Venous thrombosis generation by means of high-intensity focused ultrasound

Ultrasound

Pergamon

lOrigina1

in Med. & Biol., Vol. 21, No. I, pp. 113-I 19, 1995 Copyright 0 1995 Elsevier Science Ltd Printed in the USA. All rights reserved 03Ol-5629/95 $9.50 + .OO

Contribution

VENOUS THROMBOSIS HIGH-INTENSITY C.

DELON-MARTIN,+

C.

GENERATION BY MEANS FOCUSED ULTRASOUND VOGT,+E.

CHIGNIER,~

OF

C. GUERS,+

J. Y. CHAPELON+ and D. CATHIGNOL~ ‘INSERM, Unit.5 281, Lyon, France; and *INSERM, UnitC 331,

Facultt5 de MCdecine Alexis Carrel, Lyon, France (Received 22 July 1993; in final

form 16 May 1994)

Abstract-Sclerotherapy of superficial varicose veins is now performed with chemical agents since physical agents have given only poor clinical results. We investigated the possibility of using high intensity focused ultrasound energy to achieve this goal in an animal model, the rat femoral vein. A specially designed probe delivering ultrasonic energy at a central frequency of 7.31 MHz was constructed and evaluated. Femoral veins of six rats were surgically exposed to a set of between four and seven 3-s exposures at l-mm increments at a power level of 167 W/cm’. At 2 days following the irradiation, control veins were patent while occlusive thrombus was documented by Doppler flow and histological studies in all six of the irradiated veins. No damage to the surrounding soft tissues was noted. We conclude that high-intensity focused ultrasound can be used to induce thrombosis in this animal model.

Key Words: High-intensity Thrombosis,

V&lcose

focused ultrasound, disease, Sclerotherapy.

Therapeutic

INTRODUCTION

ultrasound,

Nonlinear

effects,

Cavitation,

and frequencies in the range of 0.5 to 4 MHz. Recently, Yang et al. (1992) have investigated the responses of rabbit abdominal aorta and inferior vena cava to highintensity focused ultrasound (HIFU) (4 MHz, 1500 W/ cm2). Their treatment consists of 20 X 20 insonation points (5-s exposure/point) with a l-mm increment covering the vessels. They did not observe significant changes in these vessels with Doppler ultrasound and angiography examination. The application of therapeutic HIFU as a physical sclerosing agent of varicose veins is the subject of this article. At present, the classical treatment of varicose disease is either stripping, a surgical technique which is used for saphenous-incompetent veins, or sclerotherapy, which is used for superficial veins and residual varicose veins after stripping. The latter consists of an injection within the vein of a chemical irritant which destroys the endothelium of the vein, leading to thrombosis in about 2 days and fibrosis in about 4 weeks. Treatment with physical agents has been attempted using lasers but with only poor results (Apfelberg et al. 1987). Schultz-Haakh et al. (1989) have shown that ultrasonic exposure can induce endothelial damage in veins, but neither thrombosis nor fibrosis was further established.

The use of focused ultrasonic energy for the treatment of biological structures began in the 1960s. At this time, the aim was to destroy preselected targets located deep within the brain, without any damage to the tissue in the path or surrounding the lesion (Lele 1967; Fry et al. 1970; Gavrilov 1972). This technique had two main drawbacks: first, it was necessary to open the skull before treatment; and second, an accurate noninvasive target localization system was necessary for an accurate positioning of the treatment device. With progress in ultrasound scanners, a wider range of applications has become possible and work on the ultrasonic treatment of diseases or tumors is currently under way in many laboratories: in ophthalmic therapy (Lizzi et al. 1984); for the treatment of discrete liver tumors (ter Haar et al. 1989); of the prostate adenoma (Fry and Sanghvi 1989) and tumors (Chapelon et al. 1992); for the treatment of the bladder (Vallancien et al. 1990). These applications use intensities of about 1 kW/cm* in the focal zone, exposure durations of a few seconds Address correspondence to: Chantal Delon-Martin, INSERM, Unit6 318, CHU A. Michallon, BP 217 X, 38043 Grenoble Cedex, France. 113

