A high-intensity US probe designed for intraductal tumor destruction: experimental results

NEW METHODS & MATERIALS A high-intensity US probe designed for intraductal tumor destruction: experimental results Frédéric Prat, MD, PhD, Cyril Lafon, PhD, Jacqueline Margonari, Fabien Gorry, MD, Yves Theillère, Jean-Yves Chapelon, PhD, Dominique Cathignol, PhD

Background: Many digestive tract tumors spread inside the lumen and are not amenable to curative surgical treatment. An intraluminal method of tumor destruction would be useful for palliative or even curative purposes. Highintensity ultrasound (US) is suitable for such purposes. Our objective was to perform experiments with animal models that would lead to development of a high-intensity US probe for intraductal tumor destruction suitable for insertion through a large-channel endoscope. Methods: The active part of the high-intensity US applicator consisted of a water-cooled piezoceramic plane transducer (3 × 10 mm) operating at 5 MHz for deep or 10 MHz for shallow tissue penetration. A cylinder of tissue was destroyed by means of rotating the transducer on its axis through a flexible shaft. Experiments were conducted in vitro on livers of butchered pigs (10 lesions), in vivo on exteriorized pig livers (15 lesions), and on metastatic Dunning tumors (AT2 subline) implanted subcutaneously in 28 rats (treated n = 16, controls n = 12). Results: In experiments on pig livers, high-intensity US induced highly reproducible cylinders of coagulation necrosis (diameter 20 ± 1 mm, height 8 ± 1 mm) with sharply demarcated and serrated boundaries. The exposure duration to achieve such lesions was 5 minutes. Regions of coagulation necrosis obtained in vivo were similar in size and shape. All 12 control rats died or were killed because of diffuse cancer by day 15 after implantation; 64% of the treated rats were tumor free 30 days after treatment, and 36% had local recurrences. Conclusion: This high-intensity US probe induces highly reproducible cylinders of coagulation necrosis and is effective against tumors in animals.

A large proportion of digestive tumors develop initially inside the circumference of the lumen. This is Received July 1, 1998. For revision October 15, 1998. Accepted March 11, 1999. From INSERM, Lyon, France, and Service des Maladies du Foie et de l’Appareil Digestif, Le Kremlin-Bicêtre, France. Supported in part by ARC grant 6833 and by the Assistance Publique, Hôpitaux de Paris. Reprint requests: F. Prat, Maladies du Foie et de l’Appareil Digestif, CHU Bicêtre, 78 Rue du Général Leclerc, 94275 Le Kremlin-Bicêtre, France. Copyright © 1999 by the American Society for Gastrointestinal Endoscopy 0016-5107/98/$8.00 + 0 37/69/98451 388

GASTROINTESTINAL ENDOSCOPY

the case for many esophageal and colorectal tumors, and for almost all tumors of the biliary tract, such as ampullary cancer and cholangiocarcinoma. Most of these cancers are not amenable to curative resection because of extent of tumor at diagnosis, patient age, or the presence of comorbid conditions.1 A palliative combination of chemotherapy and radiation therapy therefore is offered to these patients. However, these treatments are neither always nor rapidly effective, and they have not proved useful in the management of biliary tumors. Laser photocoagulation sometimes is used, but it is relatively expensive and demanding, because tissue destruction is superficial, and several sessions generally are needed to achieve symptomatic improvement. Patients with these diseases might benefit from a rapidly effective intraluminal debulking technique in terms of longlasting palliation and perhaps improved survival. In the case of biliary tumors, the need for stent exchanges might be markedly reduced. High-intensity US can produce radical and controllable tissue necrosis.2 A high-intensity US probe has been developed and tested in animal models. A modified version of the experimental probe has been specifically designed for intraductal tumor therapy during digestive endoscopy. MATERIALS AND METHODS Description of the device The US applicator used in this preliminary study has been described in detail elsewhere,3,4 and a brief description is as follows: The applicator consists of a stainless steel tube 8 cm long and 3.8 mm in diameter with a cone brass tip at one end. The active part embedded in the tip is a water-cooled piezoceramic plane transducer (3 × 10 mm2) that can be operated at 5 or 10 MHz, depending on the desired depth of heating. In theory, deeper lesions can be obtained by use of a lower frequency, provided that US intensity is increased to compensate for lower attenuation. Water-cooling eliminates heating of the transducer and facilitates acoustic coupling between transducer and targeted tissues. A US transparent envelope is sealed over the applicator tip to ensure water-tightness of the cooling circuit. A flexible shaft for remote rotational motion control is bonded to the end of the probe opposite the active section. Electrical connections are secured with miniaturized 50-ohm (Ω) coaxial cable. A thermocouple is bonded to the external face of the transducer to monitor temperature during treatment and to warn the operator of possible damage. The connecting end of the probe can be attached to a holder and rotated on its axis as many times as desired. The rotational angle of the probe is controlled with a micrometric screw. Electrical power is delivered in a continuous mode with a sinusoidal wave generator (Hewlett-Packard GmbH, Böblinger, Germany) through a power amplifier (Kalmus 150C RF; Engineering International, Wondinville, Wash.) VOLUME 50, NO. 3, 1999

