Chapter 1. Non-metastasized esophageal cancer. Biere SSAY, van Weyenberg SJ, Verheul HM, Mulder CJ, Cuesta MA, van der Peet DL

Transcription

1 Chapter 1 Non-metastasized esophageal cancer Biere SSAY, van Weyenberg SJ, Verheul HM, Mulder CJ, Cuesta MA, van der Peet DL Nederlands Tijdschrift voor Geneeskunde 2010; 154: A820 [translated from Dutch to English for this thesis)

2 ABSTRACT The incidence of esophageal carcinoma has increased recently predominantly due to a rise in the incidence of adenocarcinoma. A relation with the increasing prevalence of Barrett-esophagus is suggested. Staging must include esophago-gastro-duodenoscopy, transesophageal endo-echography and computed tomography. A higher sensitivity and specificity for metastasis can possibly be achieved with additional positron emission tomography. Neoadjuvant chemoradiotherapy followed by surgery for adenocarcinomas is associated with a better survival. This effect has not been proven for squamous cell carcinomas. Distal and gastro-esophageal tumors are particularly suitable for a transhiatal approach. Intrathoracic tumors probably require a transthoracic resection. Minimally invasive esophagectomy seems to be associated with a lower morbidity and a shorter hospital stay. Randomized trials are needed to substantiate these results. Chapter 1 18

3 Non-metastasized esophageal cancer 19 INTRODUCTION In recent years, the incidence of esophageal cancer has increased significantly. In the Netherlands 807 patients were diagnosed with esophageal cancer in In 2005 that number had doubled to 1546 (Figure 1). 1 This increase is due to the rising incidence of adenocarcinomas. Although the causes are not known, a changing diet, smoking and the increasing prevalence of obesity play a major role. 2,3 Moreover, there is a relation with the increasing prevalence of the so-called Barrett s esophagus (Table 1). 4,5 This is an acquired disease of the distal esophagus, in which the normal squamous epithelium is replaced by metaplastic columnar epithelium. The cause is probably the response to chronic reflux of gastric contents into the esophagus. The presence of a Barrett s esophagus increases the risk of adenocarcinoma by approximately 0.5% a year. 6 A Barrett s esophagus is also associated with an altered diet, obesity and it is possible related to an increase of reflux symptoms. 2,3,7 The severity of reflux symptoms is not related to the presence of a Barrett s esophagus. The number of years that a patient has symptoms of reflux, is a better predictor for the presence of a Barrett s esophagus. Figure 1. The rising incidence of esophageal cancer caused by the rising incidence of adenocarcinoma. Data of the Netherlands Cancer Registry. 1 Incidence (total) Year Esophageal cancer Adenocarcinoma Squamous cell carcinoma In this article we give an overview of the current situation concerning the treatment of esophageal cancer in relation to the value of neoadjuvant treatment, various surgical approaches with an emphasis on minimally invasive techniques. A search was conducted in the databases of Medline and Embase. Several relevant English-language search terms were combined, including esophagus, staging, treatment, neoadjuvant, minimally invasive, resection.

4 Table 1. Risk factors for esophageal carcinoma. Risk factor Squamous cell carcinoma Adenocarcinoma Tobacco use Alcohol use Barrett s esophagus Weekly reflux symptoms Obesity Achalasia Caustic injury to the esophagus Plummer-Vinson syndrome History of head and neck cancer History of breast cancer treated with radiotherapy Frequent consumption of extremely hot beverages + - A single plus sign indicates an increase in the risk by a factor of less than two, two plus signs an increase by a factor of two to four, three plus signs an increase by a factor of more than four to eight, and four plus signs an increase by a factor of more than eight. The plus minus sign indicates that conflicting results have been reported, and the dashes indicate that there is no proven risk. 2 DIAGNOSIS AND STAGING Chapter 1 20 Patients with alarm symptoms require further evaluation. These alarm symptoms are dysphagia, weight loss, severe vomiting, hematemesis, anemia with reflux or dyspeptic symptoms. Dysphagia is the most common symptom of esophageal cancer. The difficulty of swallowing food is the most common complaint in the general practitioners (GP) clinic. Dinner is often left aside resulting in weight loss and consequently a poorer prognose. 8 The first diagnostic tool should be an esophago-gastro-duodenoscopy with biopsies. If there is an esophageal carcinoma staging according to the Tumor-Node-Metastasis classification (TNM classification) is of importance (Table 2). The most appropriate treatment is chosen based on this classification. The staging includes endoscopic ultrasound for determining the clinical T and N stage. The T stage represents the growth of the tumor in the wall of the esophagus and surrounding structures. The N-stage represents the presence of suspicious regional lymph nodes. In a systematic review of 27 articles the value of the endoscopic ultrasound varies from 71 to 100% sensitivity and specificity of 66 to 100%. 9 A Computed Tomography (CT) scan of the neck, thorax and abdomen is primarily used to determine the M stage. This allows detection of any metastasis in other organs or in distant lymph nodes. This includes the cervical lymph nodes and lymph nodes around the celiac trunk. The imaging of the cervical lymph nodes can also be performed with a external ultrasound of the neck with possible cytological needle puncture. A Dutch study showed that an ultrasound of the neck in patients with normal CT scan might be useful. 10 Additional assessment could include a Positron Emission Tomography (PET), which has a higher sensitivity (78%) and specificity (98%) for organ metastasis or distant lymph node

5 Non-metastasized esophageal cancer 21 Table 2. Tumor-Node-Metastasis classification Stage Tumor invasion Node Metastasis Stage 0 Tis 0 0 Stage I T1 0 0 Stage IIa T2-T3 0 0 Stage IIb T1-T2 1 0 Stage III T3 T Stage IV T1-T4 0-1 M1 Stage IVa T1-T4 0-1 M1a Stage IVb T1-T4 0-1 M1b The primary tumor (T) is classified as follows: Tis, carcinoma in situ; T1, invasion of lamina propria or submucosa; T2, invasion of muscularis propria; T3, invasion of adventitia; and T4, invasion of adjacent structures. Regional lymph node metastases (N) are classified as follows: N0, no regional lymph node metastases; and N1, regional lymph node metastases. Distant metastases (M) are classified as follows: M0, no distant metastases; M1a, metastasis to cervical nodes in the case of cancer of the upper thoracic esophagus and metastasis to celiac nodes in the case of cancer of the lower thoracic esophagus; and M1b, other distant metastases. 2 metastasis compared to an endoscopic ultrasound and CT combined (sensitivity 37%, specificity 87%). 11 With the introduction of a multi-detector CT scan the additional value of PET is less convincing (detection of (lymph node) metastasis in 4% of the patients who had normal CT scan). 12 NEOAJDUVANT TREATMENT The role of chemoradiotherapy (chemoradiation) for the treatment of esophageal cancer was mainly determined by a study published in Patients with a squamous cell carcinoma or adenocarcinoma of the intrathoracic esophagus were randomized to radiotherapy (65Gy) or chemotherapy (cisplatin and 5-fluorouracil) combined with radiotherapy (50Gy). These patients did not undergo surgical resection. The trial had to be terminated prematurely due to a significantly better survival seen in the patients who were treated with combined therapy (five year survival 26% versus 0%). 13,14 Some randomized trials where chemoradiotherapy was used as neoadjuvant treatment were published in the years after this the abovementioned trial. Two meta-analysis analyzed the results these studies. 15,16 In the most recent meta-analysis, 10 randomized trials included a total of 1209 patients, a significantly better survival was observed for patients with an adenocarcinoma who underwent neoadjuvant chemoradiotherapy than the group who was treated with surgery alone (Hazard Ratio (HR) 0.75, Confidence Interval ). 16 For the squamous cell carcinomas, the survival benefit was less obvious due to the confidence interval (CI) which is close to 1 (HR 0.84, CI ).

6 The CROSS trial was completed at the end of 2008 in which patients were randomized to concurrent neoadjuvant chemotherapy (paclitaxel and carboplatin) and radiotherapy (41.4Gy) followed by surgery or surgery alone. 17 There were 175 patients randomized per study arm. Patients with either squamous cell carcinoma or adenocarcinoma were included in this study. The results of the trial are expected soon. Patients with T4 tumors (invasion of adjacent organs / structures such as pericardium, bronchus) are a special group. Basically, they do not qualify for a primary surgical resection. There seems to be evidence that locally advanced carcinomas can be treated with palliative chemoradiotherapy. 13,14 A five year survival rate of 26% can be reached. In addition, a good response to chemoradiotherapy could downstage the TNM stage. This allows patients after chemoradiotherapy who have a T4 tumor to be possible re-staged to a curative resection. The latter seems particularly the case for squamous cell carcinomas. 18,19,20 Optimization of peri-operative systemic therapy or either combined with radiotherapy is the challenge of the near future. Much is expected of new combination therapies including so-called targeted agents. CONVENTIONAL SURGERY Chapter 1 22 To date, surgery is the only curative option. Carcinoma in situ and T1 tumors may in some cases be treated by endoscopic mucosal resection. Currently, the vagal-sparing esophageal resection is investigated for this type of tumors. 21 For other resectable stages generally three surgical approaches are commonly used worldwide: the transthoracic resection with an intrathoracic anastomosis, the transthoracic resection with a cervical anastomosis and the transhiatal resection. Leakage of an intrathoracic anastomosis is potentially life threatening, as this could lead to mediastinitis and pleural empyema. Therefore, the transthoracic approach with an anastomosis at the level of the cervical esophagus was developed. A cervical anastomosis results in a higher anastomotic leakage and stenosis rate. These complications are however relatively easy to treat. In order to reduce the pulmonary morbidity caused by the transthoracic approach a transhiatal resection was introduced. Through a laparotomy the esophagus can be mobilized transhiatally and connected with a cervical anastomosis. This transhiatal approach limits the mediastinal lymph node dissection compared to the transthoracic approach. The impact of this difference in mediastinal dissection was studied in a large Dutch randomized trial (n = 220). 22,23 There were no significant difference in survival between the two approaches. Therefore, generally a transhiatal approach is used for distal esophageal cancer or gastro-esophageal junction carcinomas. In determining the surgical approach, there are centers that choose a transthoracic approach for distal esophageal carcinomas if there are suspicious lymph nodes present in the mediastinum on endoscopic ultrasound.

7 Non-metastasized esophageal cancer 23 All these three aforementioned conventional approaches are associated with significant morbidity and hospital stay. One of the potential methods for reducing morbidity and hospital stay, is the so-called minimally invasive surgery. To date, no randomized trials have been performed that compare minimally invasive esophageal resection with open esophageal resection. 24 Some large series and case-control studies are however published. MINIMALLY INVASIVE SURGERY Minimally invasive transthoracic esophageal resection. In 2003 an American group published the results of minimally invasive transthoracic resection in 222 patients with esophageal cancer. 25 A thoracoscopy and laparoscopy, in left lateral position, with a cervical anastomosis was performed. In total, 18% of the patients developed pulmonary complications (pneumonia 7.7%, 4.5% atelectasis by bronchoscopy, pleural drainage 6.3%). Anastomotic leakage occurred in 26 (11.7%) patients. The median length of stay in the intensive care unit (ICU) and hospital stay was 1 and 7 days respectively. Both the resection margin and the number retrieved lymph nodes were not reported in this study. In the 2-year follow-up the survival of patients with stage I tumors was 85%, 38% with stage II tumors and 26% with stage III tumors. This is similar to the 2-year survival rate of patients who underwent a conventional approach. One of the criticisms of this study was the fact that relatively many patients were selected for a minimally invasive approach with an esophageal cancer at an early stage. The largest series of patients undergoing minimally invasive transthoracic esophageal resection in prone position was published by an Indian team. 26 A total of 130 patients were described. All had a squamous cell carcinoma of the middle third of the esophagus. With this technique, thoracoscopic mobilization of the esophagus is performed while the patient lies on the stomach. There is no need for selective intubation, allowing the lung on the operated side to be ventilated. This is associated with fewer pulmonary complications compared with the approach in the left lateral position. After the thoracoscopic phase, the patient is turned to a supine position for the the abdominal laparoscopic phase. The anastomosis is made through a cervical incision. Three patients (2.3%) developed pulmonary complications. The incidence of anastomotic leakage was 2.3%. The median length of ICU stay was one day. The median total hospital stay was 8 days. None of the patients had a positive resection margin. The average number of lymph nodes retrieved was 18 (range 11 32) in this series. They reported a 3-year survival of stages I, IIa, IIb and III of 75%, 50%, 45% and 18% respectively. Striking in this study is the absence of a positive resection margin and the low overall morbidity. Minimally invasive transhiatal esophageal resection. One of the largest series of minimally invasive transhiatal esophageal resection is described by the VU university medical center in Fifty patients were eligible for a laparoscopic transhiatal approach. Nine patients (18%) developed pulmonary complications. Anastomotic leakage occurred in 4 patients (8%). The

8 median length of ICU stay and total hospital stay were 1 and 13 days respectively. In 82% of the patients a microscopic radical resection performed. The median number of retrieved lymph nodes was 14 (range 10 19). The 3-year survival in this study corresponds to series of the open approach. DISCUSSION Chapter 1 24 The results of minimally invasive esophageal resection are promising. This approach appears to lead to a lower incidence of pulmonary complications, a shorter stay in the intensive care and a shorter hospital stay. A long learning curve, longer operative time and higher costs are potentially hindering in the implementation of the technique. Although the oncological results appear to be similar to the conventional series, this should be critically examined in future studies. In addition, the presence of many different approaches uniformity in the surgical treatment is needed to perform multicenter trials. The localization of the tumor plays a decisive role. The transthoracic approach seems particularly suited to for intrathoracic esophageal carcinomas whereas the transhiatal approach is used for tumors of the gastro-esophageal junction. 28,29 The presence of (endoscopic ultrasound) suspected intrathoracic lymph nodes may also play a role in the choice of the surgical approach for distal esophageal carcinomas. In the aforementioned studies minimally invasive esophageal resection with dissection of lymph nodes appear to be feasible. A significant proportion of patients with esophageal cancer are not eligible for a curative resection at the time of diagnosis. Most patients who can undergo a curative treatment, surgical resection is the only option. Optimization of surgical treatment can reduce postoperative morbidity and a shortening of hospital stay. The rising incidence of esophageal cancer is an additional reason to perform randomized studies on optimization of surgical treatment. At this moment, the VU university medical center started a multicenter randomized trial: the TIME-trial (Traditional Invasive vs. Minimally invasive Esophagectomy). In this study, the conventional transthoracic esophageal resection is compared to the minimally invasive esophageal resection in the prone position. All patients with an intrathoracic esophageal carcinoma are eligible for inclusion. The first results are expected in about two years.

9 Non-metastasized esophageal cancer 25 REFERENCES 1. Nederlandse Kanker Registratie. Incidentiecijfers oesofaguscarcinomen. Integrale Kankercentra < 2. Enzinger PC, Mayer RJ. Esophageal Cancer. New England Journal of Medicine. 2003; 349: Souza RF, Spechler SJ. Concepts in prevention of adenocarcinoma of the distal esophagus and proximal stomach. CA: A Cancer Journal for Clinicians. 2005; 55: Van Soest EM, Dieleman JP, Siersma PD, Sturkenboom MC, Kuiper EJ. Increasing incidence of Barrett s oesophagus in the general population. Gut 2005; 54: Post PN, Siersma PD, van Dekken H. Rising incidence of clinically evident Barrett s oesophagus in the Netherlands: a nation-wide registry of pathology reports. Scandinavian Journal of Gastroenterology. 2007; 42: Shaheen NJ, Crosby MA, Bozymski EM, Sandler RS. Is there publication bias in the reporting of cancer risk in Barrett s esophagus? Gastroenterology. 2000; 119: Shaheen NJ, Richter JL. Barrett s oesofagus. Lancet. 2009; 373: Deans DA, Wigmore SJ, de Beaux AC, Paterson-Brown S, Garde OJ, Fearon KC. Clinical prognostic scoring system to aid decision-making in gastro-oesofageal cancer. British Journal of Surgery. 2007; 94: Kelly S, Harris KM, Berry E, Hutton J, Roderick P, Cullingworth J et al. A systematic review of the staging performance of endoscopic ultrasound in gastro-oesophageal carcinoma. Gut. 2001; 49: Omloo JM, van Heijl M, Smits NJ, Phoa SS, van Berge Henegouwen MI, Sloof GW et al. Additional value of external ultrasonography of he neck after CT and PET scanning in the preoperative assessment of patients with esophageal cancer. Digestive Surgery. 2009; 26: Heeren PAM. Staging and patient selection in cancer of the oesophagus and the oesophagogastric junction [dissertation]. Groningen: Universiteit van Groningen; Van Westreenen HL, Westerterp M, Sloof GW, Groen H, Bossuyt PMM, Jager PL et al. Limited additional value positron emission tomography in staging esophageal cancer. British Journal of Surgery. 2007; 94: Herskovic A, Martz K, Al-Sarraf M, Leichman L, Brindle J, Vaitkevicius V et al. Combined chemotherapy and radiotherapy compared with radiotherapy alone in patients with cancer of the esophagus. New England Journal of Medicine. 1992; 326: Cooper JS, Guo MD, Herskovic A, Macdonald JS, Martenson JA Jr, Al-Sarraf M et al. Chemoradiotherapy of locally advanced esophageal cancer: long term follow-up of a prospective randomized trial (RTOG 85-01). Journal of the American Medical Association 1999; 281: Urschel JD, Vasan H. A meta-analysis of randomized trials that compared neoadjuvant chemoradiation and surgery to surgery alone for resectable esophageal cancer. American Journal of Surgery 2003; 185: Gebski V, Burmeister B, Smithers BM, Foo K, Zalcberg J, Simes J. Survival benefits from neoadjuvant chemoradiotherapy or chemotherapy in esophageal carcinoma: a meta-analysis. Lancet Oncology 2007; 8: van Heijl M, van Lanschot JJ, Koppert LB, van Berge Henegouwen MI, Muller K, Steyerberg EW et al. Neoadjuvant chemoradiation followed by surgery versus surgery alone for patients with adenocarcinoma or squamous cell carcinoma of the esophagus (CROSS). Biomed Central Surgery. 2008; 8: Kleinberg L, Forastiere AA. Chemoradiation in the management of esophageal cancer. Journal of Clinical Oncology. 2007; 25: Bedenne L, Michel P, Bouche O, Milan C, Mariette C, Conroy T et al. Chemoradiation followed by surgery compared with chemoradiation alone in squamous cancer of the esophagus: FFCD Journal of Clinical Oncology. 2007; 25: Stahl M, Stuschke M, Lehmann N, Meyer HJ, Walz MK, Seeber S et al. Chemoradiation with and without surgery in patients with locally advanced squamous cell carcinoma of the esophagus. Journal of Clinical Oncology. 2005; 23: Peyre CG, DeMeester SR, Rizzetto C, Bansal N, Tang AL, Ayazi S et al. Vagal-sparing esophagectomy: the ideal operation for intramucosal adenocarcinoma and barrett with high-grade dysplasia. Annals of Surgery. 2007; 246:

10 22. Hulscher JB, van Sandwick JW, de Boer AG, Wijnhoven BP, Tijssen JG, Fockens P et al. Extended transthoracic resection compared with limited transhiatal resection for adenocarcinoma of the esophagus. New England Journal of Medicine. 2002; 347: Omloo JM, Lagarde SM, Hulscher JB, Reitsma JB, Fockens P, van Dekken H et al. Extended transthoracic resection compared with limited transhiatal resection for adenocarcinoma of the mid/distal esophagus: five-year survival of a randomized clinical trial. Annals of Surgery. 2007; 246: Biere SSAY, Cuesta MA, van der Peet DL. Minimally invasive versus open esophagectomy for cancer: a systematic review and meta-analysis. Minerva Chirurgica. 2009; 64: Luketich JD, Alvelo-Rivera M, Buenaventura PO, Christie NA, McCaughan JS, Litle VR et al. Minimally invasive esophagectomy: outcome in 222 patients. Annals of Surgery 2003; 238: Palanivelu C, Prakash A, Senthilkumar R, Senthilathan P, Parathasarthi R, Rajan PS et al. Minimally invasive esophagectomy: thoracoscopic mobilization of the esophagus and mediastinal lymphadenectomy in prone position-experience of 130 patients. Journal of the American College of Surgeons. 2006; 203: Scheepers JJG, Veenhof AAFA, van der Peet DL, van Groeningen C, Mulder C, Meijer S et al. Laparoscopic transhiatal resection for malignancies of the distal esophagus: outcome of the first 50 resected patients. Surgery. 2008; 143: Cuesta MA, van den Broek WT, van der Peet DL, Meijer S. Minimally invasive esophageal resection. Seminars in Laparoscopic Surgery. 2004; 11: Scheepers JJG, Mulder CJ, van der Peet DL, Meijer S, Cuesta MA. Minimally invasive oesophageal resection for distal oesophageal cancer: a review of literature. Scandinavian Journal of Gastroenterology. 2006; 41 suppl 243: Chapter 1 26

11 Chapter 2 Minimally invasive versus open esophagectomy for cancer: a systematic review and meta-analysis Biere SSAY, Cuesta MA, van der Peet DL Minerva Chirugica 2009; 64:

12 ABSTRACT Chapter 2 28 Background Evidence on the benefits of minimally invasive surgery over open procedures in gastro-intestinal surgery is continuing to accumulate. This is also the case for esophageal surgery. Esophageal cancer often requires extensive surgery and is therefore considered to be one of the most invasive elective gastro-intestinal procedures. Clinical studies investigating means to reduce the invasive nature of the surgery are currently being received with great interest. A systematic review and meta-analysis of present literature was performed to evaluate the effects of minimally invasive esophagectomy (MIE) versus open esophagectomy on outcome. Methods All comparative studies comparing MIE with open esophagectomy for cancer were included. Eligible studies were identified from three electronic databases (Medline, Embase, Cochrane) and through a cross-reference search. Three comparative groups were created for (meta-) analysis: 1. Total MIE vs. open transthoracic esophagectomy (TTE) 2. Thoracoscopy and laparotomy vs. open TTE. 3. Laparoscopy vs. open transhiatal esophagectomy (THE). Results Ten studies were identified after a comprehensive search. One controlled clinical trial and 9 case-control studies, comprising 1061 patients, were retrieved. Trends were observed in the various studies in favor of MIE for the following outcome parameters: major morbidity, pulmonary complications, anastomotic leakage, mortality, length of hospital stay, operating time and blood loss. The meta-analysis in group 1 showed no significant differences between the groups for major morbidity or pulmonary complications OR 0.88 (95% CI , p=0.78) and OR 1.05 (95% CI , p=0.91) respectively. In group 2 significantly fewer cases of anastomotic leakage were reported in the MIE group OR 0.51 (95% CI , p=0.03). In both group 1 and 2 a trend toward a reduced mortality was seen in the MIE group, although no statistical significance was reached (group 1: OR 0.58 (95 % CI , p=0.64), group 2: OR 0.59 (95% CI , p=0.34)). No meta-analysis could be performed for group 3 due to incomplete data of the selected outcome parameters in the various studies. Conclusion A faster post-operative recovery and therefore a reduction in morbidity could be achieved with MIE. Furthermore, a lower mortality rate with the implementation of MIE can be realised. MIE is investigated in case-control studies and bias may have been introduced simply by study design. Therefore, randomized trials comparing MIE with open esophagectomy are necessary in order to evaluate outcome more efficiently.

13 Minimally invasive versus open esophagectomy 29 INTRODUCTION Surgery for malignant disease of the esophagus is considered to be one of the most extensive and traumatic gastro-intestinal surgical procedures. The classic resection is associated with a significant morbidity which necessitates post-operative care in the intensive care unit and carries a significant risk of death. Minimally invasive esophagectomy (MIE) is considered to be an alternative in order to reduce the extensive nature of the required surgery and the associated risk of morbidity and mortality. Throughout the last decade, evidence of the benefits of minimally invasive surgery over open procedures in gastro-intestinal surgery has continued to accumulate 1. Faster post-operative recovery, fewer peri-operative complications and a shorter duration of hospital stay appear to be the main advantages for patients. The learning curve and higher costs however seem to be the main drawbacks. Importantly, the oncological outcomes of minimally invasive surgery and open procedures are similar. The implementation of minimally invasive surgery in esophagectomy is gradually becoming more and more widespread. Clinical studies evaluating the use of laparoscopy and/ or thoracoscopy in the transthoracic esophagectomy (TTE) and laparoscopy in the transhiatal esophagectomy (THE) are ongoing. Luketich et al 2 and Palanivelu et al 3 reported on their experience with total MIE in TTE. Scheepers et al 4 reported on 50 procedures of minimally invasive THE. Gemmill and McCulloch 5 previously systematically reviewed minimally invasive resection for gastro-esophageal cancer. No direct comparison was however made with the traditional open approach and the outcome parameters were not stratified for the different operative approaches. Furthermore, their analysis also included minimally invasive gastrectomies. Therefore, we conducted a systematic review including only studies directly comparing minimally invasive esophagectomy versus conventional open esophagectomy. METHODS Literature search The following electronic databases were used for a literature search: Medline (1950 to September 2008), Embase (1980 to September 2008), and the Cochrane Library (2008 issue 3). A comprehensive search was performed using the following search terms: esophageal cancer, cardia cancer, minimally invasive, laparoscopy, thoracoscopy. Logical combinations of these and related terms (e.g. oesophagus, carcinoma, neoplasms) were used to maximize sensitivity. Furthermore, a truncation symbol was used in each database in order to allow retrieval of all suffix variations of a root word. After identifying relevant titles, the abstracts of these studies were read to decide if the study was eligible. The full article was retrieved when the information in the title and/ or abstract appeared

14 to meet the objective of this review. A manual cross-reference search of the bibliographies of relevant articles was conducted to identify studies not found through the computerized search. The related articles feature of Pubmed was simultaneously used. All published studies comparing minimally invasive esophagectomy with open resection for esophageal or gastro-esophageal (GE) junction cancer were included. Study selection criteria Given the paucity of the available evidence addressing the study question, the search was not restricted to randomized controlled trials (RCTs). All comparative studies (including case-matched studies) were included as well. Studies comparing minimally invasive esophagectomy, i.e. thoracoscopy and/ or laparoscopy, with open esophagectomy for carcinoma of the esophagus and GE junction were eligible for inclusion. Only studies in English were included for analysis. There were no further restrictions for inclusion. The following primary outcome parameters were assessed: major morbidity (e.g. pulmonary complications, anastomotic leakage, pulmonary embolism, gastric volvulus, cardiac failure, haemorrhage, chyle leakage), pulmonary complications, anastomotic leakage, mortality, length of hospital stay. Pneumonia and atelectasis were used as the main definition for pulmonary complications. Secondary outcome parameters were also assessed: operating time, blood loss, conversion to open surgery, lymph node retrieval, quality of life, costs. Chapter 2 30 Quality assessment of retrieved articles Each included article was appraised by two reviewers, who assessed the methodological quality of the selected studies independently. A critical review checklist of the Dutch Cochrane Centre was used to appraise RCTs [ Last accessed September 2008]. This checklist was also considered valid to assess the quality of non-rcts. The first six items of the checklist were negative for all the studies and were therefore not included for further appraisal (items: 1. Randomization? 2. Treatment allocation concealed? 3. Eligibility criteria specified? 4. Patient blinded? 5. Outcome assessor blinded? 6. Care provider blinded?). Data extraction and analysis Data was extracted on preformatted sheets. The studies were tabulated and methodologically evaluated to assess homogeneity. In the case of heterogeneity between the studies, it would not be justified to pool the assessed outcomes. In general, two approaches are used in esophagectomy: 1. transthoracic esophagectomy (TTE) which consists of: a. mobilization of the esophagus through thoracotomy or thoracoscopy; b. abdominal approach i.e. laparotomy or laparoscopy for reconstruction of the gastric conduit; c. cervicotomy for gastroesophageal anastomosis. 2. transhiatal esophagectomy (THE) which consists of: a. laparoscopy or laparotomy for mobilization of the intrathoracic esophagus and reconstruction of the gastric conduit; b. cervicotomy for gastroesophageal anastomosis.

15 Minimally invasive versus open esophagectomy 31 Because of expected clinical heterogeneity, three comparative groups were created in order to allow a (meta-)analysis of the data for the (primary) outcome parameters. These comparative groups were: 1. thoracoscopy and laparoscopy (total MIE) vs. open TTE, 2. thoracoscopy and laparotomy (TO) vs. open TTE, 3. laparoscopy vs. open THE. Statistical analysis Meta-analyses were performed for the five primary outcome parameters: major morbidity, pulmonary complications, anastomotic leakage, mortality, length of hospital stay. Quantative data for the primary outcome parameters were entered into the software Cochrane Review Manager (RevMan) version 4.3 and analyzed using RevMan Analyses (Cochrane Collaboration, Oxford, UK). Summary estimates, including 95% confidence intervals, were calculated. For continuous outcome data (e.g. length of hospital stay) means and standard deviations were used to calculate a weighted mean difference in the meta-analysis. For dichotomous outcomes (e.g. morbidity, mortality) the odds ratio (OR) was calculated. Statistical heterogeneity was tested using X 2 and I 2 tests. Data were pooled using the random effect model, because moderate heterogeneity was suspected (e.g. indication for surgery, major morbidity). RESULTS The initial search yielded 1388 potential literature citations (Figure 1). Of these, 1336 were excluded after scanning titles; another 30 studies were excluded after reading the abstracts. Of the 22 publications considered potentially relevant, 12 were subsequently excluded. Of these, 6 comparative studies 6 11 were excluded due to non-english language publication and 2 studies 12,13 which reported on an overlapping patient population were also excluded. A total of 10 studies were included for the systematic review (table 1) These studies involved 1061 patients, of which 560 underwent minimally invasive TTE and 59 underwent minimally invasive THE. The number of patients in the various studies ranged from 4 to 309. One controlled clinical trial 22 and 9 case-control studies 14 21,23 were retrieved. No randomized trials were found. The quality assessment of all studies is shown in table 1. Three studies had non-similar baseline characteristics 14,15,17. In addition, only three studies had similar non-surgery related treatment Table 2 shows data on the studied outcome parameters. No meta-analysis could be performed for the primary outcome parameters in the THE group due to lack of information on these parameters. The mean length of hospital stay was reported in several studies 22,23, whereas in some studies medians were reported for the duration of hospital stay 14,15,19. Therefore, no meta-analysis could be performed for the length of hospital stay. No study investigated either quality of life and or costs involved in esophagectomy.

16 Figure 1. Search strategy Search strategy 1388 references 52 studies selected on title 1336 considered not relevant based on title 17 lack of control group 13 review or tutorial articles 22 studies selected on abstract 3 studies compared two minimally approaches 2 studies reporting on overlapping patient data 1 review article 6 comparative non-english language studies excluded Chapter 2 10 full-text articles included in SR 32 Two studies did not specify whether laparoscopy or laparotomy was used in the minimally invasive TTE 18,19. A meta-analysis would not be justified as heterogeneity is clearly present. The study of Bernabe et al 23 compared hand assisted laparoscopic (HALS) transhiatal esophagectomy with open. Therefore, this study was not included in the meta-analysis. Primary outcome parameters Major morbidity TTE Two studies reported on major morbidity for total MIE TTE vs. open TTE 14,15. The meta-analysis did not show any significant difference in major morbidity between the two groups (OR 0.88 (95% CI , p=0.78), figure 2). Only 1 of the 3 thoracoscopic and laparotomic (TO) TTE studies reported on major morbidity, there was no significant difference between TO TTE and open TTE (62% and 66.7% respectively) 15. Braghetto et al 18 observed significantly less major morbidity in the MIE group compared to the open TTE group (29.7% vs. 56.6%, p=0.004). Fabian et al 19 reported no differences in major morbidity.

17 Minimally invasive versus open esophagectomy 33 Table 1. Characteristics of the included studies. Study Study design Numbers Indication Groups similar at baseline? TTE Total MIE vs. open Follow-up? Intention to treat? Similar nontrial treatment? Morris Case-control study 4 vs. 7 - upper aerodigestive tract carcinoma ± Yes Yes Not stated - advanced thyroid cancer - middle esophageal carcinoma - Barret s high grade dysplasia - distal esophageal obstruction Smithers Case-control study 23 vs Carcinoma of the esophagus ± Yes ± Not stated - GE junction Thoracoscopy and laparotomy vs. open Osugi Case-control study 77 vs Carcinoma of the esophagus Yes Yes Yes Not stated Shiraishi Case-control study 78 vs Carcinoma of the esophagus ± Yes Yes ± Smithers Case-control study 309 vs Carcinoma of the esophagus - GE junction ± Yes ± Not stated Thoracoscopy and laparoscopy/laparotomy vs. open Braghetto Case-control study 47 vs Carcinoma of the esophagus Not stated Yes Yes Not stated Fabian Case-control study 22 vs Carcinoma of the esophagus Yes Yes Yes Not stated THE Laparoscopy vs. open Bresadola Case-control study 8 vs. 8 - Carcinoma of the esophagus Yes Yes Yes Yes Van den Broek Case-control study 25 vs Carcinoma of the esophagus Yes Yes Yes Yes - GE junction Scheepers Controlled clinical trial 9 vs. 8 - Carcinoma of the esophagus Yes No Yes Yes Hand-assisted laparoscopy vs. open Bernabe Case-control study 17 vs Barrett s high grade dysplasia Yes Yes Yes Not stated - Carcinoma of the esophagus TTE: Transthoracic Esophagectomy, THE: Transhiatal Esophagectomy, MIE: Minimally Invasive Esophagectomy. GE, Gastro-Esophageal.

18 Chapter 2 34 Table 2a. Primary outcome parameters. Study Major morbidity Pulmonary complications Anastomotic leak Mortality (30-day) Length of hospital stay days (mean (SD)) TTE Total MIE vs. open MIE Open MIE Open MIE Open MIE Open MIE Open (NR) * (NR) * 12,441 NR 0 14% (1/7) 14% (1/7) 25% (1/4) 71% (5/7) Morris % (4/4) (7-49) * (8-44) * 0 2.6% (3/114) 8,7% (10/114) 4% (1/23) 27.8% (35/114) 30% (7/23) 66.7% (76/114) Smithers % (14/23) Thoracoscopy and laparotomy vs. open 0 0 NR NR 2.8% (2/72) 1.4% (1/77) 19,4% (14/72) Osugi 2003 NR NR 15.6% (12/77) NR NR 8.1% (3/37) 2.6% (2/78) 24.3% (9/37) 11.5% (9/78) 32.4% (12/37) Shiraishi 2006 NR NR 20.5% (16/78) (8-123) * (8-44) * 2.6% (3/114) 2.3% (7/309) 8,7% (10/114) 5.5% (17/309) 27.8% (35/114) 26% (80/309) 66.7% (76/114) Smithers % (193/ 309) Thoracoscopy and laparoscopy/ laparotomy vs. open NR NR 10.9% (13/119) 6.3% (3/47) 14.3% (17/119) 6.4% (3/47) 18.5% (22/119) 14.8% (7/47) 37.8% (45/119) Braghetto % (10/47) 9.5 * 11 * 9.8% (4/43) 4.5% (1/22) 7% (3/43) 14% (3/22) 19% (8/43) 5% (1/22) 30% (13/43) Fabian % (5/22) THE Laparoscopy vs. open Bresadola 2006 NR NR NR NR NR NR NR NR NR NR NR NR (NR) * (NR) * 10% (2/20) 8% (2/25) 72% (18/20) Vd Broek % (14/25) Scheepers 2008 NR NR NR NR NR NR NR NR 13 (11-16) 16 (14-20) Hand-assisted laparoscopy vs. open 11.6 (2.9) Bernabe 2005 NR NR NR NR NR NR 0 NR 9.1 (3.2) TTE: Transthoracic Esophagectomy, THE: Transhiatal Esophagectomy, MIE: Minimally Invasive Esophagectomy, NR: Not Reported, * : Median (range), : grams

19 Minimally invasive versus open esophagectomy 35 Table 2b. Secondary outcome parameters. Study Operating time min (mean (SD)) Blood loss ml (mean (SD)) Conversion Lymph node retrieval (mean (SD)) TTE Total MIE vs. open MIE Open MIE Open MIE Open Morris (NR) 549 (NR) 900 (NR) 967 (NR) Quality of life Costs NR NR NR NR NR Smithers Excluded NR NR ( ) * ( ) * ( ) * (0-3000) * (9-33) * (1-44) * Thoracoscopy and laparotomy vs. open Osugi (80) Shiraishi (87.1) 186 (35) (110.5) NR 33.9 (330) (170) (12) (561.1) (429.7) 32.8 (14) NR NR NR NR NR NR NR Smithers Excluded NR NR ( ) * ( ) * (0-1500) * (0-3000) * (2-59) * (1-44) * Thoracoscopy and laparoscopy/ laparotomy vs. open Braghetto 2006 NR NR NR NR NR NR NR NR NR Fabian (72) 270 (87) THE Laparoscopy vs. open Bresadola (59) Vd Broek (52) (19) 257 (34) Scheepers % ( ) * ( ) * ( ) * ( ) * (3/6) Hand-assisted laparoscopy vs. open 178 (96) 356 (136) 5% (1/20) 15 (6) 8 (7) NR NR NR NR 0 NR NR NR NR 600 (216) 900 (13) 36% (9/25) 7 (4.9) 6.5 (4.9) NR NR NR NR (9-16.5) * (8-15) * Bernabe (53) 388 (102) 331 (220) 542 (212) (NR) 9.8 (NR) NR NR TTE: Transthoracic Esophagectomy, THE: Transhiatal Esophagectomy, MIE: Minimally Invasive Esophagectomy, NR: Not Reported, SD: Standard Deviation, * : Median (range), : gram

