Best Practice & Research Clinical Gastroenterology 28 (2014) 41–52
Contents lists available at ScienceDirect
Best Practice & Research Clinical Gastroenterology
4
Minimally invasive surgery for oesophageal cancer Maarten C.J. Anderegg, MD, PhD Candidate 1, Suzanne S. Gisbertz, MD, PhD, Assistant Professor of Surgery 2, Mark I. van Berge Henegouwen, MD, PhD, Assistant Professor of Surgery * Department of Surgery, Academic Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands
a b s t r a c t Keywords: Oesophageal cancer Minimally invasive surgery Oesophagectomy Thoracoscopy Laparoscopy
Worldwide an increasing part of oncologic oesophagectomies is performed in a minimally invasive way. Over the past decades multiple reports have addressed the perioperative outcomes and oncologic safety of minimally invasive oesophageal surgery. Although many of these (retrospective) case–control studies identified minimally invasive oesophagectomy as a safe alternative to open techniques, the clear benefit remained subject to debate. Recently, this controversy has partially resolved due to the results of the first randomized controlled trial that compared both techniques. In this trial short-term benefits of minimally invasive oesophagectomy were demonstrated in terms of lower incidence of pulmonary infections, shorter hospital stay and better postoperative quality of life. However, the current lack of long-term data on recurrence rate and overall survival precludes a comprehensive comparison of minimally invasive and open oesophagectomy. Proclaiming minimally invasive oesophagectomy as the standard of care for patients with resectable oesophageal cancer would therefore be a premature decision. Ó 2013 Elsevier Ltd. All rights reserved.
* Corresponding author. Tel.: þ31 20 5662766; fax: þ31 20 5669243. E-mail addresses:
[email protected] (M.C.J. Anderegg),
[email protected] (S.S. Gisbertz), m.i.vanbergehenegouwen@ amc.nl (M.I. van Berge Henegouwen). 1 Tel.: þ31 20 5665953; fax: þ31 20 5669243. 2 Tel.: þ31 20 5661480; fax: þ31 20 5669243. 1521-6918/$ – see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.bpg.2013.11.002
42
M.C.J. Anderegg et al. / Best Practice & Research Clinical Gastroenterology 28 (2014) 41–52
Introduction Due to a steadily increasing incidence, oesophageal cancer is by now the eighth most common malignancy in the world amounting to nearly half a million new patients annually [1]. Since the majority has advanced, inoperable or metastatic disease at the time of diagnosis, less than 50% of patients are eligible for curative treatment [2,3]. Oesophagectomy is the mainstay of this curative treatment, but the procedure is associated with a considerable risk of (severe) complications and the highest mortality rate among all elective gastrointestinal surgical interventions [4]. In an attempt to reduce complication- and mortality rates progress has been made in patient selection, perioperative care and surgical techniques. Among the technical advances minimally invasive oesophagectomy represents the most important one in terms of utilization and scientific foundation. Minimally invasive oesophagectomy was introduced in 1992, when the first report on thoracoscopic oesophagectomy was published by Cuschieri et al [5]. From that moment, new developments in the field of oesophageal surgery rapidly succeeded each other. A laparoscopic transhiatal approach was introduced in 1994 by Sadanga et al [6] and in 1999 Kawahara et al demonstrated the feasibility of combining an oesophageal resection with an extended lymphadenectomy in a video-assisted thoracoscopic surgical (VATS) setting [7]. By now, multiple minimally invasive ways to perform an oesophageal resection have been developed for both the transthoracic and the transhiatal approach. The extent to which they are used has increased dramatically, as was clearly shown by a nationwide study from England in 2010. This review of 18,673 oesophagectomies performed over 12 years revealed that the use of minimally invasive techniques had risen from 0.6% in 1996 to 24.7% in 2009 [8]. In the Netherlands, the national upper gastrointestinal cancer registry has shown that in 2012 41% (37% in 2011) of the oncologic oesophageal resections was performed in a minimally invasive way [9]. Along with this rise in utilization came a steady increase in scientific output on this topic resulting in the embracement of minimally invasive techniques in national guidelines on oesophageal cancer [10,11]. Despite the growing interest in minimally invasive oesophagectomy, concern about clinical safety and oncologic efficacy are still under discussion. In this review we address this concern by providing an overview of the literature on minimally invasive surgery for oesophageal cancer with respect to commonly used techniques, patient selection and (post)operative outcomes.
