Ultrasonic epithelial ablation of the lower esophagus without stricture formation

Transcription

1 Surg Endosc (1998) 12: Springer-Verlag New York Inc Ultrasonic epithelial ablation of the lower esophagus without stricture formation A new technique for Barrett s ablation R. M. Bremner, R. J. Mason, C. G. Bremner, T. R. DeMeester, P. Chandrasoma, J. H. Peters, J. A. Hagen, M. Gadenstätter Department of Surgery, University of Southern California School of Medicine, 1510 San Pablo Street, Suite 514, Los Angeles, CA , USA Received: 3 April 1997/Accepted: 6 October 1997 Abstract Background: The premalignant potential of Barrett s esophagus has stimulated efforts to find a way to ablate the columnar epithelium in order to reheal the area with squamous epithelium, thus obviating the cancer risk. This study describes and evaluates a new technique using ultrasonic energy to ablate the epithelium of the lower esophagus in a porcine model. Methods: Eight young farm pigs were used to develop the technique of applying a laparoscopic Cavitron Ultrasonic Surgical Aspirator (CUSA) to the lower esophageal mucosa through an operating gastrostomy. A further 11 Yakutan minipigs then underwent CUSA epithelial ablation, followed by a laparoscopic Nissen fundoplication or postoperative acid suppression therapy. We then assessed the healing response in these subjects. Results: Optimal CUSA energy settings enabled complete ablation of the squamous epithelium with preservation of the muscularis mucosa and submucosa. The integrity of the aspirated cells was sufficient for cytological analysis. Healing occurred by squamous regeneration without stricture formation. Conclusions: The CUSA technique holds promise for complete ablation of the Barrett s epithelium in a single setting. The unique tissue-selective nature of the ablative process allows complete mucosal reepithelialization without stricture formation. Key words: Barrett s esophagus Epithelial ablation The risk of developing adenocarcinoma in a segment of Presented at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), San Diego, California, USA, March 1997 Correspondence to: T. R. DeMeester columnar-lined Barrett s esophagus is estimated at 0.8% per year. This means that of 100,000 patients with Barrett s, 800 will develop adenocarcinoma. This risk is sufficient to encourage annual endoscopic surveillance in patients without dysplasia. An alternative approach to surveillance is to ablate the Barrett s epithelium and encourage reepithelialization of the ablated area with squamous epithelium. Ablation has been performed with photodynamic therapy (PDT), laser therapy, and bipolar electrocoagulation [2 5, 7, 9]. Although all of these methods have had variable success, they also have demonstrated significant limitations that have hindered their wider application. This study evaluates a new technique that utilizes the tissue-selective nature of ultrasonic energy to ablate the epithelium of the lower esophagus in a porcine model. Methods CUSA ultrasonic generator The ultrasonic energy used in this experiment is generated by a Cavitron Ultrasonic Surgical Aspirator, or CUSA (CUSA Model 200, Valleylab Inc, Boulder, CO, USA). This generator is commonly used for liver resections and removal of brain tumors. The energy is delivered to the esophageal mucosa using an instrument that has been modified for laparoscopic access. The instrument is 30.1 cm long and consists of an energy tip measuring 2.54 mm in diameter, coupled with an irrigation and suction channel (Fig. 1). The tip vibrates at 23 khz. The operator s ability to adjust the amplitude varies the energy transferred to the tissue. A patent cavipulse setting on the instrument causes an oscillation of pulses at lower frequencies, which enhances the selective tissue disruption and fragmentation. Pilot and acute studies An initial pilot study was performed in eight young farm pigs (60 80 lb) to ascertain the optimal CUSA settings for ablation of esophageal squamous epithelium and to determine the optimal port site placement to access the esophageal lumen by the CUSA probe. By trial and error, it was

2 343 Fig. 1. Photograph of the laparoscopic CUSA instrument. Fig. 2. Diagram illustrating placement of the operating gastrostomy port. The laparoscopic CUSA is inserted through the port. An endoscope is used to direct the CUSA into the lumen of the esophagus. determined that placement of the port near the greater curvature of the stomach in the midline ensured easy access to the esophageal lumen and enabled the ultrasound tip to be applied to the full circumference of the luminal mucosa (Fig. 2). Optimal settings for epithelial ablation were determined by removing the esophagus from four animals, opening it longitudinally, and pinning the opened esophagus to a board. The CUSA was applied to the strips of esophageal mucosa using variable settings of amplitude and cavipulse. The extent and depth of ablation was determined by histologic examination. A cavipulse of 3 with % amplitude enabled ablation of the epithelium while leaving the underlying muscularis mucosa and muscularis propria intact (Fig. 3). Higher cavipulse settings resulted