Esophagectomy remains the standard of care for localized

Transcription

1 ORIGINAL ARTICLES: GENERAL THORACIC Minimally Invasive Esophagectomy James D. Luketich, MD, Philip R. Schauer, MD, Neil A. Christie, MD, Tracey L. Weigel, MD, Siva Raja, BS, Hiran C. Fernando, MD, Robert J. Keenan, MD, and Ninh T. Nguyen, MD Section of Thoracic Surgery and the Minimally Invasive Surgery Center, University of Pittsburgh Medical Center Health System, Pittsburgh, Pennsylvania Background. Open esophagectomy can be associated with significant morbidity and delay return to routine activities. Minimally invasive surgery may lower the morbidity of esophagectomy but only a few small series have been published. Methods. From August 1996 to September 1999, 77 patients underwent minimally invasive esophagectomy. Initially, esophagectomy was approached totally laparoscopically or with mini-thoracotomy; thoracoscopy subsequently replaced thoracotomy. Results. Indications included esophageal carcinoma (n 54), Barrett s high-grade dysplasia or carcinoma in situ (n 17), and benign miscellaneous (n 6). There were 50 men and 27 women with an average age of 66 years (range 30 to 94 years). Median operative time was 7.5 hours (4.5 hours with > 20 case experience). Median intensive care unit stay was 1 day (range 0 to 60 days); median length of stay was 7 days (range 4 to 73 days) with no operative or hospital mortalities. There were four nonemergent conversions to open esophagectomy; major and minor complication rates were 27% and 55%, respectively. Conclusions. Minimally invasive esophagectomy is technically feasible and safe in our center, which has extensive minimally invasive and open esophageal experience. Open surgery should remain the standard until future studies conclusively demonstrate advantages of minimally invasive approaches. (Ann Thorac Surg 2000;70:906 12) 2000 by The Society of Thoracic Surgeons Esophagectomy remains the standard of care for localized esophageal cancer. The surgical options include thoracotomy, laparotomy, or both, and can be associated with significant morbidity and mortality and a delay in return to preoperative activity level [1 3]. Advances in minimally invasive instrumentation, increasing applications of less invasive techniques to a variety of disorders, and the potential for less morbidity are encouraging surgeons to explore the role of less invasive techniques for esophagectomy. Numerous reports have been published on the application of videothoracoscopy to esophagectomy without apparent advantage [4, 5]. However, this was not surprising because most of these cases included an access thoracotomy or laparotomy, which was similar in size to standard incisions used by most surgeons. Total laparoscopic esophagectomy has been reported by our group and two others [6 8]. These initial reports demonstrated that minimally invasive esophagectomy was technically feasible and safe in centers with extensive laparoscopic esophageal experience and that there may be advantages over conventional surgery. The goal of this report is to summarize our growing experience with a combined thoracoscopic and laparoscopic approach to esophagectomy in our first 77 cases. Presented at the Thirty-sixth Annual Meeting of The Society of Thoracic Surgeons, Ft Lauderdale, FL, Jan 31 Feb 2, Address reprint requests to Dr Luketich, Section of Thoracic Surgery, University of Pittsburgh Medical Center, 200 Lothrop St, Pittsburgh, PA 15213; luketichjd@msx.upmc.edu. Material and Methods Patient Population Minimally invasive esophagectomy was attempted in 77 patients over a 3-year period (August 1996 to September 1999). This represents approximately 60% of patients undergoing esophagectomy during this period. Initial patient selection included only small T1 tumors or highgrade dysplasia. As experience was gained, T2 and T3 tumors were included. Due to the retrospective nature of this review, more specific selection criteria were not applied. Most patients had undergone minimally invasive surgical staging or endoscopic ultrasound preoperatively, which we have previously described [9]. The first 5 patients were approached totally laparoscopically; thoracoscopy was subsequently added to facilitate thoracic esophageal mobilization and lymph node dissection [10]. Surgical Technique The patient is intubated with a double-lumen tube for single lung ventilation and positioned in the left lateral decubitus position. Four thoracoscopic ports are introduced (Fig 1). The camera port (10 mm) is placed at the seventh intercostal space, mid-axillary line. Two 5-mm ports are placed, one at the eighth or ninth intercostal space, 2 cm posterior to the posterior axillary line for the ultrasonic coagulating sheers (U.S. Surgical, Norwalk, CT), and one posterior to the tip of the scapula. One 10-mm port is placed at the fourth intercostal space anteriorly for retraction of the lung and esophageal countertraction during dissection. Next, a single retract by The Society of Thoracic Surgeons /00/$20.00 Published by Elsevier Science Inc PII S (00)

