled cavitation produce irregular lesions, called hemorrhagic necroses, with damage outside the target area. This point has been clearly noted by several authors (Fry et al. 1970, Sibille et al. 1993, Chapelon et al. 2000), who recommended avoiding these non-linear effects if the intention was to induce a coagulation necrosis with a well-defined geometry. Studies on gels (Holt and Roy 2001) and on ex vivo (Lele 1987, Clarke and ter Haar 1997) and in vivo

    (Hynynen 1991) tissue have nevertheless demonstrated that the appearance of cavitation in the tissue causes a large temperature increase, which it may be judicious to use.

    These observations were the starting point for research involving the use of controlled cavitation to enlarge the dimensions of the necrosed region. For this, some authors (Brayman et al. 1995, Chang et al. 2001, Chen et al. 2003) injected contrast agents, generally used to improve the imaging, into the tissue before starting treatment. Adopting a different approach, Sokka et al. (2002) proposed emitting a very high intensity ultrasound wave (around 300 acoustic watts) of brief duration (0.5 s) to generate microbubbles by cavitation. This exposure was followed immediately by a lower intensity insonification (around 20 acoustic watts) for 19.5 s. This procedure could be used to combine cavitation and the thermal effect and to induce larger necroses than those observed when no cavitation was produced. However, although the volume of such lesions was larger, this approach could not induce a necrosis beyond the focal point.

    Among the explanations advanced to account for the increase in the size of the lesions, Chavrier et al. (2000) proposed that the energy absorbed by the bubbles generated during cavitation was redistributed by spherical waves, the frequencies of which corresponded to harmonics above the excitation frequency. This implied a local increase in the attenuation


    around the bubbles, which then absorbed much of the energy emitted by the transducer and prevented a deposit of heat beyond the focal area. Holt et al. (2002) proposed that the temperature increase following cavitation was due to the viscous damping between the bubbles and the tissue.

    With the same idea of necrosing greater volumes, this paper presents a new approach using the cavitation effect to increase the depth of necrosis. For this, an endocavitary applicator equipped with a plane transducer is used. Cavitation is obtained with an acoustic

    22intensity of only 60 W/cm, which is then reduced to 14 W/cm, corresponding to an energy

    level sufficient to create a coagulation necrosis. The aim of the tests presented in this study is to compare the size of the lesions in the presence or absence of cavitation and to ensure, by temperature measurements, that the thermal dose required for necrosis is reached over the whole of the affected depth when the cavitation effect is used.


    The applicator (Fig. 1) consisted of a 3.8-mm O.D. stainless steel tube with a 0.1-mm thick wall ending in a cone-shaped brass tip. The plane active surface embedded in the brass

    2section is a 310 mm P762-type PZT piezoceramic air-backed transducer (Quartz & Silice, Nemours, France) operating at 5.16 MHz. The central lumen of the tube provided electrical RF connections and a path for the cooling water. RF connections were made via a miniaturized 50 cm long, 50 Ω coaxial cable with a 0.9-mm O.D. The outer ground conductor

    was connected to the external face of the transducer. The inner conductor reached the internal face of the transducer through the lumen of the tube. The transducer was glued with Stycast 2651 epoxy resin (Emerson & Cuming Europe n.v., Westerlo-Oevel, Belgium) to ensure that it was watertight and to hold it in place. The stainless steel tube was filled with RTV 143


    silicone (Rhône Poulenc Italia, Milano, Italy) to hold in place all the elements passing through it and so to strengthen the applicator assembly. To limit energy loss during transmission of the electric signal, a connection with a capacitor-inductor network was included to match the transducer. The inductor (1.8 µH) and the capacitor (300 pF) were placed respectively in parallel and in series with the coaxial-transducer assembly. The initial electrical impedance of the applicator (Zr=20 Ω (real part of the impedance); Zi=-50 Ω (imaginary part of the

    impedance)) measured with an HP 4285A impedance meter was Zr=50.1 Ω and Zi=-3 Ω after

    electrical matching. The reflected power was then close to zero. Using the acoustic balance technique (Davidson 1991), the electroacoustic efficiency of the applicator was measured to 65% at 5.16 MHz.