in submucosal hemorrhage and occasional disruption of the muscularis mucosa; lower amplitudes or cavipulse settings resulted in incomplete ablation of the epithelium. The chosen settings were then used to ablate the epithelium of the esophagus in the intact animals. After esophageal epithelial ablation was attempted in four animals, the esophagus of each animal was removed and examined histopathologically to establish the completeness of ablation. Operative studies The mucosal ablation technique was used on 11 Yakutan minipigs, who were allowed to recover. Five animals underwent CUSA ablation of the distal 5 cm of the esophagus and were given Omeprazole 20 mg per day for 7 days to prevent acid injury to the raw esophageal surface. Six animals underwent CUSA ablation followed by laparoscopic Nissen fundoplication under the same anesthetic. The Nissen fundoplication was performed by enveloping the lower esophagus, which was intubated with a 30-Fr bougie with a tongue of gastric fundus, and then suturing it in place with two Ethibond sutures. It was not necessary to take down the short gastric vessels because the pig has a large, floppy fundus. The diaphragmatic crura were sutured to close the dissection defect at the hiatus. The animals were endoscoped at 2 4 weeks to assess healing of the mucosa. They were killed at 6 weeks using 6 cc of Euthasol (390 mg/ml pentobarbital) for euthanasia. The esophagus in each animal was resected and intubated with a Penrose drain containing barium sulfate for radiographic examination to identify any luminal narrowing or stricture formation. All specimens underwent histological examination. Fig. 3. Histologic view of the interface between the normal epithelium and the area of ablation ( 10 magnification). Note the absence of any epithelial cells in the ablated area. The muscularis mucosa is completely intact, and there is no evidence of injury to the submucosa or muscularis propria. The muscularis propia is not visible in the magnification; it is deep to the area shown. Anesthesia was maintained with 0.5 2% isoflurane. The abdominal wall was cleaned with betadine solution and draped in a standard sterile fashion. An Olympus endoscope was passed through the esophagus into the stomach and directed anteriorly. The room was darkened, and the light from the tip of the endoscope was visualized through the anterior abdominal wall. A 16-gauge needle was passed percutaneously into the stomach using the light source as a target. The entrance of the needle into the stomach was monitored visually through the endoscope. An operating gastrostomy tube (Cook Surgical, Cook Inc., Bloomington, IN, USA) was placed into the stomach using a standard percutaneous endoscopic gastrostomy technique (Fig. 2). The balloon of the port was insufflated with air to secure the anterior gastric wall to the anterior abdominal wall. Correct placement of the operating gastrostomy port was critical to enable the passage of a rigid laparoscopic CUSA instrument through the gastroesophageal junction and into the lumen of the lower esophagus. If the placement was incorrect, the CUSA instrument could not be applied to the complete circumference of the esophageal lumen. An endoscope was placed in the upper esophagus to visualize the insertion of the laparoscopic CUSA instrument into the esophagus. The tip of the CUSA probe was guided in a sweeping motion over the mucosal lining, beginning cephalad and moving progressively caudad. The ablated epithelial cells were collected from the central aspiration port and preserved for cytology. The epithelium was ablated over a 5-cm circumferential segment of lower esophagus just proximal to the gastroesophageal junction. Epithelial ablation could be easily confirmed endoscopically by loss of the pearly-white appearance of the epithelium, thus revealing the underlying dark red muscularis mucosa. There was no visible blood loss associated with the ablation. Bleeding during the ablation usually indicated an injury to the muscularis mucosa and disruption of submucosal capillaries. It could be avoided by using the correct CUSA power settings and avoiding repeated contact of the ultrasonic tip with previously ablated areas. Care was taken to avoid excessive luminal insufflation, since it tended to cause distention of the small bowel. At the conclusion of the ablation process, the operating gastrostomy port was removed and the gastrotomy closed with a running silk suture through a slightly enlarged skin incision at the gastrostomy site. The fascia was closed with interrupted Prolene sutures, and the skin was closed with staples. The animals were recovered and given buphrenex for postoperative pain management. Oral feeding was started on the 1st postoperative day. Operative technique Prior to the procedure, the animals were fasted for 48 h to ensure an empty stomach, since Yakutan minipigs are known to have a prolonged gastric emptying time. Animals were anesthetized with 75 mg Telazol (tiletamine with zelozepam 100 mg/ml), 35 mg xylozine, and 0.2 mg glycopyrolate. Ethics This study was approved by the animal Use Committee of the University of Southern California (protocol number 8976).