20 Figure 2. Major morbidity total minimally invasive esophagectomy (MIE) Study or Subgroup Morris 2007 Smithers 2007 Total MIE Open Odds Ratio Odds Ratio Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI % 92.7% 4.09 [0.15, ] 0.78 [0.31, 1.96] Total (95% CI) % Total events Heterogeneity: Tau² = 0.00; Chi² = 0.92, df = 1 (P = 0.34); I² = 0% Test for overall effect: Z = 0.29 (P = 0.78) 0.88 [0.36, 2.14] Favours MIE Favours open THE Major morbidity was reported in only 1 study, i.e. van den Broek et al 21. They reported comparable morbidity in the laparoscopy and open THE groups (72% vs. 70%). Pulmonary complications Chapter 2 36 TTE No significant difference in pulmonary complications was seen in the meta-analysis of total MIE vs open TTE (OR 1.05 (95% CI , p=0.91), figure 3). Although a trend was seen for fewer pulmonary complications in the TO TTE group compared to the open TTE group in the metaanalysis, this did not reach statistical difference (OR 0.73 (95% CI , p=0.10), figure 4). Braghetto et al 18 observed more pulmonary complications in the open approach (14.8% MIE vs. 20% open, NS). Fabian et al 19 reported significantly fewer respiratory failures in the MIE group compared to the open procedure (5% and 23% respectively). Figure 3. Pulmonary complications total minimally invasive esophagectomy (MIE) Study or Subgroup Morris 2007 Smithers 2007 Total MIE Open Odds Ratio Odds Ratio Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI % 91.0% 2.00 [0.09, 44.35] 0.99 [0.37, 2.61] Total (95% CI) % Total events 8 36 Heterogeneity: Tau² = 0.00; Chi² = 0.18, df = 1 (P = 0.67); I² = 0% Test for overall effect: Z = 0.11 (P = 0.91) 1.05 [0.42, 2.66] Favours MIE Favours open Figure 4. Pulmonary complications thoracoscopy and laparotomy, transthoracic esophagectomy (TO TTE) Study or Subgroup Osugi 2003 Shiraishi Smithers 2007 Thoracoscopy/laparotomy Open Odds Ratio Odds Ratio Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI % 18.0% 62.6% 0.76 [0.33, 1.79] 0.54 [0.22, 1.30] 0.79 [0.49, 1.26] Total (95% CI) Total events Heterogeneity: Tau² = 0.00; Chi² = 0.58, df = 2 (P = 0.75); I² = 0% Test for overall effect: Z = 1.64 (P = 0.10) 100.0% 0.73 [0.50, 1.06] Favours MIE Favours open

21 Minimally invasive versus open esophagectomy 37 THE Only one study specified data on pulmonary complications, they reported comparable results (8% vs. 10%) 21. Bernabe et al 23 did not report on pulmonary complications. Anastomotic leakage TTE Smithers et al 15 observed no significant difference between total MIE vs. open surgery. There was however, significantly less anastomotic leakage in TO TTE category compared to the open procedure (OR 0.51 (95% CI , p=0.03), figure 5). Braghetto et al 18 reported anastomotic leakage in 6.4% of the MIE procedures compared to 16.6% cases in the procedures with an open approach (NS). Fabian et al 19 observed no significant differences in anastomotic leakage (14% vs. 7%). Figure 5. Anastomotic leakage thoracoscopy and laparotomy, transthoracic esophagectomy (TO TTE) Study or Subgroup Osugi 2003 Shiraishi Smithers 2007 Thoracoscopy/laparotomy Open Odds Ratio Odds Ratio Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI % 36.2% 57.4% 0.46 [0.04, 5.19] 0.41 [0.15, 1.13] 0.61 [0.27, 1.36] Total (95% CI) Total events Heterogeneity: Tau² = 0.00; Chi² = 0.37, df = 2 (P = 0.83); I² = 0% Test for overall effect: Z = 2.12 (P = 0.03) 100.0% 0.51 [0.28, 0.95] Favours MIE Favours open THE Anastomotic leakage was not reported by any of the three THE studies. Bernabe et al 23 also did not report on anastomotic leakage. Mortality TTE No significant differences in mortality were seen in the total MIE compared to the open TTE or the TO TTE compared to the open TTE categories in the meta-analysis (OR 0.58 (95% CI , p=0.64), figure 6 and OR 0.59 (95% CI , p=0.34), figure 7 respectively). However, a trend was observed in favor of both minimally invasive approaches. The case control study by Braghetto et al 18 showed no significant differences in mortality figures between the MIE TTE and open TTE groups (6.3% vs. 10.9%) Also, no significant reduction in mortality was observed by Fabian et al 19 (4.5% vs. 9.8%).

22 Figure 6. Mortality total minimally invasive esophagectomy (MIE) Study or Subgroup Morris 2007 Smithers 2007 Total MIE Open Odds Ratio Odds Ratio Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI % 56.6% 0.48 [0.02, 14.70] 0.68 [0.03, 13.57] Total (95% CI) % Total events 0 4 Heterogeneity: Tau² = 0.00; Chi² = 0.02, df = 1 (P = 0.88); I² = 0% Test for overall effect: Z = 0.47 (P = 0.64) 0.58 [0.06, 5.56] Favours MIE Favours open Figure 7. Mortality thoracoscopy and laparotomy, transthoracic esophagectomy (TO TTE) Study or Subgroup Osugi 2003 (1) Shiraishi Smithers 2007 Thoracoscopy/laparotomy Open Odds Ratio Events Total Events Total Weight M-H, Random, 95% CI % 64.2% Not estimable 0.30 [0.05, 1.87] 0.86 [0.22, 3.37] Odds Ratio M-H, Random, 95% CI Total (95% CI) Total events 9 6 Heterogeneity: Tau² = 0.00; Chi² = 0.82, df = 1 (P = 0.37); I² = 0% Test for overall effect: Z = 0.95 (P = 0.34) 100.0% 0.59 [0.20, 1.76] Favours MIE Favours open (1) Not reported Chapter 2 THE No mortality was observed in any of THE studies. Bernabe et al 23 also did not report on procedures leading to fatality in either the hand-assisted laparoscopic category or in the open THE group. 38 Length of hospital stay TTE Smithers et al 15 reported a median length of hospital stay in the group of total MIE which was significant shorter than in open TTE category (11 vs. 14 days). Furthermore, there was a significantly shorter duration of hospital stay in the total MIE TTE category compared to the TO TTE category (11 vs. 13 days) and the length of hospital stay was also significantly shorter in the TO TTE group compared to the open TTE category (13 vs. 14 days). Morris et al 14 reported no significant difference in hospital stay (22 vs.18.5 days, p=0.63). Median length of hospital stay did not significantly differ between MIE and open surgery in the study by Fabian et al 19. THE Scheepers et al 22 reported a significantly shorter median length of hospital stay for patients undergoing laparoscopic THE compared to those undergoing open surgery (13 and 16 days respectively). Van den Broek et al 21 did not observe any differences in duration of hospital stay (16 days). The mean duration of hospital stay in the study by Bernabe et al 23 was 9.1 days in HALS THE group and was significantly shorter than to that of the open THE procedure (11.6 days, p=0.037).

23 Minimally invasive versus open esophagectomy 39 Secondary outcome parameters Operating time TTE Morris et al 14 reported no significant differences in operating time in the total MIE category compared to open TTE procedure (615 min vs. 549 min). Smithers et al 15 observed a significantly longer median operating time in the total MIE category compared to the open approach (330 min vs. 300 min, p=0.01). In comparison to the open TTE procedure, a significantly longer operating time was reported in the TO TTE categories by Osugi et al 16, Shiraishi et al 17 and Smithers et al 15. Braghetto et al 18 did not report on operating time. Fabian et al 19 reported a significantly longer operating time for MIE as compared to open MIE (333 min and 270 min respectively). THE Operating times were comparable in the studies by Bresadola et al 20 and Scheepers et al 22 when comparing laparoscopic THE vs. open THE. Van den Broek et al 21 reported a significantly longer operating time for MIE THE (300 min vs. 257 min, p=0.003). Operating time was comparable in the study by Bernabe et al 23 when comparing HALS vs. open THE. Blood loss TTE No significant differences between the two groups were observed with regard to blood loss in the study of Morris et al 14. In contrast, Smithers et al 15 reported significantly less blood loss in the total MIE TTE category compared to the open procedure (300 ml and 600 ml respectively). Significant less blood loss was also reported in two studies comparing TO TTE to open TTE 15,16. Braghetto et al 19 did not report on blood loss. Fabian et al 19 observed significantly less blood loss in the MIE group (178 ml vs. 356 ml). THE Significantly less blood loss was described in the laparoscopic THE procedure compared to the open procedure in the study by van den Broek et al 21 and Scheepers et al 22. When comparing open THE to HALS, Bernabe et al 23 reported significantly more blood loss in open THE category than in the HALS category (542 ml and 331 respectively, p=0.037).

24 Conversion to open surgery TTE Smithers et al 15 excluded all procedures from their analysis which were converted to open surgery. Morris et al 14 did not report on conversion. No study evaluating the benefits of TO TTE reported on conversion rate. THE No conversion to open surgery was necessary in any of the three THE studies Similarly no conversion to open surgery was reported in the study by Bernabe et al 23. Lymph node retrieval Chapter 2 TTE Lymph node retrieval was comparable in total MIE and open MIE categories in the study by Smithers et al 15 (17 and 16 median respectively). When describing the TO TTE procedures, 2 of the 3 studies reported the number of lymph node retrieved 15,16. There were no significant differences. Braghetto et al 18 did not report on lymph node retrieval. Fabian et al 19 reported significantly more lymph node retrieval in the MIE group compared to the open THE group (15 and 8 respectively). 40 THE Lymph nodes retrieval was comparable in both categories in the studies by van den Broek et al 21 and Scheepers et al 22. Bernabe et al 23 reported on mean number of 8.7 LN retrieved for HALS THE group, compared to 9.8 LN retrieved in the open THE category. No statistical significance was reported. Quality of life No study investigated the quality of life. Costs No analysis was performed of the costs involved in the included studies. DISCUSSION A faster post-operative recovery resulting from a reduction of peri-operative morbidity is the main purpose of minimally invasive surgery. Although numerous studies on esophagectomy for esophageal cancer have been performed, few studies compare MIE with traditional open

25 Minimally invasive versus open esophagectomy 41 esophagectomy. Furthermore, no randomized trials have been performed to date. The overall quality of the studies is moderate as the case-control design inevitably introduces selection bias. Various studies showed a trend in favor of several investigated outcome parameters MIE (e.g. pulmonary complications, mortality). In the meta-analysis MIE did not seem superior to open esophagectomy. The retrospective design of the included studies could be the cause of this result. In addition, anastomotic leakage was a less frequent complication in the combined thoracoscopy and laparotomy TTE than in the procedure entailing thoracotomy and laparotomy TTE. This could be explained by a better step by step thoracoscopic mobilization in the former approach. With the introduction of total MIE a faster post-operative recovery is expected. This is primarily caused by fewer pulmonary complications. In this review only 2 studies compared total MIE TTE with open TTE 14,15. However, no significant differences in the rate of pulmonary complications were found in the meta-analysis. Morris et al 14 included esophagectomy for upper aerodigestive tract cancer and advanced thyroid cancer. This could have influenced the outcome. Furthermore, their study was a report of their first experiences with total MIE. This could explain their high incidence of pulmonary complications, as case volume and experience are related to outcome 24. Luketich et al 2 reported on the largest total MIE TTE series. In their series of 222 (214 total MIE TTE) patients in lateral decubitus, the reported incidence of pulmonary complications was 7.7%. Palanivelu et al 3 reported 1.5% pulmonary complications in their series of 130 TTE MIE in prone position. In contrast, Hulscher et al 25 reported 57% pulmonary complications in the open TTE. Further prospective comparative studies are needed to establish the rate of reduction of pulmonary complications. In addition, faster post-operative recovery is generally measured by the length of hospital stay. Smithers et al 15 reported a significantly shorter hospital stay for the TO TTE category compared to open procedure. This suggest that thoracoscopy reduces the post-operative recovery. They also reported on a significantly shorter hospital stay for total MIE TTE when compared to TO TTE. With further implementation of minimally invasive surgery a faster post-operative recovery is therefore possible. Smithers et al 15 discontinued the total MIE approach because in the preliminary results total MIE did not seem to be superior to the TO TTE approach. However, in the total MIE group they used a right upper abdominal quadrant incision which could have influenced their preliminary results. THE was introduced by Orringer and Sloan in the seventies in order to reduce the morbidity and mortality of TTE 26. Following their initial report, Orringer et al 27 reported THE in their series of 2007 patients. However, they did not include laparoscopic THE in their series. The VU medical center reported on the outcome of 50 laparoscopic THE 4. A better outcome was observed for various outcome parameters e.g. pulmonary complications and hospital stay. No difference in pulmonary complications was observed in the study of van den Broek et al 21 in a preliminary report of laparoscopic THE. A significant shorter hospital stay was seen for laparoscopic THE compared to open in the study of Scheepers et al 22. In concordance with other laparoscopic gastro-intestinal procedures, laparoscopic THE also seems to benefit the post-operative recovery.

26 In conclusion, a faster post-operative recovery and therefore a reduction in morbidity can be achieved with MIE. Furthermore, less mortality with implementation of MIE could be realised. MIE was investigated in case-control studies and bias may have been introduced simply by the study design. Randomized trials comparing MIE with open esophagectomy are necessary in order to evaluate outcome more efficiently. This is the first systematic review including only studies directly comparing minimally invasive esophagectomy versus conventional open esophagectomy and stratifying this for operative approach. Chapter 2 42

27 Minimally invasive versus open esophagectomy 43 REFERENCES 1. Greene FL, Kercher KW, Nelson H, Teigland CM, Boller AMl. Minimal access cancer management. CA Cancer J Clin. 2007;57(5): Luketich JD, Alvelo-Rivera M, Buenaventura PO, Christie NA, McCaughan JS, Litle VR et al. Minimally invasive esophagectomy: outcomes in 222 patients. Annals of Surgery 2003;238(4): Palanivelu C, Prakash A, Senthilkumar R, Senthilanathan P, Parthasarathi R, Rajan PS et al. Minimally invasive esophagectomy: thoracoscopic mobilization of the esophagus and mediastinal lymphadenectomy in prone position--experience of 130 patients. Journal of the American College of Surgeons 2006;203(1): Scheepers JJ, Veenhof XA, van der Peet DL, van Groeningen C, Mulder C, Meijer S et al. Laparoscopic transhiatal resection for malignancies of the distal esophagus: outcome of the first 50 resected patients. Surgery 2008;143(2): Gemmill EH, McCulloch P. Systematic review of minimally invasive resection for gastro-oesophageal cancer. British Journal of Surgery 2007;94(12): Badessi F, Gusai GP, Fais D, Huscher C. [Traditional esophagectomy and esophago-gastrectomy vs. laparoscopic surgery. Evaluation and results]. [Italian]. Tumori 2003;89(4 Suppl): Caputo M, Fumagalli RU, Bona S, Rosati R, Peracchia A. Laparoscopic gastroplasty for esophagectomy. Rays 2005;30(4): Kaminski M, Majewski W, Sulikowski T, Sienko J, Romanowski M, Ostrowski M et al. Comparison of esophageal squamous cell cancer treatment results using two resection techniques: Transthoracic vs modified esophagectomy by means of thoracoscopy. [Polish]. Polski Przeglad Chirurgiczny 2003;75(10): Takemura M, Osugi H, Lee S, Taguchi S, Kaneko M, Tanaka Y et al. Validity of Thoracoscopic Esophagectomy and Lymph Node Dissection as a Radical Operation for Thoracic Esophageal Cancer Invading into the Adventitia. [Japanese]. Japanese Journal of Gastroenterological Surgery 2003;36(10): Valenti V, Fares R, Reynolds N, Cohen P, Theodore N, Martinez-Isla A. [Open and laparoscopic transhiatal oesophagectomy for cancer of the oesophagus: analysis of resection margins and lymph nodes]. [Spanish]. Chirugia Espanola 2008;83(1): Zhu CC, Chen SL, Ye MH. [Esophagectomy combined with radical lymphadenectomy by video-thoracoscopy]. Zhonghua wai ke za zhi [Chinese journal of surgery] 2005; 43: Taguchi S, Osugi H, Higashino M, Tokuhara T, Takada N, Takemura M et al. Comparison of three-field esophagectomy for esophageal cancer incorporating open or thoracoscopic thoracotomy. Surgical Endoscopy 2003;17(9): van der Peet DL, Scheepers JJG, Cuesta MA. Laparoscopic versus open transhiatal resections for esophagealcancer. Journal of Clinical Gastroenterology 2006; 40 Supplement 4:S Morris LGT, Tran TN, DeLacure MD. Early experience with minimally invasive esophagectomy in head and neck surgical patients. Otolaryngology Head and Neck Surgery 2007;137(6): Smithers BM, Gotley DC, Martin I, Thomas JM. Comparison of the outcomes between open and minimally invasive esophagectomy. Annals of Surgery 2007;245(2): Osugi H, Takemura M, Higashino M, Takada N, Lee S, Kinoshita H. A comparison of video-assisted thoracoscopic oesophagectomy and radical lymph node dissection for squamous cell cancer of the oesophagus with open operation. British Journal of Surgery 2003;90(1): Shiraishi T, Kawahara K, Shirakusa T, Yamamoto S, Maekawa T. Risk analysis in resection of thoracic esophageal cancer in the era of endoscopic surgery. Annals of Thoracic Surgery 2006;81(3): Braghetto I, Csendes A, Cardemil G, Burdiles P, Korn O, Valladares H. Open transthoracic or transhiatal esophagectomy versus minimally invasive esophagectomy in terms of morbidity, mortality and survival. Surgical Endoscopy 2006;20(11): Fabian T, Martin JT, McKelvey AA, Federico JA. Minimally invasive esophagectomy: a teaching hospital s first year experience. Diseases of the Esophagus 2008;21(3):220-5.