High-intensity US probe for intraductal tumor destruction

A

F Prat, C Lafon, J Margonari, et al.

B

Figure 1. A, Photograph shows US applicator projecting from the accessory channel of a therapeutic duodenoscope with its brass tip and piezoceramic plane transducer. Water-tight envelope is not shown. B, Overview of the device: 2-meter long flexible shaft, with active part at one end and electrical connection, guidewire port, and handle at the other. The generator is controlled with a timer to allow manual triggering of emission and automatic cutoff. A probe adapted for digestive endoscopy has been built for future clinical trials (Fig. 1). This new probe has the same functional characteristics as the prototype. The transducer has been slightly reduced in diameter and length so that it can be accepted by the 4.2 mm accessory channel commonly available on therapeutic duodenoscopes. It is connected to a 2 m long flexible shaft and has a channel for passage of a 0.021-inch guidewire. In vitro experiments In vitro experiments were conducted with pig livers from a butchery. Livers were kept frozen at –40°C. Before experiments, the liver was thawed, degassed with a vacuum pump (0.3 bars for 15 minutes), and immersed into a bath of degassed water thermostatically maintained at 37°C. The transducer was inserted deeply into the liver. To obtain cylindrical lesions, sequential US shots were applied. The transducer was rotated between two shots, each single shot inducing an elementary lesion. Rotations were remotely controlled with a micropositioning system (Microcontrole, Evry, France). Cylindrical necrosis was obtained through summation of 20 shots separated from each other by an 18-degree rotation angle (18 degrees was found in previous experiments3,4 to allow homogeneous tissue necrosis while avoiding uncontrollable build-up of overlapping lesions). The duration of the first shot was set at 20 seconds to establish the lesion; the duration of the subsequent 19 shots was reduced to 10 seconds to take into account the increase in temperature caused by preceding shots and prevent lesion build-up. Five seconds was needed between shots to rotate the transducer. Livers were frozen after treatment to facilitate examination of the lesions, which were examined macroscopically and measured with a calliper square. The lesions were easily differentiated from the uncoagulated surrounding tissue by a whitish color that contrasted with the brownish color of the intact liver. VOLUME 50, NO. 3, 1999

In vivo experiments In vivo experiments were performed on exteriorized pig livers. Male pigs with an average weight of 30 kg were used for the study. Anesthesia was obtained under airway intubation through a tracheostomy with a mixture of 4 mL ketamine (50 mg/mL) and 6 mL 2% xylazine administered intramuscularly and followed by slow intravenous infusion of nesdonal (0.4 gm/100 mL). A large laparotomy was performed and the liver was exteriorized. The livers were treated in the same manner as for the in vitro experiments with introduction of the applicator into the liver parenchyma to a depth of approximately 2 cm. To take into account previous reports that indirect damage to the liver tissue surrounding the treated volume might occur because of the lack of blood supply after treatment,5 pigs were sacrificed 3 hours post-treatment while still under general anaesthesia. Histologic sections were stained with hemalun and eosin and saffron. Experiments with animal tumors R3327 Dunning tumors were produced in Fischer Copenhagen rats (Harlan, Gannat, France). Carcinoma cells (2 × 106; AT-2 subline) were injected subcutaneously in the abdominal wall of the rats. After 8 days, a solid tumor was present that was excised and minced; 20 mg pieces were then implanted in 28 rats. Under these conditions and without treatment, death due to tumor progression is expected to occur within 4 weeks of implantation in 100% of animals.6 Sixteen animals were treated and 12 were used as controls. Treated animals received high-intensity US shots 8 days after implantation, when the subcutaneous tumor had reached a diameter of 8 mm. A single treatment was performed with a 5 MHz transducer mounted on a handpiece with the rat abdomen immersed in degassed water. Because of the tumor location and small size, shots were performed with the transducer applied to the skin surface and over both sides of the tumor, tangentially to the abdomen to avoid intra-abdominal injury. The treated tisGASTROINTESTINAL ENDOSCOPY

389

F Prat, C Lafon, J Margonari, et al.