3 344 Fig. 4. The microscopic specimen of the cell block recovered from the aspiration port during the ablation ( 100 magnification). The cells have maintained their integrity for cytologic examination. Fig. 5. Endoscopic pinch biopsy specimen obtained at 2 weeks after ablation showing the early regeneration of squamous epithelium with three to six cell layers ( 40 magnification). Results Acute studies Complete ablation of the epithelium of the lower esophagus was possible through the operating gastrostomy port in all animals. In all specimens, only the epithelium was ablated, leaving a thin coagulum overlying the muscularis mucosa. Macroscopically and histologically, no islands of epithelium were left in the area of ablation. The muscularis mucosa was intact in all animals, and no injury was observed either in the submucosa or muscularis propria. The integrity of the aspirated cells was excellent and allowed for cytological evaluation (Fig. 4). Operative studies Two of the 11 animals died during the 6-week observation period. One animal who underwent laparoscopic Nissen fundoplication died suddenly during the night on the 7th postoperative day from a cecal volvulus with bowel infarction. Pathology of the esophageal specimen revealed granulation tissue superficial to the muscularis mucosa; there was no evidence of epithelial islands, confirming the effectiveness of the ablation technique. A second animal was noted at endoscopy 4 weeks after surgery to have herniated the fundoplication into the chest. The injury was due to failure to close the diaphragmatic crura at the time of fundoplication. We found it prudent to kill the animal at this time even though she had been eating well and had not lost weight. On inspection, the esophageal mucosa was completely reepithelialized and free from stricture. The remaining animals were endoscoped at 2-week intervals. At 2 weeks, endoscopy showed areas of translucent white epithelium covering the granulation tissue in the ablated area. Histology of pinch biopsies showed epithelialization with three to six layers of squamous cells (Fig. 5). At 4 weeks, epithelium covered the whole of the ablated areas, but it was slightly pink in color. By 6 weeks, complete healing had occurred without stricture formation, and the ablated area could not be identified. Radiographic examination of the barium-filled Penrose drain passed through the intact specimen showed no evidence of stricturing (Figs. 6, 7). On histological examination, all specimens were free of fibrosis in the submucosa or the muscularis propria, and it was difficult to distinguish normal from regenerated epithelium. Discussion Currently, there is no effective means of reversing the metaplastic process associated with Barrett s esophagus. High doses of proton pump inhibitors or antireflux surgery have occasionally resulted in total or partial regression of the columnar epithelium, but the occurrence is rare and unpredictable. Further, acid suppression therapy has not been shown to prevent malignant progression of the columnar lining, and there is insufficient evidence at present that antireflux surgery is protective [8, 12]. Consequently, the International Society of Diseases of the Esophagus (ISDE) recommends that patients with Barrett s metaplasia be endoscoped and biopsied at yearly intervals in order to detect the emergence of dysplasia, a sign of movement toward invasive cancer [10]. The cost of surveillance over a lifetime is significant. Added to this is the cost of esophagectomy if high-grade dysplasia or cancer develops. Consequently, alternative approaches to the problem are being investigated. Photodynamic therapy (PDT), laser ablation, and multipolar electrocoagulation have all been used in humans to remove columnar epithelium with variable success [1 5, 7, 9]. The results of published studies have shown that squamous reepithelialization of ablated areas occurs with squamous cells provided gastric acid secretion is suppressed. One limitation of these techniques is that the ablation of Barrett s epithelium is incomplete. There are reports of squamous overgrowth of partially ablated Barrett s epithelium, as well as development of adenocarcinoma beneath squamous reepithelialized areas [6]. Another drawback of these techniques is the inability to control the depth of tissue injury, with resultant stricture formation in 50% of patients. The great advantage of ultrasonic energy is the tissue-