28 20. Bresadola V, Terrosu G, Cojutti A, Benzoni E, Baracchini E, Bresadola F. Laparoscopic versus open gastroplasty in esophagectomy for esophageal cancer: a comparative study. Surgical Laparoscopy, Endoscopy & Percutaneous Techniques 2006;16(2): van den Broek WT, Makay O, Berends FJ, Yuan JZ, Houdijk AP, Meijer S et al. Laparoscopically assisted transhiatal resection for malignancies of the distal esophagus. Surgical Endoscopy 2004;18(5): Scheepers JJ, Sietses C, Bos DG, Boelens PG, Teunissen CM, Ligthart-Melis GC et al. Immunological consequences of laparoscopic versus open transhiatal resection for malignancies of the distal esophagus and gastroesophageal junction. Digestive Surgery 2008;25(2): Bernabe KQ, Bolton JS, Richardson WS. Laparoscopic hand-assisted versus open transhiatal esophagectomy: a case-control study. Surgical Endoscopy 2005;19(3): Birkmeyer JD, Stukel TA, Siewers AE, Goodney PP, Wennberg DE, Lucas FL. Surgeon volume and operative mortality in the United States. New England Journal of Medicine 2003;349(22): Hulscher JBF, Van Sandick JW, De Boer AGEM, Wijnhoven BP, Tijssen JG, Fockens P. Extended transthoracic resection compared with limited transhiatal resection for adenocarcinoma of the esophagus. New England Journal of Medicine 2002;347(21): Orringer MB, Sloan H. Esophagectomy without thoracotomy. Journal of Thoracic & Cardiovascular Surgery 1978;76(5): Orringer MB, Marshall B, Chang AC, Lee J, Pickens A, Lau CL. Two thousand transhiatal esophagectomies: changing trends, lessons learned. Annals of Surgery 2007;246(3): Chapter 2 44

29 Chapter 3 Cervical or thoracic anastomosis after esophagectomy for cancer: a systematic review and meta-analysis Biere SSAY, Maas KW, Cuesta MA, van der Peet DL Digestive Surgery 2011; 28: 29-35

30 ABSTRACT Chapter 3 Background Cervical anastomosis and thoracic anastomosis are used for gastric tube reconstruction after esophagectomy for cancer. This systematic review was conducted in order to identify randomized trials that compare cervical with thoracic anastomosis. Methods A literature search for randomized trials was performed in the databases: Medline, Embase, and Cochrane Library. Results A total of 4 trials were included. All studies had a small sample size and were of moderate quality. One trial was excluded from the meta-analysis. The following outcomes were significantly associated with a cervical anastomosis in the meta-analysis: recurrent laryngeal nerve trauma (Odds Ratio (OR) 7.14 (95% confidence interval (c.i.) ); p=0.006) and anastomotic leakage (OR 3.43 (95% c.i ); p=0.03). None of the following outcomes were associated with the location of the anastomosis: pulmonary complications (OR 0.86 (95% c.i ); p=0.87), peri-operative mortality (OR 1.24 (95% c.i ); p=0.74), benign stricture formation (OR 0.79 (95% c.i ); p=0.79) and tumor recurrence (OR 2.01 (95% c.i ); p=0.21). Conclusion Cervical anastomosis could be associated with a higher leak rate and recurrent nerve trauma. However, the currently available randomized evidence is limited. Further randomized trials are needed to provide sufficient evidence for the preferred location of the anastomosis after esophagectomy. 46

31 Cervical or thoracic anastomosis after esophagectomy 47 INTRODUCTION In the 1940s a 72% peri-operative mortality was associated for esophagectomy for esophageal cancer 1. In order to reduce this mortality Ivor Lewis introduced a standardized approach to esophageal resection in 1946 for carcinoma of the middle third of the esophagus 2. This approach involved a two-stage procedure that included a laparotomy and 1 to 2 weeks later a thoracotomy with thoracic anastomosis. Risks of anastomotic leakage in the thorax with fatal consequences resulted in a three-stage approach with a cervical anastomosis 3. In case of a leakage a cervical fistula is a manageable complication 4. However, an increased incidence of leakage and stricture formation could be associated with a cervical anastomosis. Nowadays both thoracic anastomosis and cervical anastomosis are used worldwide for gastric tube reconstruction after esophagectomy. Advocates of the cervical anastomosis favor this location despite the possible increased incidence of leakage, benign stricture formation and damage to the recurrent laryngeal nerve because of better tumor eradication and reduced mortality and morbidity associated with anastomotic breakdown. Several randomized controlled trials (RCTs) have been published in which cervical anastomosis are compared with thoracic anastomosis. To date no systematic review and meta-analysis of literature has been performed in order to identify the randomized trials performed in the past and to designate the appropriate anastomotic approach after esophagectomy. This systematic review was conducted in order to identify randomized trials that compare cervical with thoracic anastomosis after esophagectomy for carcinoma. A critically appraisal was performed and short- and long-term outcome were assessed. METHODS Literature search The following electronic databases were used for a literature search: Medline (1950 to March 2010), Embase (1947 to March 2010), and the Cochrane Library (2010 issue 1). A comprehensive search was performed using the following search terms: esophageal cancer, cervical, thoracic, anastomosis. Logical combinations of these and related terms (e.g. oesophagus, carcinoma, intrathoracic) were used to maximize sensitivity. Furthermore, a truncation symbol was used in each database in order to allow retrieval of all suffix variations of a root word (e.g. anastomos*). A manual cross-reference search of the bibliographies of relevant articles was conducted to identify studies not found through the computerized search. The related articles feature of Pubmed was also used. Two reviewers executed the search independently (SSAYB and KWM). Study selection criteria After identifying relevant titles, the abstracts of these studies were read to decide if the study was eligible. The full article was retrieved when the information in the title and/ or abstract appeared

32 to meet the objective of this review. All published randomized trials comparing cervical with thoracic anastomosis after esophagectomy for cancer were included. There were no restrictions with regard to the language of the published study. Quality assessment of retrieved articles The two reviewers appraised each included article independently, assessing the methodological quality of the selected studies. A modified critical review checklist of the Dutch Cochrane Centre was used to appraise RCTs [ Last accessed November 2009]. The following items of the checklist are negative for all surgical studies and were therefore not included for further appraisal (items: 1. Patient blinded? 2. Care provider blinded?). Data extraction and analysis Data was extracted on preformatted sheets. Two reviewers also performed this independently. The studies were tabulated and methodologically evaluated to assess homogeneity. In the case of heterogeneity between the studies, it would not be justified to pool the assessed outcomes. The following outcome parameters were assessed: pulmonary complications, anastomotic leakage, mortality, recurrent nerve trauma, positive resection margin (R1), length of hospital stay, stricture of anastomosis, tumor recurrence, pre-defined quality of life assessment. Chapter 3 48 Statistical analysis Meta-analyses were performed for outcome parameters when possible. Quantative data for the primary outcome parameters were entered into the software Cochrane Review Manager (RevMan) version 4.3 and analyzed using RevMan Analyses (Cochrane Collaboration, Oxford, UK). Summary estimates, including 95% confidence intervals, were calculated. For continuous outcome data (e.g. length of hospital stay) means and standard deviations were used to calculate a weighted mean difference in the meta-analysis. For dichotomous outcomes (e.g. mortality) the odds ratio (OR) was calculated with corresponding confidence interval (c.i.). Statistical heterogeneity was tested using X 2 and I 2 tests. Data were pooled using the random effect model, because moderate heterogeneity was suspected (e.g. stapler devices). RESULTS The initial search yielded 224 potential literature citations (figure 1). Of these, 218 were excluded after scanning the title. Of the six studies considered potentially relevant, 2 were excluded. One study was excluded which compared manual anastomosis with mechanical anastomosis 5. Another study was identified as a duplicate report in a non-english language journal 6. A total of 4 randomized trials were included for the systematic review (table 1) These studies involved 267 patients, of which 132 patients randomized for a cervical anastomosis and 135 patients had a thoracic anastomosis (Table 1).

33 Cervical or thoracic anastomosis after esophagectomy 49 Figure 1. Identification of eligible randomized controlled trials. Search strategy 224 references 218 considered not relevant based on title 6 studies selected on title 1 manual vs. mechanical anastomosis comparison 1 duplicate study 4 full-text articles included in Systematic review Table 1. Characteristics of the included randomized trials. Reference Year Journal Single center or multicenter trial Surgical approach Type of anastomosis Cervical group Thoracic group Number of patients included Cervical group Thoracic group 1989 Surgery, Gynaecology 1992 Journal of Thoracic Chasseray Single center Transthoracic Stapled* Stapled et al. 10 and Obstetrics Ribet Single center Transthoracic Hand-Sewn Stapled et al. 8 and Cardiovascular Surgery Walther 2003 Annals of Surgery Single center Transthoracic Hand-Sewn Stapled et al. 7 Okuyama et al Surgery Today Single center Transthoracic Hand-Sewn Stapled * Seven patients in the cervical group had a hand-sewn anastomosis. Quality assessment The quality assessment of the included studies is shown in table 2. Chasseray et al. excluded 31 patients because a transhiatal approach was subsequently chosen for some patients, there was a nonresectable tumor or because the surgeon did not want to place the gastric tube in the posterior mediastinum 10. Furthermore, concealment of allocation was not reported in this study. In the cervical group 8 patients had a retrosternal positioning of the gastric tube, in 35 patients the gastric tube was placed in the posterior mediastinum. Importantly only 7 of 43 patients in the cervical anastomosis group had a hand-sewn anastomosis.

34 Table 2. Methodological quality of the included randomized trials. Reference Excluded patients Treatment allocation concealed? Eligibility criteria specified? Groups Follow-up? similar at baseline? Intention to treat? Chasseray et al Not stated Yes Yes Yes Yes Yes Ribet et al. 8 2 Yes Yes Yes Yes No Yes Walther et al Yes Yes Yes Yes Yes Yes Okuyama et al. 9 Not stated Yes Yes Yes Yes Yes Yes Similar non trial treatment? Chapter 3 50 Ribet et al. excluded 2 patients in the thoracic anastomosis group because a thoracic anastomosis could not be made 8. These patients were excluded from the analysis; therefore no intention to treat analysis was performed. Walther et al. excluded 47 patients, 17 patients had a nonresectable tumor, of which 3 underwent a bypass procedure 7. Another 8 patients had tumor located high in the esophagus, 8 patients had a jejunal or colon interposition, 6 patients refused participation, 4 had an palliative resection, 2 patients had a redo esophagectomy and 1 patients was operated in another hospital. The results of the last 29 excluded patients were reported separately in this study. Patients with benign disease were also randomized and analysed (2 patients in the cervical group and 7 patients in the thoracic group). Okuyama et al. did report the exclusion of patients. Eligibility was however described in this study 9. Overall, the quality of the included studies is moderate. Chasseray et al. only performed a hand-sewn cervical anastomosis in 7 of the 43 patients; this study was therefore excluded from the meta-analysis. The outcome parameters are depicted in table 3. The meta-analyses are shown in table 4. Outcome parameters Pulmonary complications The location of the anastomosis did not influence the incidence of pulmonary complications defined as pulmonary infection (OR 0.86 (95% c.i ); p=0.87). Chasseray et al. reported a 16% (n=7) incidence of respiratory complications in the cervical group. In the thoracic anastomosis group the incidence was 29% (n=15). This difference in incidence did not reach a statistical significance. Anastomosis leakage A significant difference in incidence of anastomotic leakage is seen when comparing cervical anastomosis with thoracic anastomosis (OR 3.43 (95% c.i ); p=0.03). Chasseray et al. reported significant more anastomotic leaks in the cervical group (26% vs. 4%; p<0.02).

35 Cervical or thoracic anastomosis after esophagectomy 51 Table 3. Outcome parameters. Reference Pulmonary complications Cervical group Thoracic group Anastomosis Leakage Cervical group Thoracic group Cervical group Mortality Thoracic group Positive resection margin Cervical group Thoracic group Chasseray et al (16%) 15 (29%) 11 (26%) 2 (4%) 4 (9%) 7 (14%) 1 (2%) 1 (2%) Ribet et al (70%) 11 (37%) 9 (30%) 3 (10%) 5 (17%) 4 (13%) 3 (10%) 10 (33%) Walther et al. 7 2 (5%) 4 (10%) 1 (2%) 0 (0%) 1 (2%) 1 (2%) 0 (0%) 0 (0%) Okuyama et al. 9 2 (11%) 5 (36%) 3 (17%) 1 (7%) 0 (0%) 0 (0%) Not stated Not stated Table 3. continued. Reference Chasseray et al. 10 Recurrent nerve trauma Cervical group Not stated Ribet 6 et al. 8 (20%) Walther 1 et al. 7 (2%) Okuyama 8 et al. 9 (39%) Thoracic group Not stated 1 (3%) 0 (0%) 1 (7%) Hospital stay (days) Cervical group 19.5 (3-71)* Thoracic group 18 (2-122)* Stricture of anastomosis Cervical group 10 (23%) (13%) 14 (6-68)* Not stated 14 (0-83)* Not stated 8 (20%) 0 (0%) Thoracic group 7 (14%) 1 (3%) 12 (29%) 2 (14%) Tumor recurrence Cervical group 0 (0%) Thoracic group 0 (0%) Quality of Life assessment Cervical group Not assessed 9 5 Not (60%) (36%) assessed 1 (2%) 4 (22%) 1 (2%) 2 (14%) Not assessed Thoracic group Not assessed Not assessed Not assessed No No significant significant difference difference * reported in median; reported in mean, no standard of mean was reported; follow-up data only known for 29 patients Table 4. Meta-analysis of outcome parameters. Pulmonary complications Anastomotic leakage Mortality Recurrent nerve trauma Odds Ratio (95% confidence interval) P Odds Ratio (95% confidence interval) P Odds Ratio (95% confidence interval) P Odds Ratio (95% confidence interval) P Meta-analysis 0.86 ( ) ( ( ) ( ) Table 4 continued. Meta-analysis of outcome parameters. Stricture of anastomosis Tumor recurrence Odds Ratio (95% confidence interval) P Odds Ratio (95% confidence interval) P Meta-analysis 0.79 ( ) ( ) 0.21

36 Mortality Situation of a cervical anastomosis was not associated with a lower mortality (OR 1.24 (95% c.i ); p=0.74). Okuyama et al. did not observe any peri-operative mortality. Chasseray et al. reported no significant difference in mortality between cervical and thoracic anastomosis groups (9% (n=4) vs. 14% (n=7)). Positive resection margin Only 2 of 3 RCTs eligible for the meta-analysis investigated the resection margins. One study did not observe any positive resection margin and the other study observed more positive margins in the thoracic group (10% (n=3) vs. 33% (n=10)). Okuyama et al. did not report the status of the resection margin. Chasseray et al. reported 1 positive resection margin in each anastomotic group. Recurrent nerve trauma A cervical anastomosis was associated with significant more recurrent nerve trauma (OR 7.14 (95% c.i ); p=0.006). Chasseray et al. did not report on recurrent nerve trauma in their RCT. Chapter 3 52 Hospital stay Two of the 3 RCTs eligible for the meta-analysis reported length of hospital stay. Ribet et al. calculated a mean hospital stay of 24.2 days for the cervical anastomosis and 16.6 days for thoracic anastomosis. Walther et al. calculated median hospital stay which was not significant different between the two groups (14 vs. 14 days). Benign stricture of anastomosis requiring dilatation A cervical anastomosis was not associated with a significant higher incidence of stricture of the anastomosis requiring dilatation (OR 0.79 (95% c.i ); p=0.79). Chasseray et al. observed no significant difference between the cervical group compared with the thoracic group (23% (n=10) vs. 14% (n=7)). Tumor recurrence The location of the anastomosis did not significantly influence the incidence of tumor recurrence (OR 2.01 (95% c.i ); p=0.21). Quality of life Only Okuyama et al. investigated the quality of life at 6 months after surgery. They assessed heartburn, regurgitation, stenotic sense, abdominal fullness, cough and wound pain. There was no significant difference between the cervical and thoracic anastomosis groups.

37 Cervical or thoracic anastomosis after esophagectomy 53 DISCUSSION This systematic review aimed to compare cervical anastomosis with thoracic anastomosis after esophageal resection for cancer. Four randomized trials were identified of which 3 were eligible for the meta-analysis. The small sample sizes and moderate quality of the included studies brings limitations for the meta-analysis. A cervical anastomosis was associated with significant more recurrent nerve trauma and anastomosis leakage. No significant difference was observed in pulmonary complications, peri-operative mortality, benign strictures of the anastomosis requiring dilatation and tumor recurrence. There is insufficient data in this meta-analysis to report about the involvement of the resection margin, hospital stay and the quality of life after a cervical or thoracic anastomosis. The traditional perception is that a thoracic anastomosis is associated with less frequent leaks, but if occurring have fatal consequences for the patient 11. Although the confidence interval of this meta-analysis almost reaches 1, a significant higher incidence of anastomotic leaks is seen when comparing cervical with thoracic anastomosis after esophagectomy. The location of the anastomosis could be a risk factor for leakage. Cervical anastomoses could suffer more from ischemia of the gastric conduit than thoracic anastomoses 13. Other possible factors that could contribute to leakage could be of technical nature, related to gastric drainage or low institution and surgeon volume 14,15. Walther and coworkers performed a well designed single center randomized trial analysing both included patients and excluded patients 7. No significant difference in leak rate was identified between the groups. Futhermore, the higher leak rate for cervical anastomoses are generally more accepted because of the lower risk for the patient. The significant higher leak rate for cervical anastomosis did not result in a significant higher mortality in the meta-analysis. An improved peri-operative care could account for the absence of the direct relationship between anastomosis leakage and mortality in this analysis 16,17. Okuyuma and coworkers did not even observe any mortality in their trial 9. Most surgical groups in the world use a handsewn technique for cervical anastomosis and a mechanical technique for thoracic anastomosis. There seems to be no direct relationship with regard to the influence of the applied technique to outcome 18. However, in order to minimize potential bias only studies comparing handsewn cervical which compared with mechanical anastomosis were included in this systematic review. A higher incidence of recurrent laryngeal nerve trauma is associated with cervical anastomosis in this meta-analysis. A narrow anatomical relationship exists between the recurrent nerve and the esophagus at cervical level. Careful dissection and active identification of the recurrent nerve could minimize trauma to the nerve 19,20. Furthermore, an under-appreciation of recurrent nerve trauma exists in literature, because not all patients with this trauma have symptoms of hoarseness and inter-observer variation exists in diagnosis 19. None of the included trials indicated if a standard laryngoscopy or laryngoscopy on indication was performed post-operatively for recurrent nerve trauma. In literature, recurrent nerve trauma is associated with pulmonary complications. This is