II4

Ultrasound

in Medicine

and Biology

In varicose disease, the veins to be treated by sclerotherapy are generally small in diameter, located at a shallow point under the skin. Ultrasonic treatment can be proposed as an alternative treatment for such veins. The treatment probe requires a small focal area and low depth of penetration. For ergonomic considerations, a moderately sized probe is also required. This study aims at showing the feasibility of inducing venous thrombosis by means of a high-intensity ultrasonic system. The design, construction and evaluation of our ultrasonic probe for high-intensity emission are first introduced. Before the irz viva work, in vitro experiments are carried out to overcome difficulties caused by cavitation. The animal model and the experimental protocol used in the in viva work are described. Efficacy of the device in the chosen experimental model is given for a set of acoustic parameters. MATERIALS

AND

METHODS

Therapeutic ultrasonic probe A piezoelectric ceramic (PI-89 from Quartz & Silice) with a diameter of 20 mm and focal length of 20 mm is mounted in a brass holder containing the electrical impedance matching. Its resonance frequency is 7.31 MHz. The acoustic field generated by the probe is measured using a wideband hydrophone with a 0.3-mm diameter active element and a sensitivity of 1.6 mV/bar, specially designed for the measurement of shock waves (Mestas and Cathignol 1990). The acoustic focus is found 19.2 mm in front of the transducer face, and its measured focal length is 0.6 mm wide and 4 mm long. These values are an overestimation of the actual values since the hydrophone active element is of the same order as the focal area. Indeed, theoretical calculations of the focal zone give a diameter of 0.5 mm and a length of 1.3 mm. Afterwards, the available acoustic intensity is measured at the focus to take into account the effects of nonlinear propagation. We used the following procedure: the hydrophone is located at the acoustic focus; an RF signal with a frequency of 7.31 MHz, and an amplitude ranging from 20 mV to 1 V, is applied through a wideband RF amplifier (Kalmus Engineering International) to the transducer which emits the ultrasonic wave; the pressure wave received on the hydrophone is displayed on a digital oscilloscope (2430 A, Tektronix) and recorded on an IBM-PC for further calculation of the intensity. The temporal peak (TP) intensity is calculated using the following formula: ITF =

s

1. T,

where c represents

‘[l+Tp(t)* - dt PC

the sound velocity

in soft tissues,

Volume

2 1, Number

1, 1995

p(t) = V(t)lS, with S being the hydrophone sensitivity and V(t) the voltage signal provided by the hydrophone. V(t) represents the response of the system, the medium and the wideband hydrophone to an RF signal. It is composed of a transitory part and of a stationary periodic part (duration T, beginning to). The TP intensity is calculated during this stationary part. In case of linear propagation, as intensity is proportional to the square of the pressure, it is also proportional to the square of the input voltage. The relationship between TP intensity and the square of the input voltage is roughly linear up to intensities of 400 W/ cm’, and nonlinear above, due to the nonlinear propagation phenomenon. The system can provide TP intensities of up to 548 W/cm*. TP values of intensity are spatially averaged, since they are recorded from the 0.3-mm active element wideband hydrophone. Thus, they are underestimations of the instantaneous spatial peak intensity, a pertinent value when considering cavitation occurrence. Experimental setup The probe can be controlled either in the diagnostic mode for the site localization and subsequent positioning of the probe, or in the therapeutic mode (Fig. 1). In the diagnostic mode, an ultrasonic transducer analyzer controls the transducer in the pulse mode so that the backscattered signal from the tissues gives an A-mode image on the oscilloscope. After a rough positioning of the acoustic focus of the probe near the target tissue, manual displacements in the X, Y and 2 directions are used to ensure an accurate positioning of the probe. Then, the probe is switched to the therapeutic mode. A frequency generator emits an RF signal at the resonance frequency of the probe, to the wideband RF amplifier, and thus to the ultrasonic probe. The TP intensity available at the focus depends on the input voltage of the RF signal. Cavitation experiments In vitro experiments were carried out in pig muscle located behind rabbit skin in order to determine insonation parameters (input voltage, exposure duration) inducing bum lesions. A transmission medium, consisting of a water-filled plexiglass container, closed on one side with a latex membrane (acoustic impedance close to that of human tissues), was inserted between the transducer and the tissues. A thin coupling film was inserted between the membrane and the biological tissues. Different qualities of water, different kinds of membrane and different films were tested to determine the coupling media giving a minimum loss of available energy. During high power acoustic emission, transient acoustic cavitation may occur, resulting