4 345 Fig. 6. Necropsy specimens of resected esophagi after complete mucosal healing. A Photograph of the esophagi intubated with barium-filled Penrose drains of group 1 animals. B A radiograph of these esophagi demonstrates the absence of any luminal narrowing or stricture formation. C Four of the five esophagi from the animals that had a Nissen fundoplication again demonstrating the absence of stricture formation. The fifth animal was killed at a different time but showed similar findings. selective nature of the ablative process. Tissue that has a high elastin or collagen content or that consists of muscle cells is resistant to damage by the ultrasound energy. This explains why the CUSA instrument can selectively ablate hepatic parenchymal cells while preserving the vascular and ductal tissues during liver resections. The aim of the present study was to develop a technique whereby ultrasonic energy could be used to ablate the esophageal epithelium without injury to the underlying tissue layers. The technique that we developed utilizes recent advances in both endoscopy and laparoscopy. Selective ablation of the distal esophageal epithelium through the operating gastrostomy port was easily accomplished once optimal CUSA settings were defined and the proper positioning of the gastrostomy site was determined. The impressive histology of the acute experiments showed that complete ablation of all squamous epithelium was possible without injury to deeper layers of the esophageal wall. This explains why healing occurred in the absence of stricture formation. It is assumed that the columnar epithelium that occurs in Barrett s metaplasia will act similarly, but this still needs to be demonstrated. A valuable feature of this technique is the ability to collect the ablated epithelium by aspiration so that cytologic examination can be performed. All techniques utilizing chemical or thermal means to destroy the epithelial cells make it impossible to detect occult carcinoma in the ablated epithelium. In effect the aspiration specimen of the ultrasonic ablative technique represents the ultimate biopsy. Although there was no objective difference in the healing of the animals with a Nissen fundoplication, this is not to be interpreted that a Nissen fundoplication should not be performed following ablation. The animals in this study had normal lower esophageal sphincters and therefore minimal esophageal exposure to gastric contents. This is not the situation in patients with Barrett s esophagus where the presence of a defective sphincter is the rule (12). The structurally defective lower esophageal sphincter in patients with Barrett s esophagus makes the medical management of reflux tenuous at best. Even if the epithelium was successfully ablated in these patients, lifelong acid suppression therapy would still be required, and the potential for re-injury and metaplasia would always remain. Consequently, without the addition of a Nissen fundoplication, the need for strict sur-

5 346 impact of therapy on extent of Barrett s esophagus in 67 patients. Dig Dis Sci 35: Sampliner RE, Hixson LJ, Fennerty MB, Garewal HS (1993) Regression of Barrett s esophagus by laser ablation in an anacid environment. Dig Dis Sci 38: Skinner DB, Siewert JR (1996) Results of the consensus on esophageal cancer held at the VIth World Congress of the International Society for Diseases of the Esophagus Dis Esophagus 9: Stein HJ, Hoeft S, DeMeester TR (1993) Functional foregut abnormalities in Barrett s esophagus. J Thorac Cardiovasc Surg 105: Williamson WA, Ellis FH Jr, Gibb SP, Shahian DM, Aretz HT (1990) Effect of antireflux operation on Barrett s mucosa [see Comments]. Ann Thorac Surg 49: Fig. 7. An opened necropsy specimen of a healed esophagus 6 weeks after mucosal ablation shows complete healing with no evidence of stricture formation. veillance would continue. It is felt that the technique presented in this study holds promise for the permanent eradication of Barrett s columnar epithelium and with it the risk of malignancy. Acknowledgments. This work was supported in part by grants from SAGES and Valleylab Inc. (Boulder, CO, USA). We would like to acknowledge Harry Valenta of Valleylab for his technical assistance and enthusiasm with the project, and, Linda and Paul Kirkman of Animal Unit at USC for their care and help with the animals both intraoperatively and postoperatively. References 1. Barr H, Shepherd NA, Dix A, Roberts DJ, Tan WC, Krasner N (1996) Eradication of high-grade dysplasia in columnar-lined (Barrett s) oesophagus by photodynamic therapy with endogenously generated protoporphyrin IX [see Comments]. Lancet 348: Berenson MM, Johnson TD, Markowitz NR, Buchi KN, Samowitz WS (1993) Restoration of squamous mucosa after ablation of Barrett s esophageal epithelium [see Comments]. Gastroenterology 104: Brandt LJ, Blansky RL, Kauvar DR (1995) Repeat laser therapy of recurrent Barrett s epithelium: success with anacidity [Letter]. Gastrointest Endosc 41: Overholt BF, Panjehpour M (1995) Barrett s esophagus: photodynamic therapy for ablation of dysplasia, reduction of specialized