Techniques Traditionally, a distinction in open techniques is drawn between the transhiatal oesophagectomy and the two main transthoracic oesophagectomies: the 2-incisional ‘Ivor-Lewis’ approach and the 3incisional ‘McKeown’ approach [3,12]. Choice of technique depends on tumour location, extent of lymphadenectomy and surgeon’s preference. Despite the theoretical advantage of transthoracic resection regarding extended (mediastinal) lymphadenectomy and a wider circumferential resection margin, consensus on the ideal approach has not been reached yet. In the largest randomized trial on this issue Hulscher et al assigned 220 patients with an adenocarcinoma of the mid/distal oesophagus or gastric cardia (involving the gastroesophageal junction) to open transhiatal oesophagectomy or open transthoracic oesophagectomy with extended en-bloc lymph node dissection [13,14]. Transhiatal resection was associated with a shorter operative time, lower median blood loss, fewer pulmonary complications, decreased chylous leaks, shorter duration of mechanical ventilation and shorter stay in the intensive care unit and hospital. No significant difference in in-hospital mortality was observed [13]. In a follow-up study of this trial including the complete five-year survival data no significant survival benefit of the transthoracic resection was observed. However, in subgroup analyses based on tumour location and number of positive lymph nodes a five-year survival benefit of 14% (51% vs. 37%, P ¼ 0.33) and 41% (23% vs. 64%, P ¼ 0.02) was seen with the transthoracic approach for patients with mid/distal oesophageal malignancies and patients with 1–8 positive lymph nodes in the resection specimen, respectively [14]. Based on these results we currently consider transthoracic resection as the standard surgical treatment for all oesophageal tumours with the exception of gastroesophageal junction tumours without intrathoracic lymphadenopathy or patients that are unfit to undergo a transthoracic dissection.
M.C.J. Anderegg et al. / Best Practice & Research Clinical Gastroenterology 28 (2014) 41–52
43
Table 1 Overview of surgical techniques in the treatment of oesophageal cancer. Technique Open transhiatal
Open transthoracic (Ivor-Lewis)
Open transthoracic (McKeown)
Minimally invasive transhiatal Minimally invasive transthoracic (Ivor-Lewis)
Minimally invasive transthoracic (McKeown)
Indication - Gastroesophageal junction tumours without intrathoracic lymphadenopathy - Distal oesophageal tumours - Gastroesophageal junction tumours with intrathoracic lymphadenopathy - Oesophageal tumours - Gastroesophageal junction tumours with intrathoracic lymphadenopathy - Gastroesophageal junction tumours without intrathoracic lymphadenopathy - Distal oesophageal tumours - Gastroesophageal junction tumours with intrathoracic lymphadenopathy - Oesophageal tumours - Gastroesophageal junction tumours with intrathoracic lymphadenopathy
Thoracic part
Abdominal part
Location of anastomosis
n/a
Open
Cervical
Open
Open
Intrathoracic
Open
Open
Cervical
n/a
Laparoscopic
Cervical
Thoracoscopic
Laparoscopic
Intrathoracic
Thoracoscopic
Laparoscopic
Cervical
As shown in Table 1, minimally invasive alternatives exist for the three open techniques. In this section we will discuss these minimally invasive techniques after a brief recapitulation of the traditional open procedure.
Transhiatal oesophagectomy Classic transhiatal oesophagectomy involves an upper midline laparotomy to mobilize the stomach and to dissect paracardial, lesser-curvature, left-gastric-artery, celiac trunc, common-hepatic-artery and splenic artery lymph nodes. Subsequently, the thoracic oesophagus and the distal paraoesophageal lymph nodes are bluntly dissected through the diaphragmatic hiatus. Further mobilization of the oesophagus is performed through a left neck incision after which a cervical anastomosis is created with a gastric pull-up approach. During the minimally invasive variant of this procedure the patient is operated in supine position with a five-port laparoscopy set-up. Through the diaphragmatic hiatus dissection of the oesophagus and paraoesophageal lymph nodes is performed up to the level of the inferior pulmonary vein. After mobilization of the stomach and the described lymphadenectomy a left cervical incision is made to dissect the proximal oesophagus. As soon as the cervical oesophagus is divided the resection specimen is stripped en retrieved via a 5–6 cm accessory supra-umbilical incision. The next step is extracorporeal generation of the gastric conduit, which is brought up prevertebrally to the cervical region where a hand-sewn or stapled anastomosis is created.