38 Chapter 3 54 thought to be caused mainly by aspiration 19,21. In this meta-analysis no significant difference in pulmonary complications is seen when comparing cervical anastomosis with thoracic anastomosis. Other factors could be of more influence in the incidence of pulmonary complications, e.g. adequate post-operative analgesics and physiotherapy. Only 2 studies eligible for the meta-analysis reported on the involvement of the resection margin 7.8. Walther and coworkers did not observe any involvement of the resection margins in both cervical and thoracic anastomosis groups 7. Ribet and coworkers observed more positive resection margins in the thoracic group. Adequate pre-operative and per-operative assessment probably improves the low incidence of positive resection margins 22,23. These assessments also influence the tumor recurrence. No significant difference in tumor recurrence was seen in the meta-analysis. It is therefore probable that the involvement of the resection is comparable between the cervical and the thoracic anastomosis groups. There are some reports that associate stricture formation with the use of stapler devices 24,25. There is also evidence that anastomotic stricture is associated with anastomotic leakage 25,26. Although there is a significant higher incidence of anastomosis leakage with cervical anastomosis, benign stricture formation requiring dilatation is not associated with the location of the anastomosis in this meta-analysis. Other factors such as reflux could have more impact on stricture formation. The possible role of proton pump inhibitors (PPIs) in reducing reflux and therefore benign stricture formation has recently been investigated in a RCT 27. A lower prevalence of benign stricture formation was observed in the group using PPIs. The present systematic review has several limitations. There were only 4 RCTs selected for inclusion. These included studies were performed with a relative small sample size. Only three trials were included for the meta-analysis as Chasseray et al. used a stapled anastomosis in the cervical group. Two of these 3 trials were performed in the last 2 decades 7,9. The other trial was conducted from 1983 to Changes in techniques and peri-operative care can influence outcome. A bias due to change patterns of care, surgical techniques and improved suture devices could be present in this meta-analysis. RCTs from different time periods and continents could include some degree of heterogeneity in clinical approach. To address this issue a random-effect model was used to meta-analyse the data. Furthermore, one should be aware that wide confidence intervals for some comparisons do not completely rule out the possibility of big increases in risk, or big decreases in risk. With regard to exclusion of patients, Chasseray et al. excluded 31 patients for debatable reasons. Okuyama et al. did not even report the exclusion of patients. A selection bias could be possible in these studies. Also, surgical and/ or hospital volume is well known to influence outcome. None of the included trials reported their yearly volume. Several of the aforementioned limitations of the trials have been recently addressed in literature 28. This systematic review shows that there are only 4 single center randomized trials available in literature with small sample sizes and moderate quality. In the meta-analysis of 3 trials no difference in pulmonary complications, peri-operative mortality, benign stricture formation and tumor recurrence after esophagectomy with a cervical or thoracic anastomosis for esophageal cancer is

39 Cervical or thoracic anastomosis after esophagectomy 55 seen. A cervical anastomosis could however be associated with a higher incidence of recurrent laryngeal nerve trauma and anastomotic leakage. Further multicenter randomized trials in current clinical settings are needed to provide sufficient evidence for the preferred location of the anastomosis after esophagectomy for esophageal cancer. New anastomosis techniques could also be of interest for future randomized trials 29.

40 REFERENCES Chapter Oschner A, DeBakey M. Surgical aspects of carcinoma of the esophagus. Journal of Thoracic Surgery 1941; 10: Lewis I. The surgical treatment of carcinoma of the oesophagus: with special reference to a new operation for growths of the middle third. British Journal of Surgery 1946; 34: McKeown KC. Total three-stage oesophagectomy for cancer of the oesophagus. British Journal of Surgery 1976; 63: Scheepers JJ, van der Peet DL, Veenhof AA, Heijnen B, Cuesta MA. Systematic approach of postoperative gastric conduit complications after esophageal resection. Diseases of the Esophagus 2009; 23: Valverde A, Hay JM, Fingerhut A, Elhadad A. Manual versus mechanical esophagogastric anastomosis after resection for carcinoma: a controlled trial. French Associations for Surgical Research. Surgery 1996; 120: Ribet M, Debrueres B, Lecomte HM. [Esophagectomy for advanced malpighian cancer of the thoracic esophagus. Esogastric anastomosis in the neck or in the thorax? Late results of a randomized prospective study]. Annales de chirurgie 1992; 46: Walther BJ, Johansson J, Johnsson F, Von Holstein CS, Zilling T. Cervical or Thoracic Anastomosis after Esophageal Resection and Gastric Tube Reconstruction: A Prospective Randomized Trial Comparing Sutured Neck Anastomosis with Stapled Intrathoracic Anastomosis. Annals of Surgery 2003; 238: Ribet M, Debrueres B, Lecomte HM. Resection for advanced cancer of the thoracic esophagus: cervical or thoracic anastomosis? Late results of a prospective randomized study. The Journal of Thoracic and Cardiovascular Surgery 1992; 103: Okuyama M, Motoyama S, Suzuki H et al. Hand-sewn cervical anastomosis versus stapled intrathoracic anastomosis after esophagectomy for middle or lower thoracic esophageal cancer: a prospective randomized controlled study. Surgery Today 2007; 37: Chasseray VM, Kiroff GK, Buard JL et al. Cervical or thoracic anastomosis for esophagectomy for carcinoma. Surgery, gynecology & obstetrics 1989; 169: Urschel JD. Esophagogastrostomy anastomotic leaks complicating esophagectomy: a review. American Journal of Surgery 1995; 169: Holscher AH, Vallbohmer D, Brabender J. The prevention and management of perioperative complications. Best practice & research. Clinical gastroenterology 2006; 20: Reavis KM. The esophageal anastomosis: how improving blood supply affects leak rate. Journal of Gastrointestinal Surgery 2009; 13: Whooley BP, Law S, Alexandrou A, Murthy SC, Wong J. Critical appraisal of the significance of intrathoracic anastomotic leakage after esophagectomy for cancer. American Journal of Surgery 2001; 181: Junemann-Ramirez M, Awan MY, Khan ZM, Rahamim JS. Anastomotic leakage post-esophagogastrectomy for esophageal carcinoma: retrospective analysis of predictive factors, management and influence on longterm survival in a high volume centre. European Journal of Cardiothoracic Surgery 2005; 27: Iscimen R, Brown DR, Cassivi SD, Keegan MT. Intensive Care Unit Utilization and Outcome After Esophagectomy. Journal of Cardiothoracic and Vascular Anesthesia May 6. [Epub ahead of print] 17. Ruol A, Castoro C, Portale G, Cavallin F, Sileni VC, Cagol M et al. Trends in management and prognosis for esophageal cancer surgery: twenty-five years of experience at a single institution. Archives of Surgery 2009; 144: Urschel JD, Blewett CJ, Bennett WF, Miller JD, Young JE. Handsewn or stapled esophagogastric anastomoses after esophagectomy for cancer: meta-analysis of randomized controlled trials. Diseases of the Esophagus 2001; 14: Force S. The innocent bystander complications following esophagectomy: atrial fibrillation, recurrent laryngeal nerve injury, chylothorax, and pulmonary complications. Seminars in Thoracic and Cardiovascular Surgery 2004; 16:

41 Cervical or thoracic anastomosis after esophagectomy Wright CD, Zeitels SM. Recurrent laryngeal nerve injuries after esophagectomy. Thoracic Surgery Clinics 2006; 16: Gockel I, Kneist W, Keilmann A, Junginger T. Recurrent laryngeal nerve paralysis (RLNP) following esophagectomy for carcinoma. European Journal of Surgical Oncology 2005; 31: McManus K, Anikin V, McGuigan J. Total thoracic oesophagectomy for oesophageal carcinoma: has it been worth it? European Journal of Cardiothoracic Surgery 1999; 16: Blewett CJ, Miller JD, Young JE, Bennett WF, Urschel JD. Anastomotic leaks after esophagectomy for esophageal cancer: a comparison of thoracic and cervical anastomoses. Annals of Thoracic and Cardiovascular Surgery 2001; 7: Rice TW. Anastomotic stricture complicating esophagectomy. Thoracic Surgery Clinics 2006; 16: Dresener SM, Lamb PJ, Wayman J, Hayes N, Griffin SM. Benign anastomotic stricture following transthoracic subtotal oesophagectomy and stapled oesophago-gastrostomy: risk factors and management. British Journal of Surgery 2000; 87: Honkoop P, Siersema PD, Tilanus HW, Stassen LP, Hop WC, van Blankenstein M. Benign anastomotic strictures after transhiatal esophagectomy and cervical esophagogastrostomy: risk factors and management. Journal of Thoracic and Cardiovascular Surgery 1996; 111: Johansson J, Oberg S, Wenner J, Zilling T, Johnsson F, von Holstein CS, Walther B. Impact of proton pump inhibitors on benign anastomotic stricture formations after esophagectomy and gastric tube reconstruction: results from a randomized clinical trial. Annals of Surgery 2009; 250: Lagarde SM, Vrouenraets BC, Stassen LP, van Lanschot JJ. Evidence based surgical treatment of esophageal cancer: overview of high quality studies. Annals of Thoracic Surgery 2010; 89: Collard JM, Romagnoli R, Goncette L, Otte JB, Kestens PJ. Terminalized semimechanical side-to-side suture technique for cervical esophagogastrostomy. Annals of Thoracic Surgery. 1998; 65:

42

43 Chapter 4 Intrathoracic anastomosis after minimally invasive Ivor Lewis esophagectomy for esophageal cancer Maas KW, Biere SSAY, Scheepers JJG, Gisbertz SS, Turrado Rodriguez V, van der Peet DL, Cuesta MA [Submitted]

44 ABSTRACT Chapter 4 Background Minimally invasive Ivor Lewis esophagectomy is one of the approaches used worldwide for treating esophageal cancer. Optimization of this approach and especially identifying the ideal intrathoracic anastomosis technique is needed. To date, several minimally invasive Ivor Lewis anastomosis techniques have been described. A literature search of the current techniques and approaches for intrathoracic anastomosis was performed. Methods The PubMed electronic database was used for a comprehensive literature search by two independent reviewers. The included studies were evaluated on leakage and stenosis rate of the anastomosis. Results Twelve studies were included in this review. The most frequent applied technique was the stapled anastomosis. Stapled anastomoses can be achieved either through a transthoracic or a transoral introduction. This stapled approach can be performed with a circular or linear stapler. The reported anastomotic leakage rate ranges from 0% to 10%. The reported anastomotic stenosis rate ranges from 0% to 27.5%. Conclusion The most frequently used anastomosis is a stapled one: with the anvil introduced either transorally or transthoracically. Clinical trials are needed to compare different methods in order to improve the quality of the intrathoracic anastomosis after a minimally invasive Ivor Lewis esophagectomy. 60

45 Intrathoracic minimally invasive Ivor Lewis esophagectomy 61 INTRODUCTION Despite the recent developments of the minimally invasive approach for esophageal cancer, esophagectomy is still associated with a significant risk of peri-operative morbidity and mortality. 1 After a successful performed esophageal resection, the creation of a safe anastomosis is essential in order to reduce the risk of leakage and related complications. We see the resurgence of an important trend to anastomose the gastric tube with the intrathoracic esophagus, the so-called Ivor Lewis operation. 2 In the past, risk of anastomotic leakage in the thorax with possibly fatal consequence resulted in a three-stage approach with a cervical anastomosis as developed by McKeown, 3 and the transhiatal approach with a cervical anastomosis by Orringer. 4 In case of anastomotic leakage in the neck, a subsequent cervical fistula is a manageable complication. 5 There is however, some evidence that cervical anastomosis could be related to more anastomotic leakage. 6 Currently, the increased numbers of gastro-esophageal junction tumors may form an ideal indication for an intrathoracic anastomosis. This could also reduce recurrent nerve lesion associated with a cervical dissection. Moreover, a shorter gastric conduit will permit a more extended gastric resection and will, because its good vascular supply, theoretically lead to fewer anastomotic leakages. This transthoracic procedure may be performed by a conventional or minimally invasive approach. In order to achieve a perfect anastomosis, in the case of minimally invasive esophagectomy, this approach demands high skills and standards of the surgeon. The question is which type of intrathoracic anastomosis is the ideal one. According to the paper by Blackmon et al., stapled anastomosis in conventional surgery will have fewer leakages and stenosis than the manually performed one. 7 We performed a literature review of all techniques used for an intrathoracic anastomosis in minimally invasive Ivor Lewis esophagectomy in order to identify and standardize the technique. METHODS The PubMed electronic database was used for literature search. A comprehensive search was performed using the following terms: esophagus, esophagectomy, anastomosis, intrathoracic, cancer, Ivor Lewis, minimally invasive, MIE, laparoscopy, thoracoscopy, esophagogastric. Related terms and combinations were also used (e.g. thoracic, oesophagectomy). Relevant titles were identified and abstracts were read in order to decide eligibility. When the information in the title and abstract met the objectives of this review, the full article was read. A manual cross-reference search of the bibliographies of the relevant articles was carried out in order to identify other studies. Furthermore, the related articles feature of PubMed was used. Two reviewers (KWM and SSAYB) executed the search independently of each other.

46 RESULTS Twelve studies were included in this review of the different techniques. Tables 1 and 2 depict the study characteristics and the anastomotic outcomes. Intrathoracic anastomosis can be accomplished by a hand sewn technique or a stapler technique (figure 1). Chapter 4 62 Table 1. Study characteristics Study No patients Surgical approach Watson et al Cadiere et al Lee et al Nguyen et al Misawa et al Bizekis et al Thairu et al Sutton et al Nguyen et al Campos et al Ben-David et al Gorenstein et al Patient position Anastomotic technique Type of anastomosis 2 Transthoracic Prone position Handsewn End-to-side 1 Transthoracic Prone position Handsewn Side-to-end 8 Transhiatal and transthoracic 1 Transthoracic 5 Transthoracic 50 Transthoracic 45 Semi-lateral position Left lateral decubitus position Left lateral decubitus position Left lateral decubitus position 18 Transthoracic Prone position 10 Transhiatal Supine position 51 Transthoracic Left lateral decubitus position 37 Transthoracic NR 6 Transthoracic 31 Transthoracic NA, not applicable. NR, not reported. Left lateral decubitus position Left lateral decubitus position Circular stapled Circular stapled Circular stapled Circular stapled Circular stapled Transorally circular stapled Transorally circular stapled Transorally circular stapled Linear stapled Linear stapled Layers Single layer Two layers Fixation of anvil NA NA End-to-side NA Purse-string End-to-side End-to-side NA NA Handsewn purse-string Purse-string Endo-stitch End-to-side NA NR End-to-side NA Linear staple gun and Z-stitch End-to-side NA NR End-to-side NA NR End-to-side NA NR Side-to-side NA NR Side-to-side NA NR

47 Intrathoracic minimally invasive Ivor Lewis esophagectomy 63 Table 2. Anastomotic outcome Study No. patients Anastomotic leak Anastomotic stenosis Watson et al Cadiere et al Lee et al (12.5%) Nguyen et al Misawa et al Bizekis et al (6%) 6 (12%) Thairu et al Not reported Sutton et al (10%) Not reported Nguyen et al (9.8%) 14 (27.5%) Campos et al (2.7%) 5 (13.5%) Ben-David et al Not reported 0 Gorenstein et al (3.2%) Not reported Figure 1. Overview of minimally invasive anastomosis techniques MIE, minimally invasive esophagectomy.

48 Hand sewn technique The first description of a totally endoscopic Ivor Lewis esophagectomy with an intrathoracic anastomosis was reported in 1999 by Watson et al. 8 The thoracoscopic phase of the operation was performed in prone position. They described two patients in which the intrathoracic anastomosis was achieved with a hand sewn technique. Anastomosis between the upper esophagus and the gastric tube was performed by a single layer suture. Posteriorly the seromuscular gastric layer was sutured to the full thickness of the posterior wall of the esophagus and anteriorly full thickness of the stomach with the esophageal wall. Both patients recovered without complications and with a short total and hospital stay. Ten years later, Cadiere et al. described a total minimally invasive Ivor Lewis esophagectomy with a hand sewn anastomosis. 9 The patient was operated through a right thoracotomy in prone position. No post-operative complications were reported in this case report. An illustration of the hand sewn intrathoracic anastomosis technique is shown in figure 2. Figure 2. Handsewn intrathoracic anastomosis (for color figure, see color section, page 179) Chapter 4 64 Stapled technique Transthoracic circular stapled anastomosis In 1997, Lee et al. described a right lateral thoracoscopic esophagectomy with an intrathoracic stapled anastomosis in a series of eight patients with carcinoma of the lower esophagus 10.The operation was carried out with a double-lumen endotracheal tube to enable collapse of the right lung. The patient was placed in 45 degrees semilateral position with the right side up. A two-team synchronous approach was used, one for the abdominal stage to perform the mobilization of the stomach through a transverse incision laparotomy, whereas the other team performed the thoracoscopic approach. Esophagogastric anastomosis was fashioned by a stapling device using the ligature method described by Allsop and Ng. 11,12 A right-angled clamp inserted by the abdominal surgeon through the hiatus was applied onto the esophagus just proximal to the lesion. A transverse incision was placed above the tumor in the esophagus with adequate margin. The abdominal surgeon inserted the anvil (of the 28 mm circular stapler) into the thoracic cavity (through the hiatus) and the thoracoscopic surgeon manipulated it into the lumen of the esophagus thoracoscopically. A Vicryl ligature was applied around the esophagus by means of Roeder s knot to fix

49 Intrathoracic minimally invasive Ivor Lewis esophagectomy 65 Figure 3. Fixation of the anvil by a Roeder s Knot (for color figure, see color section, page 179) the anvil in the proper position and the esophagus divided as illustrated in figure 3. The specimen was delivered into the abdominal cavity and the cardio-esophageal junction was transected with a GIA linear stapler. A transverse gastrotomy was performed of the distal part of the stomach and the circular stapler was inserted and the stomach introduced into the thorax and the anastomosis performed. The operation was successful in seven patients, the remaining patient required conversion to thoracotomy. No leakages were recorded. One patient died from an intracranial bleeding. On short term, another patient developed a benign stenosis requiring dilatation. The anastomosis technique described by Nguyen et al. in 2001 in a case report was performed through a right lateral thoracoscopy with lung block after the laparoscopic phase. 13 The esophagus was divided with the endoscopic stapler 2 cm below the azygos vein. The esophageal specimen was detached from the gastric conduit and retrieved through an enlarged (4 cm) posterior trocar site. Using this small incision the anvil of 21 or 25 circular stapler was placed into the chest and inserted through the opening of the esophagus and secured with a hand sewn purse-string suture. The circular stapler was introduced into the chest and passed through an anterior gastrostomy in the gastric conduit. A stapled end-to-side esophagogastric anastomosis was made. The anastomosis was reinforced with a second layer of seromuscular interrupted sutures. The gastric conduit was sutured to the crura of the diaphragm to close the diaphragmatic defect. Most of the techniques described for circular stapled anastomosis are similar to this technique. 14