2 Ann Thorac Surg LUKETICH ET AL 2000;70: MINIMALLY INVASIVE ESOPHAGECTOMY 907 Fig 1. Video-assisted thoracoscopic surgical port sites. Fig 3. Abdominal port sites for laparoscopy. ing suture is placed near the central tendon of the diaphragm (0-surgitek; U.S. Surgical). This traction suture allows downward retraction on the diaphragm giving excellent exposure of the distal esophagus, eliminating the need for a retractor. The mediastinal pleura overlying the esophagus is divided and the entire thoracic esophagus is exposed. A penrose drain is placed around the esophagus to facilitate traction and exposure (Fig 2). The azygos vein is divided using the end-gia vascular stapler (U.S. Surgical). Circumferential mobilization of the esophagus, with all surrounding lymph nodes, periesophageal tissue and fat, and the mediastinal pleura is performed from the diaphragmatic reflection to the thoracic inlet. The entire subcarinal node packet is removed with the specimen; lymph nodes from the upper esophagus and recurrent nerve area are only sampled. The vagus nerves are identified and divided above the level of the carina, the right recurrent is gently dissected away from the upper esophagus, and the left is frequently not visualized. A single 28F chest tube is inserted through the camera port, the other port sites are closed, and the patient is turned to the supine position. In the supine position, five abdominal ports are placed on the anterior abdominal wall similar to our approach for laparoscopic Nissen fundoplication (Fig 3). The left lobe of the liver is retracted upward to expose the esophageal hiatus using a diamond flex retractor (Genzyme, Tucker, GA) and held in place with a self-retaining system (Mediflex; Velmed Inc, Wexford, PA). The operation begins with dissection of the gastro-hepatic ligament and exposure of the right crus of the diaphragm. Short gastric vessels are divided using the ultrasonic Fig 2. Thoracoscopic mobilization of the esophagus. Fig 4. Laparoscopic pyloroplasty.

3 908 LUKETICH ET AL Ann Thorac Surg MINIMALLY INVASIVE ESOPHAGECTOMY 2000;70: Fig 7. Thoracoscopic and laparoscopic port sites at the 3-week postoperative visit. Fig 5. Laparoscopic gastric tubularization. coagulating shears (U.S. Surgical). The dissection continues along the greater curvature of the stomach, preserving the right gastroepiploic arcade. The stomach is retracted superiorly, allowing lymph node dissection of the celiac and gastric vessels. Once the gastric artery and vein are exposed, they are divided using an endoscopic vascular stapler (Endo-GIA II; U.S. Surgical). Pyloroplasty is then performed using ultrasonic shears and closed transversely (Fig 4) using an Endo-stitch (U.S. Surgical). The lesser curve fat and nodes are dissected en bloc with the stomach. A limited mobilization of the duodenum and gastric antrum is performed. The gastric tube is then constructed by dividing the stomach starting at the distal lesser curve, preserving the right gastric vessels using the 4.8-mm stapler (Endo-GIA II; U.S. Surgical) (Fig 5). The gastric tube is attached to the esophageal and gastric specimen using two sutures (Endostitch; U.S. Surgical). An additional suture is placed on the anterior proximal gastric tube to facilitate correct orientation as the tube is brought up to the neck. The ligament of Treitz is identified and a laparoscopic jejunostomy tube is placed by first attaching a limb of proximal jejunum to the anterior abdominal wall using the endo-stitch. A needle catheter kit (Compat Biosystems, Minneapolis, MN) is inserted percutaneously into the peritoneal cavity, and under direct laparoscopic vision it is directed into the proximal loop of jejunum. The guide wire and catheter are threaded into the jejunum. The jejunal entry site is sutured to the anterior abdominal wall for a distance of several centimeters. The last step in the abdominal operation is the final dissection of the phrenoesophageal ligament, which opens the plane into the thoracic cavity. This