Ivor-Lewis transthoracic oesophagectomy During the traditional Ivor-Lewis oesophagectomy a midline laparotomy is combined with a right posterolateral thoracotomy in the lateral decubitus position. This approach permits direct visualization of the thoracic oesophagus and allows the surgeon to perform a full thoracic lymphadenectomy. Since an intrathoracic anastomosis is created, this technique is only indicated for distal tumours because of the limited proximal margin that can be achieved in case of tumours of the mid-oesophagus.
44
M.C.J. Anderegg et al. / Best Practice & Research Clinical Gastroenterology 28 (2014) 41–52
Fig. 1. Thoraco-laparoscopic Ivor-Lewis oesophagectomy. (a). Dissection of the celiac trunk. (A): hepatic artery. (B): splenic artery. (C): left-gastric-artery. (D): caval vein. (E): caudate lobe of the liver. (F): pancreas. (G): oesophageal hiatus (b). Dissection of the oesophageal hiatus. (A): oesophagus. (B): oesophageal hiatus (c). Formation of the gastric conduit using a linear stapler. (d). Right thoracoscopic view in prone position. Dorsally the aorta is visible (A). (B): clipped thoracic duct at diaphragm level. (C): left main bronchus. (D): oesophageal specimen e. View of the carina (A) after lymfadenectomy of the subcarinal nodes (B). The resected subcarinal nodes are retracted and visible at the right sight of the picture (C). (D): left atrium (f). Creation of the end-to-side intrathoracic anastomosis using a circular stapler.
Comparable to the described transhiatal technique the minimally invasive Ivor-Lewis oesophagectomy starts with five-port laparoscopy in supine position. After gastric mobilization and identical abdominal lymphadenectomy the gastric conduit is created by use of an endoscopic linear stapler (Fig. 1(a)–(c)). As soon as the phrenoesophageal ligament is divided, the abdominal part of the procedure is completed and the patient is repositioned in either the left lateral decubitus position or the prone position (see below). The procedure is continued with four port thoracoscopy. The first thoracic step is the division of the pulmonary ligament followed by circumferential mobilization of the oesophagus, division of the azygos vein and dissection of paraoesophageal, lower and middle mediastinal, subcarinal and right-sided paratracheal lymph nodes (Fig. 1(d)–(e)). When the gastric conduit has been brought into the thorax, the oesophagus is divided just superior to the level of the carina and an intrathoracic anastomosis can be generated (Fig.1(f)). In order to do so transoral and transthoracic staplers have been developed. Based on a
M.C.J. Anderegg et al. / Best Practice & Research Clinical Gastroenterology 28 (2014) 41–52
45
recent review comparing these two methods in terms of anastomotic leakage- and anastomotic stenosis rates both techniques can be considered equally safe and effective [15]. In the described procedure both the abdominal and thoracic stage are performed in a minimally invasive way. However, the combination of open and endoscopic techniques for transthoracic resection has also been described. When a combination of open and endoscopic approaches is used the procedure is classified as a hybrid minimally invasive technique [16]. McKeown transthoracic oesophagectomy The 3-incisional McKeown oesophagectomy combines features of the transhiatal- and the IvorLewis transthoracic technique. The abdominal and thoracic stage of the procedure are comparable to previously described Ivor-Lewis technique and allow the surgeon to perform the same two-field (upper abdominal and mediastinal) lymphadenectomy under direct vision. The main difference however, is the addition of a left cervical incision in order to generate a cervical anastomosis. Although robust scientific evidence is lacking, cervical reconstruction is considered to be associated with clinical advantages compared to an intrathoracic anastomosis like easier management of leakage and wider proximal resection margins. Supposed disadvantages include a higher rate of anastomotic leakage because of an extended gastric conduit and a higher rate of anastomotic stenosis [17,18]. The thoracoscopic and laparoscopic part of the minimally invasive McKeown technique are comparable to the descriptions above. However, the procedure usually starts with a thoracic stage to avoid the need for extra repositioning. Removal of the resection specimen and construction