High-intensity US probe for intraductal tumor destruction

Figure 2. Three-dimensional view of a cylindrical volume of US-induced necrosis (20 shots separated by 18-degree rotational angle at a frequency of 10 MHz and a power intensity of 14 W/cm2) obtained in vitro in a pig liver. The central hole results from the passage of the probe.

Figure 3. Sectional view of cylindrical US-induced necrosis (20 shots separated by an 18-degree rotational angle at a frequency of 10 MHz and a power intensity of 14 W/cm2) obtained in vivo in an exteriorized pig liver. The central dark area results from the passage of the probe. Each indentation of the lesion boundary results from an elementary shot.

sue volume was a portion of a cylinder, usually composed of 5 or 6 elementary lesions. The rotation angle between shots was set to 10 degrees rather than 18 degrees to take into account increased thermal convection at the skin surface between tumor and water. Control rats were subjected to exactly the same procedure except that US exposure was eliminated. They were observed until death or were killed if not dead in 3 weeks. Treated rats were observed until death or at least 50 days after treatment. Experiments on living pigs and on rat tumors conformed with the uniform requirements for animal experiments and were in accordance with INSERM requirements for animal experimentation and legal conditions set by the French national commission on animal experimentation.

5 were made at 5 MHz, 24 W/cm2. The intensity was increased to compensate for the perfusion effect. Lesions were the same size as those produced in vitro. The shape was slightly different, elementary lesions being better demarcated (Fig. 3). This is explained by a perfusion effect that decreases the attenuation coefficient. Microscopic examination revealed consistently diffuse and sharply delineated lesions of coagulation necrosis; hepatocytes, when still identifiable, had condensed chromatin and vacuolized cytoplasm.

RESULTS In vitro experiments Fifteen lesions were attempted and produced in butchered pig livers. Ten lesions were performed with the applicator operating at 10 MHz and a power intensity of 14 W/cm2 and 5 were obtained at 5 MHz and 19 W/cm2. The boundaries of the lesions were always easily identified from the untreated tissue by a sharp demarcation. Lesions obtained with the 10 MHz applicator were 8 ± 1 mm in height (section parallel to the applicator axis) and 20 ± 1 mm in diameter (perpendicular section) with a serrated shape (Fig. 2). Lesions were similar in shape and slightly deeper (24 mm in diameter) with the 5 MHz applicator.

All control rats died or were killed with a large abdominal tumor and diffuse metastases less than 21 days after tumor implantation. Twelve rats were alive but in poor condition at 3 weeks and were killed. Among the 16 treated rats, 4 had a small tumor just beside the main tumor 10 ± 2 days after treatment. It was considered unlikely that this pattern of evolution could be related to treatment but was caused by tumor spread along the needle track during implantation. These 4 rats had tumor progression and metastasis. One treated rat did not recover from anesthesia and died immediately after treatment. Of the remaining 11 rats, 4 (36%) had local recurrences 5 to 15 days after treatment. These tumors progressed, but the pattern of progression was delayed and was slower than in untreated rats. Seven rats (64%) had neither local recurrence nor distant metastasis after a follow-up period of 1 month and were considered cured. The survival curve is presented in Figure 4.

In vivo experiments Lesions were obtained in all 15 exteriorized pig livers that were treated. Ten lesions were made with the applicator operating at 10 MHz, 17 W/cm2, and 390

GASTROINTESTINAL ENDOSCOPY

Experiments with animal tumors

VOLUME 50, NO. 3, 1999

High-intensity US probe for intraductal tumor destruction

F Prat, C Lafon, J Margonari, et al.

Figure 4. Survival curve of treated and control rats after subcutaneous implantation of AT2 Dunning tumors and treatment with high-intensity US. Day 0 is the day of treatment, 8 days after tumor implantation. Twelve control animals alive 3 weeks after tumor implantation were in poor condition and were killed.