    The external face of the transducer was cooled by a continuous flow of degassed water circulating the length of the transducer. The water cooling circuit was maintained at 25?C and was driven by a Masterflex peristaltic pump (Cole Parmer Instrument Co., Chicago, USA) at a

    -1flow of 0.15 L.min. The temperature of the active face was maintained at about 45?C during sonication. By adjusting the cooling flow, the maximum temperature may be slightly shift away from the applicator surface, thereby increasing the depth of therapeutic heating

    (Deardorff and Diederich 2000). The applicator was fixed on a PVC support (Fig. 2) and its position was controlled by a motion stage with a control unit (Microcontrol ITL09, Newport Corporation, Irvine, CA). Three types of movement were possible: horizontally, vertically along the applicator axis and rotationally around the applicator axis. Electrical energy was supplied by a Kalmus 150 CF amplifier (Engineering International, Woodinville, WA) driven by a HP 8116A sine wave generator (HP GmbH, Böblinger, Germany). Incident and reflected power were measured using a wattmeter/reflectometer (Rohde & Schwarz, Munich, Germany) at the amplifier output. The function generator and the position controller were piloted by a computer using the Testpoint software (CEC, Billeria, Massachusetts, USA). The system


    enabled the amplitude, duration of exposures and angle of treatment of the sonications to be programmed. A Hawk 2102 XDI ultrasound scanner (B-K Medical, Norderstedt, Germany) and an imaging probe operating at 12 MHz with a linear array were used to visualize the cavitation inside the tissue.


    All tests were performed ex vivo on 43 parallelepiped-shaped samples

    3(10010030 mm) of pig's liver. The time between the animal's slaughter and the use of its liver was unknown, so each sample was degassed with a vacuum pump (0.7 bars for 30 min) before the experiment to remove any bubbles that might have formed during storage and thus reproduce in vivo conditions more closely. The samples were then immersed in a tank of degassed water thermostatically maintained at 37?C. A thermocouple was inserted in each sample to ensure that the temperature was close to 37?C before starting exposure. A hole was punched in the middle of each sample attached to a PVC support (Fig. 2). The flow of degassed water used to cool the transducer was set to 0.15 L/min.

Determination of the cavitation threshold.

    A first series of tests was conducted to determine the acoustic intensity threshold from which the applicator would produce the cavitation effect if the exposure time was 0.5 s. The

    2acoustic intensities tested to determine the threshold were 30, 40, 50, 60 and 80 W/cm.

    During these exposures, cavitation was detected with the 12 MHz ultrasound imaging probe positioned perpendicularly to the applicator (Fig. 3).


Comparison of the depth of elementary lesions with and without cavitation

    A second series of tests was conducted to show that the cavitation produced in this way could be useful to increase the depth of the necroses. Two elementary lesions were produced on three liver samples. The first elementary lesion (lesion A), regarded as a

    2reference, was induced by exposing the tissues to an acoustic intensity of 14 W/cm for 20 s.

    These parameters are known to induce coagulation necroses by purely thermal effect (Lafon et al. 1998). The second elementary lesion (lesion B) was produced by first performing a 0.5-s.

    2exposure at an acoustic intensity of 60 W/cm, followed immediately afterwards by a 20-s

    2exposure at 14 W/cm. The aim of this methodology was to determine the contribution of the cavitation generated by the short high intensity exposure in terms of increased depth of necrosis. Three samples were used during this experiment to ensure the reproducibility of the results.