50 Figure 4. Fixation of the anvil by purse-string using the Endostitch (for color figure, see color section, page 180) Chapter 4 66 Other techniques, which tried to improve the placement of the anvil in the proximal esophagus, have also been described. Misawa et al. (2005) approached the esophagus, after the laparoscopic phase, through a lateral right thoracoscopy in five patients with cancer of the middle and lower thoracic esophagus. 15 A 5 cm long thoracotomy was made in the fifth intercostal space and two more trocars were inserted. After mobilization of the esophagus, the level of transection was designated and the distal esophagus tied with a ligature string. The esophageal wall was incised anteriorly; one sixth of the wall was left. Two Allis clamps grasped the esophageal edges allowing the Endo-Stitch (US Surgical) device for achieving a purse-string suture. After this, the anvil of a 25 mm circular stapler was introduced into the proximal esophagus and the purse-string tied. An Endoloop ligature was placed over the purse-string for preserving the edges of the esophagus from falling out of the purse-string. This is illustrated in figure 4. After resection of the distal esophagus including the lesion, the proximal end of the gastric tube was pulled out of the minithoracotomy and the circular stapler was introduced through it by a gastrotomy incision. After completion of the end-to-side anastomosis, the rest of the gastric conduit was resected using an endolinear stapler. No leakage or any other post-operative complications were observed in this procedure. Bizekis et al. (2006) reported their experience with minimally invasive Ivor Lewis esophagectomy in 50 patients, the majority with tumors with extension into the cardia or a gastro-esophageal junction tumor. 16 Thirty-five patients underwent a hybrid Ivor Lewis procedure, defined as laparoscopic gastric mobilization combined with a 5 cm mini-thoracotomy and the other fifteen underwent the total minimally invasive approach (laparoscopy and thoracoscopy). In the total minimally invasive procedure the esophagus was mobilized by a lateral thoracoscopic approach with a collapsed lung up to 4 to 5 cm above the azygos vein. The esophagus was transected 2 3 cm

51 Intrathoracic minimally invasive Ivor Lewis esophagectomy 67 above the azygos vein and the specimen retrieved through an enlarged (3 4 cm) inferior intercostal port site to allow the introduction of the circular stapler and for retrieval of the specimen. The 25 mm anvil was placed into the proximal esophagus and it was secured using a purse-string suture by means of Endo-stitch (US Surgical). The stapler was introduced into the gastric conduit and a circular anastomosis was created in an end-to-side fashion. The redundant gastric conduit was removed using a linear stapler. In the hybrid procedure, after the laparoscopic phase, the patient was placed in the lateral position and a 5 cm mini-thoracotomy performed, dividing a portion of the latissimus and sparing the serratus muscle. A rib retractor was placed and the thoracic esophagus mobilized under direct vision. A stapled esophagogastric anastomosis was created in a similar fashion. There were 3 anastomotic leaks in the hybrid group, one empyema not related to leak, one chylothorax and one pulmonary embolism. They observed a 6% mortality rate. Six patients developed stenosis of the anastomosis, defined as the need for more than one dilatation post-operatively. To obviate the necessity of the placement of a purse string suture or a manually tied knot to secure the anvil, Thairu et al. (2007) described a technique in which the head of the anvil was inserted through the anterior wall of the esophagus, which was opened with dissecting scissors. 17 A linear staple gun was fired at 60º to the longitudinal axis first right and after left to the spike of the anvil thus forming a V. Around this aperture a Z-stitch was placed, which secured the anvil in place, followed by an end-to-side anastomosis. Using this technique in 18 patients, they observed no anastomotic leaks, indicating that the technique achieves a safe anastomosis. No mid-term results were reported (e.g. stenosis). Transoral circular stapled anastomosis An important development is introducing the anvil transorally into the proximal esophagus, as described originally by Wittgrove et al. for the gastrojejunostomy-construction of the gastric bypass in morbid obesity surgery. 18 In order to avoid the complications related with thoracotomy or thoracoscopy, Sutton et al. reported in 2002 a technique that allowed laparoscopic visualization of a transhiatal approach to the esophagus. 19 The esophagus is transected above the tumor with a linear Endo-GIA 60 mm stapler placed through the incision. After resection, a gastric tube was fashioned from the remaining stomach. The anvil was modified removing the spring from the underside of the platform and leaving a hook unto which a length of suture was secured. The top of the anvil was disengaged from the spike so that it fell into a flip position. This position was then maintained by tension on the suture, which was secured to a nasogastric tube. Under direct vision by an introduced endoscope, a rigid guide wire was introduced through the staple line to the abdominal operator. The delivery device was passed over the guide wire orally. Gentle traction by the abdominal operator on the delivery system brought the anvil through the mouth of the esophagus until the spike of the anvil protruded through the staple line. The anvil was positioned into an un-flipped position and the anastomosis completed. Ten patients were operated on in this

52 Chapter 4 way. One patient suffering from anastomotic leak was treated successfully with a self-expanding endoscopic stent. Three patients required esophageal dilatation after discharge. Nguyen et al. (2008) described this technique in a series of 10 patients. 20 A prepared pre-tilted anvil s head tip was attached to an oral-gastric tube that was given to the anaesthesiologist. The tilted configuration of the anvil improved the ease of transoral passage by the anaesthesiologist. After the transection of the esophagus with a linear stapler, the tube was passed transorally until it was felt within the proximal esophageal stump. A small opening at this level was made perpendicular to the staple line of the stump and the tube was advanced through it and was withdrawn until the anvil was in the right position at the end of the esophageal stump. The oral-gastric tube was removed after cutting the suture that attached it to the anvil. The head of the anvil returned to the flat position when attached to the spike of the 25 mm circular stapler in order to perform the anastomosis. The authors applied the technique without passage problems in ten patients. Moreover, tissue donuts were complete in all cases and there were no post-operative leaks. Campos et al. (2010) confirmed the good results using the transoral anvil technique in 37 patients with a distal esophageal cancer. 21 After a laparoscopic approach in 81% of the patients, the thoracic portion was performed using a muscle-sparing mini-thoracotomy in 23 patients (62%) and thoracoscopic techniques in 14 patients (37%). There were no intra-operative technical failures of the anastomosis or deaths. Five patients developed strictures (13.5%) and all were successfully treated with endoscopic dilatations. One patient had an anastomotic leak (2.7%) that was treated by re-operation and endoscopic stenting of the anastomosis. 68 Side-to-side stapled anastomosis Ben-David et al. described in 2010 six patients with gastro-esophageal junction cancers in whom, after laparoscopic dissection and formation of the gastric conduit, the thorax was approached through a lateral right thoracoscopy. 22 After dissection of the esophagus and division of the azygos vein, the esophagus was divided at the level of the vein using a 60 mm stapler. The gastric conduit was delivered from the abdominal cavity up to the transected esophagus. The transected proximal esophagus and gastric conduit were aligned with sutures. An esophagostomy was created at the distal end of the transected esophagus (with a bougie dilator inside to facilitate the esophagostomy). A gastrotomy was performed at 8 cm proximal to the end of the gastric conduit. With the aid of traction sutures, a side-to-side intrathoracic 6 cm linear stapled esophagogastrostomy was performed by placing the anvil portion of the stapler through the esophagostomy and the cartridge through the gastrostomy. After this, the common openings were sutured with absorbable layer suture followed by a non-absorbable outer layer suture. Following this, the specimen was removed through a 3 cm non-rib spreading incision along the sixth intercostal space. There were no leakages of anastomosis or other post-operative complications. At a median follow up of 9 months there were no post-operative strictures reported. Gorenstein et al. (2011) described a slight different side-to-side anastomosis technique in which the proximal esophagus was not stapled and in addition the whole lumen was used for the

53 Intrathoracic minimally invasive Ivor Lewis esophagectomy 69 construction of the side-to-side anastomosis. 23 Once a 4 cm anastomosis was constructed, the linear stapler was inserted and using traction on the stay sutures at the corners of the anastomosis, the anterior defect was closed. Of the 31 patients operated in this approach, one developed leakage that required re-operation. The leak was repaired and buttressed with an intercostal muscle flap. There were no other anastomotic complications. The side-to-side technique for stapled intrathoracic anastomosis is illustrated in figure 5. Figure 5. Side-to-side stapled intrathoracic anastomosis (for color figure, see color section, page 181)

54 DISCUSSION Chapter 4 70 Despite the new developments (e.g. minimally invasive surgery) in the surgical treatment of esophageal cancer, esophagectomy is still associated with a significant risk of peri-operative morbidity and mortality. 1,6 In order to reduce morbidity and mortality it is important that, at the end of the procedure, a low risk of leakage gastro-esophageal anastomosis is created. With the increase of gastro-esophageal junction tumors, the Ivor Lewis type of resection with intrathoracic anastomosis is increasingly being used. A shorter gastric conduit will permit an extended gastric resection and a well vascularised anastomosis. At the same time, the minimally invasive esophageal resection is increasingly being used with the potential benefits of less pain, lower incidence of respiratory infection, and a reduced hosptial stay. The combination of Ivor Lewis esophagectomy with minimally invasive surgery has the potential of improving the postoperative outcome. An interest for stapler techniques is present in literature compared to the hand sewn technique. 7,24 Stapled anastomosis in the thoracic cavity has been supported by Blackmon et al. who analysed three techniques of intrathoracic esophagogastric anastomosis; hand sewn anastomosis, circular stapled anastomosis, and side-to-side stapled anastomosis. 7 In this propensity matched analysis, no significant differences were reported concerning anastomotic leakage. However, a higher incidence of dysphagia appeared in the hand sewn method and even a fourfold higher incidence of stricture was seen in the hand sewn method. No differences were reported for both circular stapled and side-to-side stapled methods. This study suggests a superior role of the stapler technique for gastro-esophageal anastomoses. This review summarizes the different techniques used to perform an intrathoracic anastomosis safely by thoracoscopy. None of the techniques we described is found superior to the others, but stapled anatomosis seems safe with low percentage of anastomotic leakage and stenosis. There are no important differences identified in this review for the most frequently used anastomosis: stapled with the anvil introduced either transorally or transthoracically. Clinical trials are needed to compare different methods in order to improve the quality of the intrathoracic anastomosis after a minimally invasive Ivor Lewis esophagectomy.

55 Intrathoracic minimally invasive Ivor Lewis esophagectomy 71 REFERENCES 1. Kent M.S., Schuchert M., Fernando H., Luketich J.D. Minimally invasive esophagectomy: state of the art. Diseases of the Esophagus. 2006; 19: Lewis I. The surgical treatment of carcinoma of the oesophagus: with special reference to a new operation for growths of the middle third. British Journal of Surgery. 1946; 34: McKeown KC. Total three-stage oesophagectomy for cancer of the oesophagus. British Journal of Surgery. 1976; 63: Orringer MB, Sloan H. Esophagectomy without thoracotomy. Journal of Thoracic and Cardiovascular Surgery. 1978; 76: Scheepers JJ, van der Peet DL, Veenhof AA, Heijnen B, Cuesta MA. Systematic approach of postoperative gastric conduit complications after esophageal resection. Diseases of the Esophagus. 2009; 23: Biere S.S.A.Y., Maas K.W., Cuesta M.A., van der Peet D.L. Cervical or thoracic anastomosis after esophagectomy for cancer: systematic review and meta-analysis. Digestive Surgery. 2010; 28: Blackmon S.H., Correa A.M., Wynn B., Hofstetter W.L., Martin L.W., Mehran R.J., Rice D.C., Swisher S.G., Walsh G.L., Roth J.A., Vaporciyan A.A. Propensity-matched analysis of three techniques for intrathoracic esophagogastric anastomosis. Annals of Thoracic Surgery. 2007; 83: Watson, D.I., Davies N., Jamieson G.G. Totally endoscopic Ivor Lewis esophagectomy. Surgical Endoscopy. 1999; 13: Cadiere G.B., Dapri G., Himpens J., Fodderie L., Rajan A. Ivor Lewis esophagectomy with manual esogastric anastomosis by thoracoscopy in prone position and laparoscopy. Surgical Endoscopy. 2010; 24: Lee K.W., Leung K.F., Wong K.K., Lau KY, Lai K.C., Leung L.C., Lau, K.W. One-stage thoracoscopic oesophagectomy: ligature intrathoracic stapled anastomosis. Australian and New Zealand Journal of Surgery. 1997; 67: Allsop J.R. Ligature versus purse string for surgical stapled anastomoses. Australian and New Zealand Journal of Surgery. 1992; 62: Ng W.T., Lee, K., Kong C.K., Cheung C.H., Leung L.C. Ligature versus purse string for surgical stapled anastomoses (Comment). Australian and New Zealand Journal of Surgery. 1993; 63: Nguyen N.T., Follette D.M., Lemoine P.H. Roberts P.F. Minimally invasive Ivor Lewis esophagectomy. Annals of Thoracic Surgery. 2001; 72: Nguyen N.T., Longaria M., Chang K., Lee J. Wilson S.E. Thoracolaparoscopic modification of the Ivor Lewis esophagectomy. Journal of Gastrointestinal Surgery. 2006; 10: Misawa K., Hachisuka T., Kuno Y., Mori T., Shinohara M., Miyauchi M. New procedure for purse-string suture in thoracoscopic esophagectomy with intrathoracic anastomosis. Surgical Endoscopy. 2005; 19: Bizekis C., Kent M.S., Luketich J.D., Buenaventura P.O., Landreneau R.J., Schuchert M.J., Alvelo-Rivera M. Initial experience with minimally invasive Ivor Lewis esophagectomy. Annals of Thoracic Surgery. 2006; 82: Thairu N., Biswas S., Abdulaal, Ali H. A new method for intrathoracic anastomosis in laparoscopic esophagectomy. Surgical Endoscopy. 2007; 21: Wittgrove A.C., Clark W.G., Tremblay L.J. Laparoscopic gastric bypass, Roux-en-Y: preliminary report of five cases. Obesity Surgery. 2004; 4: Sutton C.D., White S.A., Marshall L.J., Berry D.P., Veith P.S. Endoscopic-assisted intrathoracic oesophagectomy without thoracotomy for tumours of the lower oesophagus and cardia. European Journal of Surgical Oncology. 2002; 28: ) Nguyen N.T., Hinojosa M.W., Smith B.R., Gray J., Reavis K.M. Thoracoscopic construction of an intrathoracic esophagogastric anastomosis using a circular stapler: transoral placement of the anvil. Annals of Thoracic Surgery. 2008; 86: Campos M.C., Jablons D., Brown L.M., Ramirez R.M., Rabl C., Theodore P. A safe reproducible anastomotic technique for minimally invasive Ivor Lewis esophagectomy: the circular-stapled anastomosis with the transoral anvil. European Journal of Cardio-thoracic Surgery. 2010; 37:

56 22. Ben-David K., Sarosi G.A., Cendan J.C., Hochwald S.N. Technique of minimally invasive Ivor Lewis esophagectomy with intrathoracic stapled side-to-side anastomosis. Journal of Gastrointestinal Surgery. 2010; 14: Gorenstein L.A., Bessler M., Sonett J.R. Intrathoracic stapled esophagogastric anastomosis: an alternative to the end to end anastomosis. Annals of Thoracic Surgery. 2011; 91: Behzadi A., Nichols F.c., Cassivi S.D., Deschamps C., Allen M.S., Pairolero P.C. Esophagectomy: the influence of stapled versus hand-sewn anatomosis on outcome. Journal of Gastrointestinal Surgery. 2005; 9: Chapter 4 72

57 PART II Minimally invasive esophagectomy in retrospect

58

59 Chapter 5 Laparoscopic vs. open transhiatal esophagectomy Maas KW, Biere SSAY, Scheepers JJG, Gisbertz SS, van der Peet DL, Cuesta MA [Submitted]

60 ABSTRACT Chapter 5 76 Background The only curative treatment for esophageal cancer is surgical resection. This treatment is associated with a high morbidity rate and long in-hospital recovery period. Both transthoracic and transhiatal esophagectomies are performed worldwide. The transhiatal approach may reduce the respiratory infection rate in compromised patients with distal esophageal and gastro-esophageal (GE) cancers. Minimally invasive esophagectomy could further improve post-operative outcome. Two cohorts of laparoscopic and open transhiatal esophagectomy for cancer were compared for short- and long-term outcome. Methods From January 2001 through December 2004, 50 patients who underwent laparoscopic transhiatal esophagectomy were compared to a historical group of 50 patients who had undergone open transhiatal esophagectomy between January 1998 and December Post-operative management was identical in both groups. Results: No significant differences were seen between the two groups with regard to baseline characteristics and oncological parameters including resection margin (R0 82% vs. 74%, p=0.334) and 5-year survival. Operation time did not differ significantly between the groups. (300 vs. 280 min, p=0.110). Median hospital stay and intensive care unit stay were significantly shorter in the laparoscopic group (13 vs. 16 days, p=0.001 and 1 vs. 3 days, p=0.000 respectively). Conclusion Minimally invasive transhiatal esophagectomy is feasible and has the same oncological outcome as open transhiatal esophagectomy. Faster recovery without a significant longer operation time could be the major benefit of the laparoscopic transhiatal approach. To our knowledge, this is the largest comparative study in literature comparing laparoscopic transhiatal with open transhiatal esophagectomy for cancers of distal and GE junction. Randomized trials are needed to further clarify the role of laparoscopic transhiatal approach for esophageal cancer.

61 Laparoscopic vs. open transhiatal esophagectomy 77 INTRODUCTION The incidence of adenocarcinoma of the esophagus and gastro-esophageal (GE) junction is rapidly rising. 1,2 The only curative therapy remains surgery. For years, the procedure of choice for esophageal cancer was the Ivor-Lewis operation, later modified by McKeown. 4 With this modified procedure, the esophagus is resected by means of a right-sided thoracotomy combined with a laparotomy using cervical esophagogastric anastomosis. 3,4 The other frequently used procedure is the transhiatal esophageal resection according to Orringer in which a thoracotomy is avoided. 5 Both procedures have high complication rates, varying from 40% to 80%, and the in-hospital mortality rate averages 7.5% to less than 5% in experienced centers. 6 A meta-analysis that compared transthoracic and transhiatal resections concluded that although transthoracic resections had significantly higher pulmonary morbidity and mortality rates, 5-year survival was about 20% after both approaches. 7 The approach and extent of the resection that is necessary is still controversial. In a prospective randomised study by Hulscher et al, transthoracic esophageal resection with systematic abdominal and mediastinal lymph node dissection (two-field lymphadenectomy) were compared with the classic transhiatal approach. 8 The transhiatal approach had a lower morbidity than the extended lymphadenectomy. Even if a trend was observed with an advantage for the transthoracic approach in tumors located in the mid-esophagus, the median survival, disease-free, and quality-adjusted survival for the most common lower esophageal cancers were not statistically significant. 8 A meta-analysis showed that minimally invasive esophagectomy could lower morbidity and shorten hospital stay. 9 This effect was only present for minimally invasive transthoracic esophagectomy as the case-control studies reporting on laparoscopic transhiatal esophagectomy had a small sample size. To date no randomized trial has been performed comparing laparoscopic and open transhiatal esophagectomy. This study compares the short- and long-term results of two cohorts of 50 consecutive patients with cancer of the distal esophageal and GE junction who were approached by a minimally invasive procedure or an open procedure. To our knowledge, this is the largest case-control study in literature comparing laparoscopic transhiatal esophagectomy with open transhiatal resection. 9 PATIENTS AND METHODS From January 2001 through December 2004, fifty consecutive patients who underwent laparoscopically assisted transhiatal esophageal resection in the VU university medical center were prospectively followed. The results were compared with an unselected historical group of fifty consecutive patients who underwent an open transhiatal esophageal resection in the VU university medical center in the pre-laparoscopic period of January 1998 through December All patients presented with a squamous cell carcinoma or an adenocarcinoma of the distal 5 centimetres of the esophagus or the GE junction.