step is completed last to minimize loss of pneumo-peritoneum into the mediastinum. We also partially divide the right and left crura to widen the hiatus to prevent gastric tube outlet obstruction. A 4- to 6-cm horizontal neck incision is made just above the suprasternal notch and the cervical esophagus is mobilized and exposed. Finger dissection is continued distally until the thoracic dissection plane is encountered. The cervical esophagus is divided and the esophagogastric specimen is pulled up out of the wound. As traction is applied to the specimen from the neck, the esophagus and attached gastric tube are laparoscopically guided in proper alignment into the mediastinum. The specimen is removed from the field through the neck. An anastomosis is performed between the esophagus and the gastric tube using standard techniques, which completes the operation (Fig 6). The gastric tube is sutured circumferentially to the hiatus to prevent subsequent thoracic herniation. Fig 6. Completed laparoscopic thoracoscopic operation. Results Minimally invasive esophagectomy was attempted in all 77 patients (57 males and 20 females). The average age

4 Ann Thorac Surg LUKETICH ET AL 2000;70: MINIMALLY INVASIVE ESOPHAGECTOMY 909 Table 1. Complications Minor Complications No. of Cases Major Complications No. of Cases Transfusion 5 Hypopharynx laceration 1 Urinary retention 1 Small bowel perforation 1 Air leak 3 Tracheal tear 1 Atelectasis 5 Recurrent laryngeal nerve injury 2 Aspiration requiring NG tube 1 Chylothorax 3 Pleural effusion 18 Anastomotic leak 7 Atrial fibrillation 8 Abscess (1 mediastinal, 1 abdominal) 2 Dislodged j-tube 1 Pyloric leak 1 j-tube infection 2 Delayed gastric empyting b 3 Delayed gastric empyting a 1 Deep vein thrombosis 1 Stricture requiring dilation 1 Pulmonary embolus 1 C-diff colitis 1 Myocardial infarction 2 Wound infection 1 Adult respiratory distress system 1 Pneumonia 4 Acute respiratory failure 2 a Resolved after pyloric dilation. b required laparoscopic pyloroplasty. NG nasogastric. was 65 years (range 30 to 89 years). The indications for esophagectomy included carcinoma (n 52), Barrett s high-grade dysplasia (n 19), and benign esophageal disorders (n 6). Twenty-two patients received preoperative chemotherapy. The benign indications for esophagectomy included two cases of end-stage achalasia, two recalcitrant strictures, and two tracheoesophageal (TE) fistulae. An elective operation was performed in 73 patients. Four patients required urgent surgery and included the two TE fistulas and two strictures with localized perforations. Fifty patients (65%) had undergone prior surgery involving the abdominal or thoracic cavity. The surgical approach included laparoscopic mobilization and thoracoscopic esophagectomy with cervical anastomosis (n 60) (Fig 7 demonstrates typical postop appearance of port access sites), laparoscopic mobilization and right mini-thoracotomy with thoracic anastomosis (n 8), and laparoscopic transhiatal esophagectomy with cervical anastomosis (n 9). In this series, 4 patients required conversion to an open procedure due to extensive adhesions in the operative field. There were no emergency conversions. Additional procedures included pyloromyotomy (n 27), pyloroplasty (n 46), and laparoscopic j-tube placement (n 56). The median operative time was 7.5 hours (range 4.0 to 13.6 hours). The size of the tumor did not significantly affect operative times but bulky, possible T4 tumors were approached by open surgery. The median intensive care unit stay was 1.0 days (range 0 to 60 days) with a median hospital stay of 7 days (4 to 73 days). The 30-day perioperative mortality was zero. Major complications occurred in 21 (27%) patients (Table 1). Six of the seven anastomotic leaks in the neck were managed by conservative measures. One developed a descending periesophageal abscess that resolved with drainage. The single hypopharnygeal perforation was secondary to a malpositioned esophageal temperature probe and was repaired primarily at the time of esophagectomy with no further sequelae. One patient suffered a tracheal tear on postoperative day 6 during reintubation for respiratory distress. The reintubation injury was due to malposition of the stylet of the endotracheal tube. This was repaired by right thoracotomy and was followed by a prolonged hospital stay of 50 days, but ultimately the patient recovered. Two patients had permanent recurrent laryngeal injury requiring vocal cord injections with Teflon. Minor complications in the perioperative