    Comparison of the depth of sector-based lesions with and without cavitation

    A final series of tests was conducted on 32 samples to see whether the size of the lesions could be increased by combining the cavitation and the thermal effect even when the applicator was rotated on its axis through angular openings of 90?. Four different rotation speeds were tested: 0.5, 0.8, 1 and 1.5?/s. Four reference groups (one for each rotation speed

    2 tested), each consisting of three samples, were treated at an acoustic intensity of 14 W/cmemitted continuously during rotation (sequence 1 see figure 4a). These groups of samples were used to determine the depth of the necroses induced by a purely thermal effect. Three samples were used for each rotation speed tested to ensure the reproducibility of the results. Then the cavitation effect was combined with the thermal effect, using a group of 20 samples.


    For this procedure, four groups (one for each rotation speed tested) of five samples were used, with the same four rotation speeds as before (0.5, 0.8, 1 and 1.5?/s). Five samples were used for each set of parameters tested to ensure the reproducibility of the results. On these twenty

    2samples, the cavitation effect was always produced with a 0.5-s exposure at 60 W/cm.

    Immediately afterwards, the ultrasound applicator was rotated through 90? and a thermal effect was produced by a continuous emission at 14 W/cm? during the rotation (sequence 2) to overlap the previous (Fig. 4b). Temperature measurements were realized when these sector-based lesions were created. Four T-type thermocouples were used to determine the maximum temperature elevation at four different positions: 2.5, 5.0, 10.0 and 15.0 mm along the axis of acoustic propagation. The temperature was measured along the axis of propagation immediately after production of the cavitation effect alone. The temperature was also measured at three angular openings (30, 60 and 90?) along the axis of propagation for all rotation speeds tested. The thermocouples were fitted into 0.7-mm external diameter needles and glued in position, with their active extremity projecting 2 mm beyond the needle. Thanks to this reinforcement of their mechanical structure, they could be introduced as far as the measurement point without being subject to any major mechanical constraints that might lead to their destruction. To avoid interference with the ultrasound and giving inaccurate results, they were inserted into the tissue immediately after sonication. The four thermocouples were fixed to a PVC support so that they could be introduced simultaneously and provide sets of linked results. They were connected to HI 9351 digital thermometers (Hanna Instruments, Tanneries, France), which stored the maximum temperature values between resets. Measurements were made both on the samples treated by sequence 1 and on those treated by sequence 2. The computer piloting the signal generation device controlled the power values transmitted and reflected to the transducer and stored the values delivered during each treatment.


    The liver samples were frozen immediately after treatment to make the tissue easier to section and examine. When the tissue temperature was close to 6?C, the sample was sliced

    along a plane perpendicular to the applicator axis at mid-height from the transducer. Lesions were inspected macroscopically. Coagulated and untreated tissue is easy to distinguish in pig's liver as the former is off-white in colour or even dark, like cooked liver, at the hottest points. Two axes issuing from the centre of the applicator and touching the outer edges of the lesion were used to assess the value of the angular section of the necrosis.


Determination of the cavitation threshold

    Figure 5a shows a sonogram of the experimental set up used to detect the cavitation effect in the tissue. The ultrasound applicator can be seen positioned opposite a piece of liver. The position of the treatment transducer (grey rectangle) and the acoustic propagation axis (white narrow) are indicated in this figure. Because the body of the applicator was made of stainless steel the long rectangular hyperechogenic area located downstream of the transducer corresponds to an imaging artifact. Figure 5b is a sonogram obtained immediately after

    2performing a 0.5-s ultrasound exposure at an acoustic intensity of 60 W/cm in the tissue.

    Hyperechogenic points appear along the acoustic axis and correspond to the cavitation phenomenon induced by this exposure. The tests conducted with acoustic intensities of below

    260 W/cm did not show the appearance of cavitation. The experiments carried out with

    2acoustic intensities of 80 W/cm for 0.5 s show images similar to those in Fig. 5b.

Comparison of the depth of elementary lesions with and without cavitation


Report this document

For any questions or suggestions please email