62 Patients with previous upper abdominal surgery did not undergo a laparoscopic approach. Patients with a colon interposition were excluded for the analysis. Pre-operative staging was performed by means of endoscopic ultrasound, Computed Tomography (CT) -scan of thorax and abdomen and a neck ultrasound. After May 1999, some patients with locally advanced esophageal cancer (T3, N0, N1) received neoadjuvant chemo-immunotherapy (Cisplatin, Gemcitabine plus GM-CSF). Operative Technique The laparoscopic transhiatal esophagectomy was described in an earlier publication by Scheepers et al. 10 An identical procedure described by Orringer and Sloan 6 was performed in the patients who underwent an open transhiatal esophageal resection. Pyloroplasty was performed in patients who had surgery before Chapter 5 78 Post-operative Management Post-operatively, patients were ventilated mechanically at the intensive care unit (ICU) and extubated when hemodynamically and respiratory stable. Extubated patients were admitted to the medium care unit (MCU) and from there to the regular ward. Patients were fed through the jejunostomy feeding tube from the first day after their operation until the oral feeding could be completely resumed. On post-operative day 5, a swallow x-ray examination was performed to assess the anastomosis and gastric tube passage. When no leakage and a good passage were seen, the nasogastric tube was removed and oral feeding was started. Patients were discharged when they were completely mobile and able to feed themselves orally. Statistical Analysis Statistical analysis was performed using the SPSS software package (SPSS, Chicago, IL, USA). Medians and interquartile ranges at the 25th and 75th percentile were calculated and subsequently depicted when relevant. Mann-Whitney U tests and chi-square tests were used when appropriate. Survival curves were obtained using the Kaplan-Meier method. Survival of both groups was compared with the log-rank test. Significance was set at p<0.05. RESULTS Between January 2001 and December 2004, fifty consecutive patients with a squamous cell carcinoma or an adenocarcinoma of the distal esophagus or GE junction underwent laparoscopic transhiatal esophageal resection. The results were compared with the results of the group of fifty consecutive patients with tumors at the same localization who underwent a conventional open transhiatal esophageal resection in the pre-laparoscopy period between January 1998 and December 2000.

63 Laparoscopic vs. open transhiatal esophagectomy 79 Table 1. Baseline characteristics Laparoscopic transhiatal esophageal Open transhiatal esophageal p resection (n=50) resection (n=50) Gender No. (%) Male 41 (82%) 33 (66%) Female 9 (18%) 17 (34%) Age * 62.5 (57-69) 65 (57-69) ASA (%) I 7 (14%) 11 (22%) II 36 (62%) 25 (50%) III 12 (24%) 14 (28%) *, median (interquartile range). There were no significant differences between the groups in terms of gender, age, and American Society of Anaesthesiologists (ASA) distribution (Table 1). Fourteen patients (28%) in the open group and no patients in the laparoscopic group underwent a pyloroplasty (p=0.000). Moreover, a significant difference was observed between the laparoscopic and open groups in the number of patients who received neoadjuvant chemotherapy (23 vs 13, p=0.037). Nine laparoscopically assisted operations (18%) were converted to open procedures. The reasons for conversions are depicted in table 2. Laparoscopic mediastinal dissection of the esophagus could be accomplished in 44 patients (88%). Table 2. Reasons for conversion from laparoscopic transhiatal esophagectomy to an open transhiatal esophagectomy Reason for conversion Number of patients Splenic bleeding necessitating splenectomy 2 Insufficient visualisation due to liver 3 Evaluation of resectability Solid celiac trunk 2 Tumor adherent to pleura 1 Tumor adherent to pancreas 1 Total 9 (18%) Tumor characteristics are listed in Table 3. There are no important differences between the groups in terms of the histologic type of tumor, TNM stage, tumor localization and tumor differentiation. Tumor-free margins were obtained in 41 (82%) of the 50 patients who underwent laparoscopic resection and in 37 patients (74%) after open resection. R1 resections were carried out in 9 (18%) and 13 (26%) respectively. No R2 resections were carried out in both groups. No significant difference was found between the two groups regarding radicality of resection (p=0.334). The median number of harvested lymph nodes was 14 (interquartile range: 10 19) in the laparoscopic group and 11 (interquartile range: 8 15) in the open group (p=0.754).

64 Chapter 5 Table 3. Tumor characteristics of patients who underwent laparoscopic or open transhiatal esophageal resection Oncologic characteristics Laparoscopic transhiatal esophageal resection (n=50) Open transhiatal esophageal resection (n=50) Histologic type Squamous cell carcinoma 12 (24%) 16 (32%) Adenocarcinoma 37 (74%) 32 (64%) Undifferentiated 1 (2%) 2 (4%) TNM stage I 3 (6%) 4 (8%) IIa 10 (20%) 12 (24%) IIb 6 (12%) 3 (6%) III 31 (62%) 31 (62%) Tumor differentiation Good 3 (6%) 5 (10%) Moderate 10 (20%) 10 (20%) Moderate/poor 8 (16%) 7 (14%) Poor 29 (58%) 28 (56%) Radicality of surgery R0 41 (82%) 37 (74%) R1 9 (18%) 13 (26%) Number of lymph nodes * 14 (10-19) 11 (8-15) *, median (interquartile range). p 80 The median operation time was longer in the laparoscopic group (300 minutes, interquartile range: ) than in the open procedure (280 minutes, interquartile range: ) but the difference was not significant (p=0.110). Median blood loss was less in the laparoscopic group (500 ml, interquartile range: ) compared with the open group (900 ml, interquartile range: ) (p=0.000). Median post-operative ICU stay was longer in the open group (3.0 days, interquartile range: ) vs. the laparoscopic group (1.0 days, interquartile range: ) (p=0.000). The median hospital stay was 13 days (interquartile range: 11 16) in the laparoscopic and 16 days (interquartile range: 14 20) in the open group (p=0.001) (Table 4). Table 4. Operative and post-operative data of patients that underwent laparoscopic transhiatal esophageal resection compared with open transhiatal esophageal resection Laparoscopic transhiatal esophageal resection (n=50) Open transhiatal esophageal resections (n=50) Operation time (minutes) * 300 ( ) 280 ( ) Blood loss (ml) * 500 ( ) 900 ( ) ICU stay (days) * 1.0 ( ) 3.0 ( ) Hospital stay (days) * 13 (11-16) 16 (14-20) *, median (interquartile range). p

65 Laparoscopic vs. open transhiatal esophagectomy 81 Morbidity and Mortality No hospital mortality was recorded for the laparoscopic group, and one patient died after an open procedure (2%) due to ARDS. The morbidity rate was comparable in the laparoscopic (42%) and open (66%) group (Table 5). Table 5. Early post-operative morbidity and mortality after laparoscopic transhiatal esophageal resection and open transhiatal esophageal resection Laparoscopic transhiatal esophageal resection (n=50) Open transhiatal esophageal resection (n=50) In-hospital mortality 0 (%) 1 (2%) Morbidity-no. (%) Patients with complications 21 (42%) 33 (66%) Pulmonary complications 9 (18%) 13 (26%) Cardiac complications 3 (6%) 6 (12%) Recurrence nerve palsy 3 (6%) 5 (10%) Chylus leakage 2 (4%) 1 (2%) Tracheal rupture 0 (0%) 1 (2%) Abdominal woundinfection 0 (0%) 3 (6%) Cervical fistula 4 (8%) 3 (6%) Herniation of small intestine into thorax 0 (0%) 1 (2%) Re-operation-no. (%) 2 (4%) 3 (6%) Inspection of cervical anastomosis 1 (2%) 1 (2%) Repositioning of small intestine 0 (0%) 1 (2%) and hiatoplasty Tracheal repair 0 (0%) 1 (2%) Correction of evisceration 1 (2%) 0 (0%) p Pulmonary (i.e. respiratory infections) and cardiac complications were seen less often in the laparoscopic group than in the open group (12 vs. 19), however not statistically significant (p=0.130). The re-operation rate was 2 (4%) and 3 (6%) in the laparoscopic and open group respectively (p=0.646). The median follow-up was shorter in the laparoscopic group (median 35 months, interquartile range 25 68) compared with the open group (median 46; interquartile range 25 68). Kaplan-Meier analysis at 36 months showed an overall survival of 36% (95% confidence interval: ) for the laparoscopic group and 38.3% (95% confidence interval: ) for the open group (Figure 1). Median disease free survival at 36 months was 31% (95% confidence interval: ) for the laparoscopic group versus 30.0% (95% confidence interval: ) for the open group. No statistical differences in mean survival and mean disease free survival were found after the cohorts were corrected for neoadjuvant therapy.

66 Figure 1. Kaplan-Meier analysis of the overall survival Months DISCUSSION Chapter 5 82 The best approach to the dissection and its extent are controversial. An approach to the esophagus through a right thoracotomy in combination with a laparotomy and cervical incision is associated with a significant rate of pulmonary complications that will account for the long ICU stay and increased morbidity and mortality. The transhiatal approach will reduce this complication rate by avoiding the collapse of the lung produced during the esophageal resection by right thoracotomy. The drawback of this approach is the blind character of the mediastinal dissection. The ongoing debate of transthoracic versus transhiatal approach includes this drawback. To date both transthoracic and transhiatal esophagectomy are performed worldwide for distal esophageal or GE junction carcinomas. The objective of this study was to investigate, in the largest case-control study in literature, the role and feasibility of laparoscopic transhiatal esophagectomy. Several minimally invasive approaches have been described to reduce operative trauma and improve dissection of the esophagus and tumor. The right thoracoscopic approach, the transhiatal approach, and microscopic endoscopic mediastinal dissection are already being performed , No randomized trials have been performed comparing laparoscopic transhiatal esophagectomy with an open resection. Randomized studies comparing an open with laparoscopic approach for colon cancer have shown important post-operative advantages of the laparoscopic approach, with at least the same oncological outcome at 4 years of follow-up. 23 Therefore, more surgical groups are performing minimally invasive esophageal resection for cancer In spite of initial high percentage of respiratory complication after thoracoscopic esophageal resection. 13,24 The systematic standardization of the procedure by Luketich et al, 14 has demonstrated that the three-stage operation can be performed safely, in an acceptable operating time, with

67 Laparoscopic vs. open transhiatal esophagectomy 83 important advantages in the post-operative recovery of the patients and an oncological outcome at least as good as that after conventional surgery. Luketich et al, described their experience with 222 patients (175 with esophageal cancer and 46 with high-grade dysplasia), most of whom underwent an esophageal resection by right thoracoscopy and laparoscopy (three-stage operation). The median ICU stay was 1 day and the hospital stay was 7 days, with an operative mortality of 1.4%. They reported 16 conversions to open surgery. Major and minor complication rates were 32% and 23.9%, respectively. The same results are found by Nguyen et al. 19,20 The laparoscopic transhiatal approach has been performed in more limited number of patients by different authors. 12,18,22 Conversion rates of 11.6% and morbidity rates of 30% have been described by Bonavina et al. in a series of 43 patients, but no comparative studies have been published. 18 The results of the series presented here, concerning morbidity and mortality are consistent with the results published in the literature for both the laparoscopic and the open transhiatal approach. 7,8,12,18,22 Furthermore, there are no differences concerning morbidity, mortality and operation time between the laparoscopic and open groups, significantly less blood loss, shorter ICU stay and hospital stay was found in the laparoscopic transhiatal approach. A point of consideration might be the high conversion rate of 18%, this could be possible caused by the learning curve. Nevertheless, in 88% of the patients, the mediastinal dissection of the esophagus could be accomplished laparoscopically without important anaesthesiological hazards, especially in relation to the opening of the pleuras that could produce a tension pneumothorax. Adaptation of PEEP pressure and an increase of minute volume of the mechanical ventilation could avoid this problem and consequent conversion in all patients. 19 Furthermore, laparoscopic transhiatal approach will permit perfect visualization of the mediastinal structures in relation to the tumor up to the carina, making this operation no longer a blind procedure, avoiding also the haemodynamic instability during the conventional dissection by the use of the retractor and manual dissection. Retrieval of the tumor through a small well protected transumbilical incision instead of through a cervical incision may avoid the appearance of port-site metastases as in the case of laparoscopic colonic surgery for cancer. Moreover, once the specimen is retrieved, dissection around the pylorus and the origin of the gastroepiploic vessels, can be accomplished followed by formation of the gastric tube, using the conventional GIA-90. In this fashion the operation is time sparing and cost-effective. The use of pyloroplasty remains controversial as well. 24 Even though many authors still include the drainage of the pylorus in the operative procedure. 14 In the current study, with the exception of the first 14 operated patients, who underwent a routine pyloroplasty procedure, the avoidance of this pyloroplasty in the following patients did not lead to any emptying problems of the gastric tube during the post-operative period. Therefore, we do not recommend a routine pyloroplasty as part of the gastric tube formation. Interpreting the results of this study one has to consider the fact that the outcome of the 9 patients whome the laparoscopic procedure was converted to an open procedure, were analyzed within the laparoscopy group.

68 The laparoscopic transhiatal approach used in this study showed important advantages over the open approach, including less operative blood loss, shorter ICU stay, and shorter hospital stay with the same oncological outcome. This makes the laparoscopic transhiatal esophageal resection for tumors of the distal esophagus a feasible procedure. A randomized study would further clarify the role of a laparoscopic approach for distal esophageal cancer. Chapter 5 84

69 Laparoscopic vs. open transhiatal esophagectomy 85 REFERENCES 1. Devesa SS, Blot WJ, Fraumeni JF Jr. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 1998; 83: Pera M, Cameron AJ, Trastek VF, Carpenter HA, Zinsmeister AR. Increasing incidence of adenocarcinoma of the esophagus and esophagogastric junction. Gastroenterology 1993; 104: Mckeown KC. Total 3-stage esophagectomy for cancer of the esophagus. Br J Surg 1976; 63: Lewis I. The surgical treatment of carcinoma of the oesophagus with special reference to a new opera tion for growths of the middle 3rd. Br J Surg 1946; 34: Orringer MB, Sloan H. Esophagectomy without thoracotomy. J Thorac Cardiovasc Surg 1978; 76: Collard JM, Otte JB, Fiasse R, Laterre PF, De Kock M, Longueville J, Glineur D, Romagnoli R, Reynaert M, Kestens PJ. Skeletonizing en bloc esophagectomy for cancer. Ann Surg 2001; 234: Hulscher JBF, Tijssen JGP, Obertop H, van Lanschot JJB. Transthoracic versus transhiatal resection for carcinoma of the esophagus: a meta-analysis. Ann Thorac Surg 2001; 72: Hulscher JBF, van Sandick JW, de Boer AG, Wijnhoven BP, Tijssen JG, Fockens P, Stalmeier PF, ten Kate FJ, van Dekken H, Obertop H, Tilanus HW, van Lanschot JJB. Extended transthoracic resection compared with limited transhiatal resection for adenocarcinoma of the esophagus 3. N Engl J Med 2002; 347: Biere SSAY, Cuesta MA, van der Peet DL. Minimally invasive versus open esophagectomy for cancer: a systematic review and meta-analysis. Minerva Chirurgica. 2009; 64: Scheepers JJ, Veenhof AA, van der Peet DL, van Groeningen C, Mulder C, Meijer S, Cuesta MA. Laparoscopic transhiatal resection for malignancies of the distal esophagus: outcome of the first 50 resected patients. Surgery. 2008;143: Azagra JS, Ceuterick M, Goergen M, Jacobs D, Gilbart E, Zaouk G, Carlier E, Lejeune P, Alle JL, Mathys M. Thoracoscopy in esophagectomy for esophageal cancer. Br J Surg 1993; 80: DePaula AL, Hashiba K, Ferreira EAB, DePaula RA, Grecco E. Laparoscopic transhiatal esophagectomy with esophagogastroplasty. Surg Laparosc Endosc 1995; 5: Dallemagne B et al. Subtotal esophagectomy by thoracoscopy and laparoscopy. Minim Invasive Ther Allied Technol 1992; 1: Luketich JD, Alvelo-Rivera M, Buenaventura PO, Christie NA, McCaughan JS, Litle VR, Schauer PR, Close JM, Fernando HC. Minimally invasive esophagectomy outcomes in 222 patients. Ann Surg 2003; 238: Buess G, Kaiser J, Manncke K, Walter DH, Bessell JR, Becker HD. Endoscopic microsurgical dissection of the esophagus (EMDE). Int Surg 1997; 82: Pinotti HW, Zilberstein B, Pollara W, Raia A. Esophagectomy without thoracotomy. Surg Gynecol Obstet 1981; 152: Makay O, van den Broek WT, Yuan JZ, Veerman DP, Helfferich DW, Cuesta MA. Anesthesiological hazards during laparoscopic transhiatal esophageal resection A case control study of the laparoscopicassisted vs the conventional approach. Surg Endosc 2004; 18: Bonavina L, Bona D, Binyom PR, Peracchia A. A laparoscopy-assisted surgical approach to esophageal carcinoma. J Surg Res 2004; 117: Nguyen NT, et al. Comparison of minimally invasive esophagectomy with transthoracic and transhiatal esophagectomy. Arch Surg 2000; 135: Nguyen NT, Roberts P, Follette DM, Rivers R, Wolfe BM. Thoracoscopic and laparoscopic esophagectomy for benign and malignant disease: lessons learned from 46 consecutive procedures. J Am Coll Surg 2003; 197: Swanstrom LL, Hansen P. Laparoscopic total esophagectomy 3. Arch Surg 1997; 132: Van den Broek WT, Makay O, Berends FJ, Yuan JZ, Houdijk AP, Meijer S, Cuesta MA. Laparoscopically assisted transhiatal resection for malignancies of the distal esophagus. Surg Endosc 2004; 18: Clinical Outcomes of Surgical Therapy Study Group. A comparison of laparoscopically assisted and open colectomy for colon cancer. N Engl J Med 2004; 350: Law S, Fok M, Chu KM, Wong J. Thoracoscopic esophagectomy for esophageal cancer. Surgery 1997; 122: 8-14.