Venous thrombosis

/

L

frequency generator

\

/

J

manual displacements ox, oy, 02 / L

therapeutic /

115

generation l C. DELON-MARTINet cd.

>

mode

\

R.F. amplifier /

\ (7

I

r X

0

I

I

diagnostic

mode 2

oscilloscope

1 degassed

Y

water

Fig. 1. Experimental setup for the femoral vein treatment. The probe operates either in the diagnostic mode for an accurate localization of the site of treatment and for a proper positioning of the probe, or in the therapeutic mode for vein treatment.

in a considerable loss of available energy in the focal zone. In the absence of cavitation, the signal received by a hydrophone positioned far behind the focus (so as not to damage it) is constant during the acoustic emission. But if such cavitation occurs, a nonconstant signal is received on the hydrophone. Thus, the envelope of the pressure wave is recorded during long acoustic emission (typically 1 s) with increasing voltage inputs. Cavitation occurs when the envelope is no longer constant. Another criterion for cavitation detection sometimes used is detection of the half harmonic of the transmitted frequency. In a set of experiments, a wideband hydrophone was placed next to the acoustic focus during an exposure sequence without the use of animals. It was connected to a spectrum analyzer (3588A Hewlett Packard) and the range around the half-resonance frequency was selected for observation. If no signal appears, it means that cavitation does not occur. In vivo experiments Two groups of animals were established: the first one for the feasibility of the method and the second one for the experimental protocol. Feasibility protocol In this group, 17 rats were used to establish the protocol of vein dissection, rat installation and posi-

tioning of the treatment probe by A-mode ultrasound imaging. All insonations were performed with the same acoustic parameters. Macroscopic and Doppler examinations were carried out to evaluate the vein damage in this group. Surgical protocol The superficial femoral vein of old male Wistar reproductive rats (500 g) was chosen for treatment. After anaesthesia with 0.11 mL/lOO g of a solution with 6% of sodium pentobarbital, about 1 cm from each of the right and left veins (control and treated, respectively) was dissected in the femoral triangle. A microsurgical sheet (a latex film, 80 pm thick) and a vein holder (a piece of wire with insulation removed, bent into the form of a gutter, 10 mm long and 3 mm wide), as sketched in Fig. 1, were positioned behind the experimental vein to help its localization by Amode imaging. The only difference in the protocol between the control and the treated veins is the high intensity irradiation performed on the latter. The rat was supported in the supine position in degassed water at 39°C with its head above water and its legs immersed. Before each insonation, A-mode imaging using the treatment probe itself was carried out, the two main echoes from the two lengths of the vein holder