of the gastric conduit usually occurs through an accessory upper midline incision of 5 cm. Subsequently the gastric conduit is delivered to the cervical region where again a hand-sewn or stapled anastomosis can be generated. Parallel to the minimally invasive Ivor-Lewis approach, hybrid minimally invasive McKeown procedures can be performed. Robot-assisted oesophagectomy A striking example of ongoing technical advances in minimally invasive surgery is the introduction of robot-assisted oesophagectomy, allowing three-dimensional visualization, improved magnification, and a greater range of instrument motion [19]. Robotic assistance has been described for gastric mobilization (in both transhiatal and transthoracic resections), mediastinal lymphadenectomy, dissection of the oesophagus and generation of an intrathoracic anastomosis [20]. A potential pitfall of robot-assisted thoracic surgery is the need for single-lung ventilation (see below). However, preliminary studies showing equality with above mentioned techniques in terms of safety and efficacy have encouraged researchers to initiate the first randomized controlled trial comparing robot-assisted thoracolaparoscopic oesophagectomy and open transthoracic oesophagectomy for resectable oesophageal cancer (ROBOT trial) [21,22]. Patient positioning: lateral decubitus position versus prone position The main advantage of the prone position is increased visualization because gravity causes the right lung to fall away from the surgical field without the need of additional retraction [23]. It has been suggested that the prone position is also associated with a lower incidence of pulmonary infection compared to the lateral decubitus position because total lung collapse resulting from single-lung ventilation can be avoided [24]. However, the downside is that putting the patient in prone position complicates airway management by the anaesthesiologist and could delay emergency conversion to open thoracotomy [25]. Patient selection Given the relatively high risk of surgery related morbidity, adequate patient selection is essential in both conventional and minimally invasive oesophagectomy. Pre-treatment staging using imaging techniques like endoscopic ultrasonography, computed tomography and positron emission
46
Study
Surgical technique
Number of included studies
Minimal number of patients
Blood loss (WMD)
Operative time (WMD)
Total morbidity (OR)
Major morbidity (OR)
Pulmonary complications (OR)
RLNP (OR)
ICU stay (WMD)
Hospital stay (WMD)
30 dayMortality
Biere-2009
MI TTO open TTO MI TTO open TTO MI O open O MI TTO open TTO
2
148
NR
NR
NR
NR
NR
163
NR
NR
0.57 (P ¼ 0.52)
NR
NR
5
216 243
5.91 (P ¼ 0.80)b NR
0.76 (P ¼ 0.71)
2
268.53a (P < 0.001)b NR
1.93a (P < 0.05) 0.52a (P ¼ 0.007) NR
1.05 (P ¼ 0.91) 1.31 (P ¼ 0.73) 0.58a (P ¼ 0.04) NR
NR
5
0.88 (P ¼ 0.78) NR
0.97a (P < 0.001) NR
2.75a (P ¼ 0.004) NR
OR 0.58 (P ¼ 0.64) RR 1.45 (P ¼ 0.47) OR 0.55 (P ¼ 0.26) HR 0.87 (P ¼ 0.88)
Sgourakis-2010 Nagpal-2010 Dantoc-2012
NR NR
NR
HR ¼ hazard ratio; ICU ¼ Intensive Care Unit; MI ¼ minimally invasive; O ¼ oesophagectomy (both transthoracic and transhiatal); OR ¼ odds ratio; NR ¼ not registered; RLNP ¼ recurrent laryngeal nerve palsy; TTO ¼ transthoracic oesophagectomy; RR ¼ relative risk; WMD ¼ weighted mean difference. a Significant result in favour of minimally invasive oesophagectomy. b Significant heterogeneity in the studies reporting this outcome.
M.C.J. Anderegg et al. / Best Practice & Research Clinical Gastroenterology 28 (2014) 41–52
Table 2 Overview of meta-analyses comparing operative and short-term outcomes between open and total minimally invasive oesophagectomy.
M.C.J. Anderegg et al. / Best Practice & Research Clinical Gastroenterology 28 (2014) 41–52
47
Table 3 Summary of outcomes of the first published randomised controlled trial comparing open and minimally invasive oesophagectomy (TIME trial) [32].
Blood loss (median) Operative time (median) ICU stay (median) Pulmonary infection within two weeks primary outcome Pulmonary infection in-hospital primary outcome Anastomotic leakage Vocal cord paralysis Reoperations Hospital stay (median) In-hospital mortality Radicality Radical (R0) resection Irradical (R1) resection Number of lymph nodes retrieved (median)
MI TTO (n ¼ 59)
Open TTO (n ¼ 56)
P value
200 mL 329 min 1 day 9% 12% 12% 2% 14% 11 days 3%
475 mL 299 min 1 day 29% 34% 7% 14% 11% 14 days 2%