DISCUSSION These results demonstrate that the US therapeutic device described in this study induces radical tissue destruction in the form of coagulation necrosis, regardless of the nature of the target tissue (normal or neoplastic). The ability of US to generate deep heating is well established. A tubular transducer is generally used for interstitial hyperthermia. However, in this study, a nondivergent but nonfocused transducer was used to destroy cylindrical or sectoral volumes. In this case, the decrease of pressure in the tissues is caused only by US attenuation. Treatment time also is reduced and heating is more easily controlled. Coagulation necrosis is typical of thermal injury and is irreversible.7 Several other modalities can induce thermal injury, such as radiofrequency, laser and argon ionized plasma, cryotherapy, or merely electrocoagulation.8 However, high-intensity US produces highly reproducible and predictable lesions, as shown in our experiments with pig livers both in vitro and in vivo. This precise tissue destruction is of particular interest when essential structures have to be preserved in the vicinity of the target tissue. The antitumor effect of high-intensity US has been studied and proved in various models, mostly with focused US.2,7,9-11 It remained to be confirmed that this effect could be obtained with nonfocused high-intensity US. Our rate of 64% tumor-free animals after treatment is a satisfactory result, and this was achieved in a tumor model that is particularly aggressive with a recurrence rate of almost VOLUME 50, NO. 3, 1999

100% after surgical resection. We can speculate that an even higher success rate can be achieved with a second treatment session a few days after the first with the purpose of complete destruction of residual tumor cells. Our group has demonstrated this property with high-intensity focused US in the treatment of patients with prostate cancer.10 With this method, the depth of tissue penetration can be predicted and depends on power intensity, shot duration, and working frequency of the transducer. A range of applicators with different frequencies could be developed. The desirable depth of tissue necrosis can be determined preoperatively with endosonography or perhaps in the future with magnetic resonance imaging. The risk for inducing complications as a result of injury to vital structures, such as blood vessels, is of particular concern. This risk is low with this device because it has been shown that blood vessels are natural thermal conductors and their walls may not be affected by US thermal effects.5,11 A prototype high-intensity US probe has been designed and built specifically for through the scope endoscopic therapy (Fig. 1). The applicator is easy to recognize and position under fluoroscopy. It can be guided over a wire, and treatment can be adapted to the length of a stricture. Successive levels of the tumor can be managed by means of moving the applicator upward or downward under fluoroscopic control. Rotation is ensured with a flexible spiral metal shaft that is remotely controlled with the same system as used in the animal experiments. This applicator has been designed for a retrograde GASTROINTESTINAL ENDOSCOPY

391

approach to the bile ducts through a sphincterotomy. Different versions can be developed for the percutaneous-transhepatic route and for other applications in the digestive tract. REFERENCES 1. Tompkins RK, Thomas D, Wile A, Longmire WP Jr. Prognostic factors in bile duct carcinoma. Ann Surg 1981;194:447-57. 2. Fry FJ, Johnson LK. Tumor irradiation with intense ultrasound. Ultrasound Med Biol 1978;4:337-41. 3. Lafon C, Chapelon JY, Prat F, Gorry F, Margonari J, Theilliere Y, et al. Design and preliminary results of an ultrasound applicator for interstitial thermal coagulation. Ultrasound Med Biol 1998;24:113-22. 4. Lafon C, Chapelon JY, Prat F, Gorry F, Theillère Y, Cathignol D. Design and in vitro results of a high intensity ultrasound interstitial applicator. Ultrasonics 1998;36:683-7. 5. Chen L, ter Haar G, Hill CR, Dworkin M, Carnochan P, Young H, et al. Effect of blood perfusion on the ablation of liver parenchyma with HIFU. Phys Med Biol 1993;38:1661-73. 6. Chapelon JY, Margonari J, Vernier F, Gorry F, Ecochard R,

392

GASTROINTESTINAL ENDOSCOPY

7. 8.

9.

10.

11.

Gelet A. In vivo effects of high intensity ultrasound on prostatic adenocarcinoma Dunning R3327. Cancer Res 1992;52: 6353-7. Ter Haar GR. Ultrasound focal beam surgery. Ultrasound Med Biol 1995;21:1089-100. Stauffer PR, Diederich CJ, Seegenschmiedt MH. Interstitial technologies. In: Seegenschmiet MH, Fessenden P, Vernon CC, editors. Thermo-radiotherapy and thermochemotherapy. Berlin: Springer-Verlag; 1995. p. 279-320. Prat F, Chapelon JY, Arefiev A, Cathignol D, Souchon R, Theilliere Y. High intensity focused ultrasound transducers suitable for endoscopy: feasibility study in rabbits. Gastrointest Endosc 1997;46:348-51. Gelet A, Chapelon JY, Bouvier R, Souchon R, Pangaud C, Abdelrahim F, et al. Treatment of prostate cancer with transrectal focused ultrasound: early clinical experience. Eur Urol 1996;29:174-83. Sibille A, Prat F, Chapelon JY, Abou el Fadil F, Henry L, Theilliere Y, et al. Characterization of extracorporeal ablation of normal and tumor-bearing liver tissue by high-intensity focused ultrasound. Ultrasound Med Biol 1993;19:803-13.

VOLUME 50, NO. 3, 1999