period occurred in 55% (Table 1). Delayed complications (greater than 30 days postoperative) occurred in 4 patients. Two patients who had an initial pyloromyotomy developed persistent delayed gastric emptying requiring laparoscopic pyloroplasty with good results. One patient developed delayed gastric emptying with hold-up of contrast at the crural level. This required laparoscopic crural division with good results. One patient presented with a moderately symptomatic diaphragmatic herniation of bowel into the right chest and underwent elective laparoscopic reduction and repair with good results. The final pathology of the cancer patients included stage 0 (4), I (10), IIA (16), IIB (10), III (30), and IV (1). Histologic subtypes included adenocarcinoma of the gastroesophageal junction, which was present in 90% of cases, and 10% had midesophageal squamous cell carcinoma. No cases of cervical esophageal cancer were included. The surgical margins were negative on frozen section in all cases. The final pathology revealed microscopic disease at the adventitial margin in 3 patients. The average number of lymph nodes removed with the specimen was 16 (range 10 to 51). Approximately 60% of the dissected nodes were from the laparoscopic dissection and 40% were from the chest. The 6 patients with benign indications were all alive at 20-month follow-up. The cancer group (71 patients) had an overall survival of 81% at a median follow-up of 20 months. There have been 18 patients with cancer recurrence. In 8, only distant disease

5 910 LUKETICH ET AL Ann Thorac Surg MINIMALLY INVASIVE ESOPHAGECTOMY 2000;70: was present; in 8, there was local and distant recurrence. In the 2 patients with local recurrence only, 1 had extensive Barretts preop and ultimately was reresected for an apparent new primary within residual Barrett s extending into the cervical esophagus. The other patient had extensive recurrence of invasive carcinoma along the gastric tube and has been palliated with photodynamic therapy. Survival by stage is illustrated in Figure 8. Comment Fig 8. Kaplan-Meier survival plots for 71 patients with cancer resections: HGD, carcinoma in situ, and stage I (n 14); stage IIA and IIB (n 26); stage III (n 30). The application of minimally invasive surgery to treat various esophageal disorders is increasing. Laparoscopic antireflux procedures have all but eliminated conventional open approaches. Now surgeons in a number of medical centers are beginning to report cases of minimally invasive approaches to more complex diseases including esophagectomy. This series of 77 patients, the largest series of minimally invasive esophagectomy reported to date, demonstrates that in experienced hands it is a technically feasible and safe operation and may be associated with a shorter hospital stay and earlier return to premorbid status. Although a randomized trial, our median hospital stay of 7 days compares favorably with that of open procedures reported in the literature, which almost universally exceed 10 to 12 days [1 3]. The operative mortality of 0% is significantly less than many open series [1 3, 11, 12]. A review of the literature has not demonstrated a consistent benefit of minimally invasive esophagectomy in previous case reports or small series [4]. However, most of these series were limited by small numbers and the majority included an access laparotomy or thoracotomy that was similar in size to standard open operations [13, 14]. A standardized approach to minimally invasive esophagectomy has not yet been accepted. In addition to our work, two other series of minimally invasive esophagectomy without an access incision or hand-assistance have been reported and suggest there may be advantages to the less invasive procedure [3, 4]. Swanstrom and Hansen [6] reported their experience with 9 patients undergoing a totally laparoscopic esophagectomy with no operative mortality and a mean hospital stay of 6.4 days. This series now includes 22 patients (personal communication). One major complication occurred in a patient with poor viability of the gastric tube requiring revision. In a report from Brazil, DePaula and associates [7] reported a favorable experience with 48 patients; although many of these patients had achalasia from Chagas disease, the hospital stay was less than 7 days with an acceptable rate of complications. Because the majority of patients undergoing esophagectomy have cancer, the adequacy of surgical margins and lymph node dissection must be considered. Two recent reports have evaluated