116

Ultrasound

in Medicine

and Biology

being used to find the vein echo between them. This method ensures an accurate positioning of the treatment probe. Then, a sequence of four to seven insonations along the vein, each separated by 1 mm when possible, was performed on the left vein. The TP intensity within the focal zone is 167 W/cm2 and the duration 3 s. Six rats were treated under these exposure conditions. Five rats were examined and sacrificed 2 days after insonation in order to determine if thrombosis occurred. One was examined 2 days after treatment and re-examined 4 weeks later, before sacrifice. Surgical procedures and animal care conformed strictly to the guidelines of the National Institute of Health and Medical Research (Decree No. 68 139 of 9 February 1968). The analysis of results was carried out before death by clinical observations under an optical surgical microscope (OPMI 1, Carl Zeiss, Germany). A Doppler with a 20-MHz probe was used to determine the patency of the vein. Pertinent samples from control and treated veins were kept for further morphological observation, using histological methods for vein wall examination and scanning electron microscopy for endothelial layer observation. Histological and ultrastructural techniques The veins were fixed in situ using a 2% buffered glutaraldehyde (0.4 M cacodylate buffer), pH 7.45,400 mOsm. The veins were then dissected and divided into two fragments, one for histological methods and the other for scanning electron microscopy. The fragments were separately fixed for another 2 h, buffer-rinsed overnight and processed differently. Fragments intended for histological observation were dehydrated in ascending ethanol series and embedded in epoxy resins. Finally, they were placed in capsules with the aid of a dissecting microscope to ensure that the tissue sections could be obtained from the full cross diameter of the vessel. Specimens were cut serially into 5-pm-thick slices using an ultramicrotome (OMU 4, Reichert Jung, Vienna, Austria). The sections were mounted in glass slides, stained with Masson’s trichrome and examined under a Polyvar microscope (Reichert Jung). Light microscopy evaluations of the samples included the analysis of the cellular and extracellular components, hemorrhage in the wall, intraluminal thrombi and their relation with the vein wall. Fragments intended for ultrastructural observations of the endothelial layer were opened, completely desiccated with increasing concentrations of acetone using the dispersion method and dried using the critical point method with liquid CO,. The whole sample was then mounted in aluminium stubs, coated with gold

Volume

21, Number

I, 1995

palladium sputtering and examined microscope at 15 kV.

on an S800 Hitachi

RESULTS Cavitation experiments The simple criterion we used allowed us to determine that degassed bidistilled water (percentage in O2 lower than 12%), a thin condom latex membrane and a degassed water film form the best coupling medium. In the same manner, it was observed that acoustic ultrasonic gels, such as those used for B-mode imaging, must not be used in these experiments. Feasibility group Occlusive or partly occlusive thrombi were observed on control veins due to the vein dissection and/ or surgical manipulation. To avoid this surgical artifact, it appears necessary to coagulate the epigastric vein with bipolar pliers at least 2 mm apart from the femoral vein and to ligature the deep femoral vein with microsurgical thread. Vein localization is the most sensitive point in this study. The method used consists first in looking for the two strong echoes due to the two lengths of the vein holder, and second, in positioning the probe between them on the weaker echo due to the vein itself. The problem arises from the surrounding structures (skin, femoral artery, tissues or microsurgical sheet) which may also emit strong echoes. These echoes must not be confused with the vein holder. Burnt zones in surrounding tissues or in the skin were obtained in this group. Moreover, it was noticed that vein holder localization was easier when the rat was lying in the supine position instead of the vertical position. In two of those experiments, the femoral arteries were insonated. When observed 2 days later, two arterial thrombi (about 6 mm long) were filling the lumen. This result suggests that HIFU may also induce arterial thrombosis. A new thrombus model could be created which may interest researchers involved in arterial thrombosis studies. Results of the experimental group Table 1 summarizes all the observations of this group. Briefly, five of six rats were sacrificed 2 days after insonation; the remaining animal was also observed 2 days after treatment and as thrombosis occurred, it was left alive to be reexamined 1 month later to see if thrombus remained or disappeared. The six control veins were macroscopically normal and Doppler examinations showed that they were patent and that no obstacle to blood flow was detected. Little adherence of the surrounding tissues was noted due to the

Venous thrombosis

v1 4

9

generation 0 C. DELON-MARTINet al.