mucosa, and treatment of superficial esophageal cancer [see Comments]. Gastrointest Endosc 42: Overholt BF, Panjehpour M (1996) Photodynamic therapy for Barrett s esophagus: clinical update. Am J Gastroenterol 91: Sampliner RE, Fass R (1993) Partial regression of Barrett s esophagus: an inadequate endpoint. Am J Gastroenterol 88: Sampliner RE, Fennerty B, Garewal HS (1996) Reversal of Barrett s esophagus with acid suppression and multipolar electrocoagulation: preliminary results. Gastrointest Endosc 44: Sampliner RE, Garewal HS, Fennerty MB, Aickin M (1990) Lack of Discussion Dr. Hunter: Dr. Bremner, tell me a little bit about how you anticipate the utilization of this, and where you see your next series of experiments going: Is this ready for human application, and if so, what patients with Barrett s are those in whom you would consider utilizing this technique? Dr. Bremner: To answer your first question first, Dr. Hunter, we anticipate using this now in Barrett s tissue and cadaveric resected esophageal specimens to insure that our assumption that this will work on Barrett s epithelium is the same as it works on squamous mucosa. It certainly, in the porcine stomach, seems to work very well; in fact, it seems to ablate the columnar cells of the stomach even easier than the squamous cells. We anticipate that assumption to hold true. Then we anticipate taking it to human studies. Obviously we would like to try patients with smaller segments of Barrett s initially. We anticipate that this would be a procedure that would be performed at the time of laparoscopic Nissen fundoplication, so under the same anesthetic ablating the mucosa and then performing the Nissen. That was one of the reasons for doing the Nissen fundoplication at the time of ablation in the porcine model here. We had wondered whether or not the raw surfaces of the epithelium, after ablation, won t coapt after the addition of a Nissen fundoplication in the lower esophagus. It was our impression, although we have no objective evidence, that the animals that underwent Nissen fundoplication healed more quickly than those without the addition of an antireflux procedure, and possibly the reason for that is that pigs normally have some physiologic reflux. We had acid suppression in these animals only for a week; possibly some of the physiologic reflux continued thereafter. We think it would need to be done at the same time as an antireflux procedure. Dr. Greene: I just had a question about the method of approach. Realizing that the ultrasonic application has to be done through a rigid system, would you anticipate that this could be done through a rigid endoscopic system from an oral approach, and would that benefit us, because we recognize that all Barrett s is not around the GE junction. In fact, we see Barrett s changes well up into the esophagus. I wonder what your thoughts would be rather than transgastric, what would be the opportunity of an oral or even a rigid approach? Dr. Bremner: Thank you very much, Dr. Greene. It s a good question. We ve been able to ablate the epithelium a whole 13 cm up from the gastroesophageal junction, but that s not

6 347 going to reach the entire esophagus in human Barrett s esophagus. The limitation is the length of the instrument. At the moment it is 30.1 cm long, and so from the orogastric approach you re only going to get the upper esophagus. But certainly from above and below you d be able to get the entire esophagus, if you needed to. Through a rigid esophagoscope, I don t anticipate that with the present technology we ll be able to reach the lower esophagus through the mouth, but that s just a technical limitation at the moment. Dr. Mulvihill: I enjoyed this very much. I wondered if you had any information on how rapidly there was restitution of this mucosa? Dr. Bremner: We do, and it s in the manuscript. We endoscoped these animals during the healing process. One of the animals had an early death from a cecal volvulus, and that was at postop day number 7. That enabled us to look at the esophageal specimen in its entirety at a week. This confirmed that there was no epithelium left, and there was just a thin layer of granulation tissue over the muscularis mucosa. That is at one week. At two weeks we had some endoscopy specimens that showed that there are approximately three to six layers of squamous epithelium; just a thin layer of squamous epithelium already at two weeks over the entire surface. At four weeks it looked almost healed, but endoscopically there was slack pink change between the normal epithelium and the regenerated. At six weeks we could hardly see any difference endoscopically, and our pathologist said it was very difficult for him to see any difference between the normal squamous and the regenerated areas.