this parameter and found that the number of nodes resected through the minimally invasive approach is comparable with that of open procedures and the ability to obtain a negative surgical margin was not compromised [13, 14]. The series by Law and associates [14] demonstrated it was possible to dissect up to 51 lymph nodes, but noted that early in their groups experience this was time consuming. We were able to obtain an average of 16 lymph nodes per case (range 10 to 51) and noted a similar learning curve for operative times. In our experience, the superior visualization of combined laparoscopy and thoracoscopy allows the surgeon to perform a very complete lymph node sampling or complete dissection. Negative surgical margins by frozen section were attainable in all cases in our experience. In three cases, the final pathology revealed positive adventitial margins, which is not different from our open experience. The stage-specific survival at 20 months compares favorably with our open experience and that of other reports [1 3, 11, 12] (Fig 7). Initially, we performed esophagectomy totally laparoscopically by the transhiatal approach or with the addition of a mini-thoracotomy. The limitations in lymph dissection and mobilization of the thoracic esophagus by laparoscopy alone led to our addition of thoracoscopy early in our experience. Akaishi and associates [13] reported performing thoracoscopic en bloc total esophagectomy with radical mediastinal lymphadenectomy on 39 patients. The mean operative time of 200 minutes and the completeness of lymph node dissection were comparable with their open experience. They attributed their success rate to extensive experience in the animal laboratory for 18 months before their first clinical case. Our initial experience suggests that minimally invasive esophagectomy is technically feasible and safe in our center, which has extensive experience with both minimally invasive and open esophageal surgery. In this preliminary report, minimally invasive esophagectomy was associated with a shorter hospital stay and a more rapid return to a normal routine than previous reviews of open esophagectomy [1 3, 11, 12]. Controlled trials will be necessary to confirm any advantages over open techniques and will require standardization of techniques, advanced minimally invasive training, and intense institutional experience.

6 Ann Thorac Surg LUKETICH ET AL 2000;70: MINIMALLY INVASIVE ESOPHAGECTOMY 911 References 1. Orringer MB, Marshall B, Stirling MC. Transhiatal esophagectomy for benign and malignant disease. J Thorac Cardiovasc Surg 1996;105: Orringer MB, Marshall B, Iannettoni MD. Transhiatal esophagectomy: clinical experience and refinements. Ann Surg 1999;230: Rindani R, Martin CJ, Cox MR. Transhiatal versus Ivor- Lewis esophagectomy: is there a difference? Aust N Z J Surg 1999;69: Luketich JD, Schauer PR, Urso K, et al. Future direction in esophageal cancer. Chest 1998;113(Suppl): Nguyen NT and Luketich JD. Minimal access approach to esophagectomy. In: Yim A, ed. Minimal access cardiothoracic surgery. Orlando: WB Saunders, 1999: Swanstrom L, Hansen P. Laparoscopic total esophagectomy. Arch Surg 1997;132: DePaula AI, Hasiba K, Ferreira EAB, et al. Transhiatal approach for esophagectomy. In: Toouli J, Gossot D, Hunter JG, eds. Endosurgery. New York: Churchill Livingstone; 1996: Luketich JD, Nguyen NT, Weigel TL, Ferson P, Keenan R, Schauer P. Minimally invasive approach to esophagectomy. J Soc Laparoendosc Surg 1998;2: Luketich JD, Schauer PR, Landreneau RJ, et al. Minimally invasive surgical staging in detecting lymph node metastases in esophageal cancer. J Thorac Cardiovasc Surg 1997;114: Nguyen NT, Schauer PR and Luketich JD. Combined laparoscopic and thoracoscopic approach to esophagectomy. J Am Coll Surg 1999;188: Anikin VA, McManus KG, Graham AN, McGuigan JA. Total thoracic esophagectomy for esophageal cancer. J Am Coll Surg 1997;185: Millikan KW, Silverstein J, Hart V, et al. A 15-year review of esophagectomy for carcinoma of the esophagus and cardia. Arch Surg 1995;130: Akaishi T, Kaneda I, Higuchi N, et al. Thoracoscopic en bloc total esophagectomy with radical mediastinal lymphadenectomy. J Thorac Cardiovasc Surg 1996;112: Law S, Manson F, Chu KM, Wong J. Thoracoscopic esophagectomy for esophageal cancer. Surgery 1997;122:8 14. DISCUSSION DR MARK B. ORRINGER (Ann Arbor, MI): Doctor Luketich, this was a well-delivered paper, and I enjoyed it very much. Your technique is an extension of the concept of esophagectomy without thoracotomy, which