117

recent dissection. Slight inflammatory process of surrounding tissues was sometimes noted. They were similar to burnt spots, probably due to incorrect probe positioning and insonations around the vein. Among the six treated veins, two had thrombi of 6 to 7 mm, three had thrombi of I 1 to 13 mm and one an 18mm thrombus. These lengths indicated that the thrombi extended beyond the exposure sites. The thrombi were contained within the vein lumen, but they did not adhere totally to the vein wall. The only rat observed twice had, at the first observation (2 days after treatment), a thrombus of 11 mm in length, but 1 month later the vein was patent at the Doppler examination and no further thrombus was observed, after vein dissection. Histological examinations (Fig. 2A-D) of control veins showed a free lumen, but some microthrombi were seen close to the endothelial layer, probably due to the in situ fixation method (Fig. 2A). In the treated veins, a general pattern was observed (Fig. 2B): the vein lumen was occluded by a thrombus, adhering partially to the vein wall (Fig. 2C). The thrombi presented zones containing noncellular material and zones rich in red cells. Channels inside the thrombi were observed in two cases. The vein wall also presented hematic infiltration in two different cases (Fig. 2D). Ultrastructural observations (Fig. 3) in control veins showed a well-preserved endothelial layer (Fig. 3A), with microthrombi in some places. In treated samples, the examination of the endothelial layer was not possible due to a coating of red cells, trapped within fibrinous material, constituting a thick thrombus (Fig. 3B). Platelets were rarely observed, indicating that the thrombus was mainly of a fibrinocruoric type.

DISCUSSION The results presented here, demonstrate the feasibility of venous thrombosis in vivo, using HIFU. The effects induced in vivo on biological tissues by ultrasonic insonations with intensities of around several hundred watts per square centimeter are not known to date. The TP intensity of the treatment used in our protocol (167 W/cm2) is about 10 times weaker than in thermal ablation techniques, where the aim is the destruction of diseased tissues with intensities of around 1 kW/cm*. The effects we observed here result from thermal damage of the vein wall without vein disruption or surrounding tissue damage. The intensity used to achieve venous thrombosis is between that used in thermal ablation techniques and that employed in ultrasonic hyperthermia treatments, which requires a small total temperature (up to 45°C) and reversible effects in healthy tissues.

Ultrasound in Medicine and Biology

Volume 21, Number 1, 1995

Fig. 2. Light microscopy observation showing: (A) control sample (nontreated vein) and (B-D) treated samples. (A) The vein wall appears to be normal. The lumen (L) is free of thrombosis. Well-preserved intimal layer (large white arrow) is observed. In places (thin white arrow), microthrombi are seen close to the endothehal layer (see text). Original magnification X48. (B) General pattern is observed: the vein lumen (L) is occluded by the thrombus (Th). Original magnification X8. (C) The thrombus (Th) partially adheres to the vein wall (white arrow). It presents zones containing noncellular material (black arrow) and zones rich in red cells (large black arrow). Original magnification ~48. (D) High magnification from (B), showing the vein wall with a hematic infiltration (H). Original magnification ~48. L: lumen; Th: thrombus; E: external side; H: hemorrhage.

The comparison between the results obtained by Yang et al. (1992) and us, suggests a different behavior of vessels with a large diameter (> 1 cm) and those with a small diameter (<2 mm). Since blood flow is about 800 times higher in aorta and vena cava than in small diameter vessels, an increase of heat due to HIFU may be compensated by heat convection in the large diameter vessels and not in the small diameter vessels, allowing thrombotic development in the latter. No evidence of a cavitation effect was observed in our samples. As there is no standard method to detect cavitation occurrence, we used a physical criterion which is not used by other authors. This criterion provides an upper limit of TP intensity of 300 W/cm2 for the treatments. Nevertheless, we also carried out experiments in degassed water from a water tank using half harmonic detection to find out whether cavitation

occurs during our in viva treatments. This method confirms that no cavitation occurs during insonations under our exposure conditions. We eliminate the possibility that venous thrombosis may be produced by a pressure radiation phenomenon. Nowadays, there is no evidence in the literature that such a phenomenon may provoke a physiological effect. These preliminary findings are observed in a model structurally similar to human vein, suggesting that the technique holds promise for the transcutaneous induction of thrombosis in human varicose veins. The dimensions of the thrombi (which are much greater than the length of the exposed portion of the vein) indicate that the thrombi probably grow from a portion of damaged endothelium. This is confirmed by the fact that most of the sections containing thrombi are nonadhering. This pattern of thrombus growth is