70 25. Mannell A, Mcknight A, Esser JD. Role of pyloroplasty in the retrosternal stomach results of a prospective, randomized, controlled trial. Br J Surg 1990; 77: Chapter 5 86

71 Chapter 6 Thoracoscopic esophageal resection for cancer in prone decibitus position: operative technique and short-term results Maas KW, Biere SSAY, Markinez I, Gisbertz SS, van der Peet DL, M.A. Cuesta [Submitted]

72 ABSTRACT Chapter 6 88 Background The minimally invasive approach for esophageal resection is increasingly being used in the world. Important advantage is avoidance of thoracotomy while obtaining the same quality of specimen. Thoracoscopic esophagectomy in prone position also avoids total collapse of the right lung. This may be associated with an important reduction in respiratory infections. We evaluated our own initial experience of a transthoracic esophagectomy with the thoracoscopic procedure in prone position (MIEprone) combined with laparoscopy or laparotomy in 40 patients. Methods A retrospective analysis was performed of 40 patients who were operated between March 2007 and March 2009 with a resectable (T3N0 1M0) esophageal cancer by a right thoracoscopy in prone position and laparoscopy or laparotomy. Patients could undergo a MIEprone if they had a thoracic or gastro-esophageal tumor, had no prior thoracic surgery and were in a good condition (Eastern Cooperative Oncology Group 2). Results Twenty-five patients had a laparoscopy and fifteen a laparotomy. Median operative time was 290 minutes with a conversion rate of 2.5% to thoracotomy and laparotomy. Median Intensive Care Unit and hospital stay were 1 day and 14 days respectively. A radical resection (R0) was achieved in 36 patients (90%). The median number of resected lymph nodes was 21 (15 33). Morbidity was observed in ten patients, three cervical anastomosis leakages and two moderate chylous leakages were treated conservatively. One patient had proximal ischemia of the gastric conduit which was resected followed by re-anastomosis and stenting. Four patients developed a respiratory infection which was treated with antibiotics. No mortality was observed in this series. Conclusion The short-term results in this series of 40 patients show the safety and feasibility of a thoracoscopic esophagectomy in prone position in our center. A multicenter randomized trial, in which this approach will be compared with the standard open thoracotomy, is needed to establish the role of this approach in daily practice.

73 Thoracoscopic esophageal resection for cancer in prone position 89 INTRODUCTION The minimally invasive approach for esophageal resection is increasingly being used in the world. Its most important advantage seems to be the reduction of complications produced by the thoracotomy and/or laparotomy. Consequences of the reduced operative trauma may be less post-operative pain, and possibly fewer respiratory complications. Initially, surgeons have used the right lateral thoracoscopic approach with total lung block in order to visualize and dissect the esophagus. 1 However, initial reports show a high conversion rate to thoracotomy of 10% to 17% and a high respiratory morbidity of 17 to 42%. 2 4 Searching for a further reduction of the conversion rate and the respiratory infection rate, Cuschieri developed the thoracoscopic approach in a prone decubitus position so that a total collapse of the lung is not necessary for dissecting the esophagus and in this way further reducing respiratory infections. 5 We evaluated our own initial experience of a transthoracic esophagectomy with the thoracoscopic procedure in prone position (MIEprone) combined with laparoscopy or laparotomy in 40 patients. METHODS A retrospective analysis was performed of 40 patients who were operated between March 2007 and March 2009 with a resectable (T3N0 1M0) esophageal cancer by a right thoracoscopy in prone position and laparoscopy or laparoscopy (MIEprone). Patients could undergo a MIEprone if they had a thoracic or gastro-esophageal tumor, had no prior thoracic surgery and were in a good condition (Eastern Cooperative Oncology Group (ECOG) 2). Operative technique 1) After induction of general anaesthesia, a standard intratracheal intubation follows. The patient is then positioned in prone decubitus position on a standard support device in order to support the head, shoulders and pelvis. The abdomen is held free for breathing excursions. The position of the arms is very important in order to get abduction of the scapula. The arms are positioned on a support device for the flexion of the shoulders and ellebows (Figure 1). In this way, the area between the spine and the inner edge of the scapula is broadened. 2) The surgeon stands on the right side of patient with the first assistant on his/her right side looking at the monitor in front of them. Scrub nurse stands on the left side of the surgeon. 3) Four trocars are placed along the inner edge of the right scapula (Figure 2): the first trocar at the level of the lowest point of the scapula, 10 mm (may be 5 mm) for the 30 degrees thoracoscope; the second trocar at the level of 4 th intercostal space, 5 mm; the third trocar at the level of 8 th intercostal space, 12 mm; and the last, at the level of 3th intercostal space, 5 mm, as work trocar for the assistant (suction, lung retraction etc). The first trocar is introduced

74 Figure 1. Prone position Figure 2. Placement of the trocars for right thoracoscopy with the patient in prone position. Chapter 6 90 Note the edge of the scapula openly in the thoracic cavity, after control by finger palpation so that the thoracic space is free of adhesions. After the introduction of the first trocar, a positive insufflation of between 6 and 8 mm Hg is initiated in order to retract enough of the right lung for an adequate visualization of the posterior mediastinum. 4) Inspection of the pleural cavity for metastases and tumor is performed in order to confirm that resection is possible. Dissection starts by dissecting the pulmonary ligament, followed by dissecting the mediastinal pleura anteriorly along the pericardial sac to the hilum of the right

75 Thoracoscopic esophageal resection for cancer in prone position Figure 3. Dissection of the pleura (for color figure, see color section, page 182) Figure 4. Visualization of the carina (for color figure, see color section, page 182) lung up to the azygos vein. In this way, the pericardial sac, the right pulmonary vein, right bronchus and carina are subsequently visualized (Figure 3 and 4). Posteriorly, the mediastinal pleura is cut longitudinally at the posterior edge of the azygos vein from the arch of the azygos vein to the costo-phrenic angle. In this way, a broad piece of pleura is resected along with the esophagus. The thoracic duct is dissected at the level of the costo-phrenic angle and ligated with clips. The rest of the duct is resected with the specimen. 5) Dissection proceeds anteriorly, first along the pericardial sac, right pulmonary vein, as deep as the left mediastinal pleura is visualized. Dissection then continues up to the right bronchus. The dissection between the triangle formed by right pulmonary vein and right bronchus is important. The lymph nodes (LN) localized here have to be dissected with the specimen. 91

76 Figure 5. Paratracheal and carinal lymphadenectomy (for color figure, see color section, page 183) Chapter ) Along the plane of the descending aorta, the esophagus with peri-esophageal tissues is dissected free. Vascular branches from the aorta to the esophagus may be cut with Ligasure or Ultracision R devices. 7) Furthermore the azygos vein is dissected and divided with a vascular endo-stapler. 8) Dissection proceeds with complete lymphadenectomy of the right bronchus, carina and left bronchus lymph nodes. The bronchial artery situated behind the azygos vein may be spared; the right vagal nerve cut at the level of the right bronchus. 9) Dissection continues between esophagus and trachea (pars membranacea) in proximal direction, to stop 3 cm from the apex of the thoracic cavity, leaving a small cuff of pleura intact. At this level, it is possible to visualize the left recurrent nerve and to perform paratracheal lymphadenectomy at both sides (Figure 5). 10) After control of haemostasis, a thoracic drain is left in the posterior mediastinum and the thoracoscopic phase is considered finalized. 11) The patient is positioned from the prone to supine position for the laparoscopic and cervical phase of the operation. The stomach is dissected free leaving intact the gastro-epiploic vessels along the greater curvature. A lymphadenectomy of the celiac trunk is now performed and dissection continues in the direction of the hiatus. This is enlarged and dissected in order to communicate the abdominal and the thoracic dissection. A second team approaches the esophagus at the cervical area. After division at the cervical area, the specimen can be retrieved through a well-protected small supra-umbilical incision. Extracorpereal resection can be completed and a gastric tube can be created. The gastric tube is smoothly directed into the cervical area and anastomosed. No pyloroplasty is required. Through the small incision, a feeding jejunostomy for direct post-operative nutrition is made. A laparoscopic intra-abdominal gastric tube reconstruction using the endo-stapler is an optional alternative method, followed by cervical retrieval of the specimen.

77 Thoracoscopic esophageal resection for cancer in prone position 93 Statistical analysis Data are expressed median with ranges for continuous variables. The Mann Whitney U test was used for analysis. Data has been analysed in SPSS software version 14. RESULTS Patient characteristics The patient characteristics are depicted in Table 1. The thirty males and ten female patients had a median age of 67 years with a range of 48 to 80 years. Tumors were located in the thoracic esophagus (34 patients) and gastro-esophageal (GE)-junction tumors (6 patients) and were found resectable by Computed Tomography- scan of thorax and abdomen, endoscopic ultrasound and Positron Emission Tomography-scan (PET). Fifteen patients had a squamous cell cancer and 25 an adenocarcinoma. Neoadjuvant therapy consisted of chemoradiotherapy with weekly administration of Paclitaxel 50 mg/m2 and Carboplatin AUC= 2 for 5 weeks and concurrent radiotherapy (41,4 Gy in 23 fractions, 5 days per week). A small group of patients were treated with pre-operative chemotherapy alone according to the MAGIC protocol 6. After the neoadjuvant treatment, an interval of six weeks was maintained before operation. Table 1. Patient characteristics. Thoracoscopic esophageal resection in prone decubitus position (n=40) Age (years) 67 (48-80) Gender Male Female Histology of tumor Squamous cell carcinoma Adenocarcinoma Tumor location Thoracic esophagus Gastro-esophageal junction TNM classification T2N1 T3N0-1 Neoadjuvant therapy Chemoradiotherapy Chemotherapy alone

78 General outcome Table 2 depicts general outcome. Median operative time of the thoracoscopic approach was 130 minutes (range 100 to 140 minutes), with a total operative time of 290 minutes (range 240 to 450 min). Blood loss was 230 ml (range 150 to 400 ml). In one patient conversion to thoracotomy was necessary because of a combination of difficulty in developing a surgical plane along the aorta and due to moderate bleeding. The patient was turned to a left lateral position and conventionally approached. Venous bleeding came from a venous plexus located between the aorta and azygos vein. In twenty-five patients, the abdomen was approached laparoscopically, and in fifteen patients it was performed through a median upper laparotomy because of relative contra-indications for laparoscopy, such as previous laparotomy, extreme obesity and PET positive lymph nodes at the prepyloric small curvature in GE-junction tumors. In one patient, conversion from laparoscopy to laparotomy was required due to the presence of extensive fibrosis in the celiac trunk after chemotherapy. Chapter 6 94 Table 2. General outcome. Thoracoscopic esophageal resections in prone decubitus position (n=40) Operation time thoracoscopy (minutes) 130 ( ) Blood loss (ml) 230 ( ) ICU stay (days) 1 (1-37) Hospital stay (days) 14 (12-69) R0 resection 36 (90%) Complete response 8 (20%) No. of retrieved lymph nodes 21 (15-33) ICU, Intensive care unit Pathological examination showed a R0 resection in 36 patients (90%) with a complete response after chemoradiotherapy in eight patients (20%). Median number of LN resected in this two field LN operation were 21 (range 15 to 33). Morbidity Post-operative complications were observed in 10 patients (25%) (table 3). In the major complications group, three patients had an anastomotic leakage at the cervical anastomosis; the next two had moderate chylous leakage and were treated conservatively by means of medium chain triglycerides (MCT) diet. In the last patient the proximal 4 cm of the gastric tube was found to be necrotic at endoscopy which was performed because of respiratory insufficiency. The patient was re-operated through a combined cervico-laparotomy, the necrosis was resected and the gastric tube-esophagus re-anastomosed after extra gastric mobilization. In addition, an endoscopic stent was placed in this patient. No in-hospital mortality was observed in this series.

79 Thoracoscopic esophageal resection for cancer in prone position 95 Table 3. Morbidity. Thoracoscopic esophageal resections in prone Management decubitus position (n=40) Anastomotic leakage 3 Conservative Moderate chylus drainage 2 Medium Chain Triglycerides diet Necrosis of proximal 4 cm of gastric tube 1 Re-operation Pneumonia (not related to other complications) 4 Antibiotics DISCUSSION This study evaluated the initial results of thoracoscopy in prone position for esophageal cancer. Short term results were analyzed in 40 patients. Both pathological and post-operative morbidity were investigated. Dallemagne introduced the thoracoscopic approach for esophageal cancer in The right lateral thoracoscopic approach with total lung block was used in order to visualize and dissect the esophagus. 1 However, initial reports showed a high conversion rate to thoracotomy of 10 to 17% and a high respiratory morbidity of 17 to 42%. 2 4 Searching for reduction of the conversion rate and the respiratory infection rate, Cuschieri et al. designed the thoracoscopic approach in prone decubitus position so that a total collapse of the lung is not necessary for dissecting the esophagus and in this way reducing respiratory infections. 5 Prone decubitus position for conventional lung resection for cancer was initially described by Overholt in 1949 and later modified by Nacleiro in ,9 In comparison with the standard right thoracotomy, the advantages of this approach were: a) the attainable range of thoracic cage and diaphragmatic excursion is greater than in the side position; b) the amplitude of mediastinal swing or displacement is less; c) exposure of the posterior aspect of the hilum and esophageal area is facilitated; d) the weight of the lung itself allows it to fall forward; and e) in the event of haemorrhage the blood flows away from its source, thus permitting its control with greater ease. The approach had not been used again prior to the introduction of the esophageal approach by prone decubitus right thoracoscopy. Controversy exists which of the two main thoracoscopic approaches offers maximal advantages to the patient. In both approaches, a change of position from lateral or prone to supine position is necessary in order to finalize the procedure. Perhaps conversion to thoracotomy will be more difficult, in the event of bleeding in the prone position. Change of position from prone to lateral may be hazardous. The most striking difference between the two approaches is the use of a complete block of the right lung by the lateral thoracoscopy (double lumen intubation) versus the limited block and normal intubation in the prone position. When comparing the prone and the lateral thoracoscopic approaches, the prone position may cause fewer respiratory complications. 10 Luketich et al reported, in a series of 222 patients, an incidence of respiratory complications

80 Chapter 6 of 7.6% of the patients and Palanivelu et al, using the prone position in 134 patients, reported 1.54% in his series. 11,12 In contrary Hulscher et al. observed 57% respiratory infections in their transthoracic group. 13 Possible factors for the better outcome in prone position could be the use of a single endotracheal tube with two-lung ventilation without total collapse of the right lung. The partial ventilation of the right lung reduces the possibility of arterio-venous shunt. Moreover, in prone position the functional residual capacity may be greater than in supine position, the ventilation-perfusion ratio is maintained and hypoxia and hypercarbia are avoided. 14 This may reduce the extent of atelectasia and infection post-operatively. Other important advantages of the prone position may be shorter anaesthesia time, a decrease lung injury, an excellent exposure of the operative field, and a better ergonomy for the surgeon. The short-term results in this series of 40 patients show the safety and feasibility of a thoracoscopic esophagectomy in prone position in our center. Pathologic outcome seems to be similar with open resection. Respiratory complications is reduced when comparing with other series. This improved morbidity seems to enhance recovery, 14 days hospital stay. Technical complications like anastomotic leakage, recurrent nerve lesion and chylothorax, related to esophageal dissection and creation of gastric tube, are expected, to remain the same as open conventional approach. It is however clear from this series that a multicenter randomized protocol has to be performed in which the thoracoscopic approach in prone position will need to be compared to the open posterolateral thoracotomy and laparotomy. 96

81 Thoracoscopic esophageal resection for cancer in prone position 97 REFERENCES 1. Cuschieri A, Shimi S, Banting S. Endoscopic oesophagectomy through a right thoracoscopic approach. Journal of the Royal College of Surgeons of Edinburgh. 1992; 37: Azagra JS, Ceuterick M, Goergen M et al. Thoracoscopy in oesophagectomy for oesophageal cancer. British Journal of Surgery. 1993; 80: Gossot D, Fourquier P, Celerier M. Thoracoscopic oesophagectomy: technique and initial results. Annals of Thoracic Surgery. 1993; 56: McAnena OJ, Rogers J, Williams NS. Right thoracoscopically assisted oesophagectomy for cancer. British Journal of Surgery. 1994; 81: Cuschieri A. Thoracoscopic subtotal oesophagectomy. Endoscopic Surgery and Allied Technologies. 1994; 2: Cunningham D, Allum WH, Stenning SP, Thompson JN, van de Velde CJ, Nicolson M et al. Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. New England Journal of Medicine. 6; 355: Dallemagne B, Weerts JM, Jehaes C. Thoracoscopic oesophageal resection. In: Minimally Invasive Surgery in gastrointestinal cancer : M. A. Cuesta and A. G. Nagy, editors p Overholt RA, Woods FM. The prone position in thoracic surgery. Journal of the International College of Surgery. 1946; 10: (222). 9. Nacleiro E.A. Operative positions and approaches in pulmonary surgery; value of the suspenden prone position. American Journal of Surgery. 1955; January: volume Fabian T, Martin J, Katigbak M, McKelvey AA, Federico JA. Thoracoscopic esophageal mobilization during minimally invasive esophagectomy: a head-to-head comparison of prone versus decubitus positions. Surgical Endoscopy. 2008; 22: Luketich JD, Alvelo-Rivera M, Buenaventura PO et al. Mnimally Invasive Oesophagectomy. Outcomes in 222 patients. Annals of Surgery. 2003; 238: Palanivelu C, Prakash A, Senthilkumar R, Senthilnathan P, Parthasarthi R, Rajan S et al. Minimally invasive esophagectomy: thoracoscopic mobilization of the esophagus and mediastinal lymphadenectomy in prone position--experience of 130 patients. Journal of the American College of Surgeons. 2006; 203: Hulscher JBF, van Sandwick JW, de Boer AG, Wijnhoven BP, Tijssen JG, Fockens P et al. Extended transthoracic resection compared with limited transhiatal resection for adenocarcinoma of the esophagus. New England Journal of Medicine. 2002; 347: Yatabe T, Kitagawa H, Yamashita K, Akimori T, Hanazaki K, Yokoyuma M. Better postoperative oxygenation in thoracoscopic esophagectomy in prone position. Journal of Anesthesia. 2010; 24:

82

83 PART III Prospective minimally invasive esophagectomy