I think has now been shown to be achievable in the majority of people needing an esophagectomy. I do have some concerns that I would like you to address. This is not just minimally invasive esophagectomy. It is minimally invasive gastric mobilization as well, and the gastric mobilization is a key part of this operation. We have learned over the years, as we stopped focusing upon the actual transhiatal esophagectomy, that mobilizing the stomach in as atraumatic a fashion as possible is extremely important, and when the stomach is delivered out onto the abdominal wall, as the stapler is progressively applied, the stomach is gradually stretched and an elongated gastric tube fashioned. You do not do this with your technique. You cannot stretch the stomach in the same way when confined by the abdominal cavity. Further, you cannot palpate the tumor to see how far away you are from an EG junction tumor in applying your stapler. You also describe taking sutures into the apex of the stomach, which we are now preaching has to be traumatized as little as possible no sutures to drains, no beating up the stomach so that you wind up with a contused, ecchymotic fundus when you bring it up to the neck for the anastomosis. The fact that you pull the divided esophagus (with its contained tumor) sutured to the gastric tip through the mediastinum without progressively lengthening the stomach is worrisome. Is this part of the reason that you have a 10% leak rate? We have just reported an anastomotic leak rate of under 3% using an end-to-end stapled anastomotic technique and strictly minimizing intraoperative trauma to the stomach. And I wonder why your patients need to go to the ICU. Our patients do not have an ICU stay, and their average hospitalization is 6 to 7 days after our open transhiatal esophagectomy without thoracotomy. So I am not convinced that the availability of the video-assisted or laparoscopic technology necessarily justifies its use for this type of operation. I certainly can understand that your visualization of the mediastinal anatomy may be more comfortable with video-assisted technology. But I am very concerned that using this approach prevents optimal elongation of the stomach, results in excessive trauma to the stomach, and requires a pyloroplasty with its gastroduodenal suture-line in the abdomen that an open transhiatal esophagectomy avoids. I would like you to address some of these issues. I enjoyed your paper very much. DR LUKETICH: Thank you, Dr Orringer. The gastric elongation; I think what has made us comfortable in that area is that we do a very significant amount of laparoscopic benign work, and so our view of the stomach, we are very comfortable with that. Now, in terms of the mobilization of the stomach, we go to great lengths to minimize trauma to the stomach. We use multiple firings of the endo-gia to uncoil the stomach. Using atraumatic noncrushing graspers to hold the stomach fundus up just above the tip of the spleen. So we are actually trying to duplicate what you have taught us and others have taught us, to be careful with the gastric tube and to gain as much length as possible. We have not actually had to convert any case for lack of gastric tube length. I think you have to be extremely careful using atraumatic graspers and pay a lot of attention to minimizing trauma to the gastric tube. In terms of the margins, we do an on-the-table esophagoscopy, and in some cases we leave the scope in place. We do frozen sections in the operating room. If a margin is close, we staple out a strip of stomach between the specimen and the tubulonginal stomach to allow additional frozen sections to be performed. In our experience, we had 77 negative frozen section margins and only three came back on permanent pathology to be positive. But that remains a definite concern of ours and that is one reason why we approach it as I mentioned. Again, in terms of the ICU stay, I hesitate to say that we have demonstrated definite advantages over conventional surgery. Clearly, even our own experience with open surgery is moving towards shorter hospital stays. So I think some of this is related to DRGs, etc. We are avoiding ICU stays in some patients. So I appreciate your concern there. I am not sure that we have demonstrated an advantage, and I hope that was my message. I think it is a safe operation, but I am not at all sure that it is a better operation or that it is more cost-effective by any means at this point in time. That will require prospective multiinstitutional trials.