Venous

thrombosis

generation

0 C. DELON-MARTIN et crl.

119

Fig. 3. Scanning electron microscopy observation of venous lumen corresponding to the control sample (A) and treated vein (B). (A) Luminal surface of the control sample. Endothelial lining is clearly observed without thrombus formation. Original magnification x3600. (B) The endothelial layer is totally hidden by an organized thrombus. The fibrin network retains red cells. Original magnification x2500.

characteristic of techniques which damage or disturb the normal functioning of the endothelium. Yet, these thrombi are potentially hazardous in that they may embolize as the endothelium is being repaired. One of the limitations of these preliminary observations is the unknown time course of these changes. The single observation done at 1 month on one of the rats from the experimental group, suggests that the ultrasonic parameters used in this work are not sufficient enough to induce a long-term vein occlusion. Another limitation of the method is the unknown dose-response curve. A threshold intensity may appear for a given exposure duration, as with HIFU techniques. The last limitation is the possible induction of cavitation due to the intervening soft tissues, when working with nonsurgically exposed veins. This technique therefore requires more evaluation before it can be proposed as an alternative to established techniques. Acknowledgemenrs-We thank Mrs. A. Rivoire and Mrs. A. Thillier for their assistance with scanning electron microscopy and histological samples and Mr. A. Mattias and Mr. R. Jarry from INSERM Il.281 for their technical assistance.

REFERENCES Apfelberg, D. B.; Smith, T.; Maser, M. R.; Lash, H.; White, D. N. Study of three laser systems for treatment of superficial varicosities of the lower extremity. Lasers Surg. Med. 7:219-223; 1987.

Chapelon, J. Y.; Margonari, J.; Vernier, F.; Con-y, F.; Ecochard, R.; Gelet, A. In viva effects of high-intensity ultrasound on prostatic adenocarcinoma Dunning R3327. Cancer Res. 52:6353-6357; 1992. Fry, F. J.; Kossof, G.; Eggleton, R. C.; Dunn, F. Threshold ultrasonic dosages for structural changes in the mammalian brain. J.A.S.A. 48:1413-1417; 1970. Fry, F. J.; Sanghvi, N. Ultrasound localization and therapy system. Patent International Publication # WO 89/07909; 1989. Gavrilov, L. R. Application of high-intensity focused ultrasound for the local treatment of biological tissues. Sov. Phys. Acoust. 17:287-301; 1972. Lele, P. P. Production of deep focal lesions by focused ultrasoundcurrent status. Ultrasonics 5: 105-I 12; 1967. Lizzi, F. L.; Coleman, D. J.; Driller, J.; Ostromogilsky, M.; Chang, S.; Greenall, P. Ultrasonic hyperthermia for ophtalmic therapy. IEEE Trans. Sonics Ultrason. 31:473-480; 1984. Mestas, J. L.; Cathignol, D. Capteurs de pression pour le controle de gtnerateurs d’ondes de choc Clectrohydraulique. I” Congrtts Francais d’Acoustique C2:1287- 1290; 1990. Schultz-Haakh, H.; Li, J. K.; Welkowitz, W.; Rosenberg, N. Ultrasonic treatment of varicose veins. Angiology 40: 129- 137; 1989. ter Haar, G.; Sinnett, D.; Rivens, I. High intensity focused ultrasound-a surgical technique for the treatment of discrete liver turnours. Phys. Med. Biol. 34:1743-1750; 1989. Vallancien, Cl.; Chopin, D.; Thibault, P. H.; D’Avila, C.; Veillon, B.; Brisset, J. M.; Andre-Bougaran, J. Extra-corporeal focalized piezo-electric hyperthermia. First experimentation. Abstracts of the IXth Congress of the European Association of Urology 1990:285. Yang, R.; Griffith, S. L.; Rescorla, F. J.; Galliani, C. A.; Ehrman, K. 0.; Fry, F. J.; Wessling, S. M.; Grosfeld, J. L. Feasibility of using high intensity focused ultrasound for the treatment of unresectable retroperitoneal malignancies, Abstracts of the 36th Congress of the A.I.U.M.. San Diego; 1992.