Robotic-assisted minimally invasive oesophagectomy for cancer: An initial experience.

BACKGROUND: The morbidity related to radical oesophagectomy can be reduced by adopting minimally invasive techniques. Over 250 thoraco-laparoscopic oesophagectomy (TLE) was done in our centre over the last 15 years, before adopting robotic surgery as the latest innovation in the field of minimally invasive surgery. Here, we share our initial experience of robotic-assisted minimally invasive oesophagectomy (RAMIE) for carcinoma oesophagus.
METHODS: A prospective observational study conducted from February to December 2017. A total of 15 patients underwent RAMIE in this period. Data regarding demography, clinical characteristics, investigations, operating techniques, and post-operative outcome were collected in detail.
RESULTS: There were 10 (66.7%) male patients and the median age of all patients was 62.9 (range 36-78) years. The median body mass index was 24.4 (range 15-32.8) kg/m2. Twelve (80.0%) patients had squamous cell carcinoma (SCC) of the oesophagus and 3 (20%) patients had adenocarcinoma (AC). Five (33.3%) patients received neoadjuvant therapy. All 15 patients underwent RAMIE. Patients with SCC underwent McKeown's procedure, and those with AC underwent Ivor Lewis procedure. Extended two-field lymphadenectomy (including total mediastinal lymphadenectomy) was done for all the patients. The median operating time was 558 (range 390-690) min and median blood loss was 145 (range 90-230) ml. There were no intra-operative adverse events, and none of them required conversion to open or total thoracolaparoscopic procedure. The most common post-operative complications were recurrent laryngeal nerve paresis (3 patients, 20.0%) and pneumonia (2 patients, 13.3%). The median hospital stay was 9 (range 7-33) days. In total, 9 (60%) patients required adjuvant treatment.
CONCLUSION: Adequate experience in TLE can help minimally invasive surgeons in easy adoption of RAMIE with satisfactory outcome.

INTRODUCTION

Minimally invasive techniques for oesophagectomy can reduce blood loss and respiratory complications with better overall survival compared to open oesophagectomy procedures.[1] Initially, with thoraco-laparoscopy and in recent years robotic oesophagectomy has gained its popularity. The first series of robotic oesophagectomy got published in 2006.[2] Over the last 10 years, there is an enormous proliferation of reports by the robotic approach.

As a traditional proponent of thoraco-laparoscopic semi-prone oesophagectomy,[3] we have recently switched to the robotic approach. We believe, the ergonomics of a robotic system coupled with our earlier experience of minimally invasive oesophagectomy, made this learning phase of robotic-assisted minimally invasive oesophagectomy (RAMIE) smoother.[45] Hereby, we share our initial experience, short-term outcomes, and technical tips/tricks of robotic oesophagectomy.

METHODS

This study is a retrospective review of our initial experience of 1-year duration. The centre is a high-volume tertiary care teaching institute for gastrointestinal (GI) surgery with an annual volume of more than 15 oesophagectomies. Robotic oesophagectomy started in early 2017 using the da Vinci Si™ surgical system (Intuitive Surgical, Inc., CA, United States). Patient population includes all the consecutive cases of mid and lower oesophageal tumours, including those involving gastro-oesophageal (GE) junction. There were no exclusions for robotic approach. All the procedures were transthoracic, with either chest or neck reconstruction. No patient was selected for purely thoraco-laparoscopy or open resection, during the said period.

All the RAMIE were done by a single surgeon with an experience of more than 250 cases of thoraco-laparoscopic oesophagectomy (TLE) over 15 years with the help of almost equally experienced assistant surgeons, anaesthetists and scrub nurses.

The whole surgical team had their basic training of robotic surgery in animal laboratory before switching into RAMIE from TLE.

The details of demography, clinical history, examination, investigations and peri-operative data were recorded. Pre-operative workup was done as per our institutional protocols, including neoadjuvant therapy. According to 8th AJCC/UICC staging of carcinoma oesophagus,[6] those patients who presented with T3 or T4a tumours (referred as ‘bulky’ lesions) with or without nodal disease were given neoadjuvant therapy followed by definitive surgery. Neoadjuvant chemoradiotherapy (NACTRT) was given for squamous cell carcinoma (SCC) while neoadjuvant chemotherapy (NACT) was given for adenocarcinoma (AC).[7] The response to neoadjuvant therapy was evaluated before surgical management. Upper GI endoscopy and contrast-enhanced computed tomography (CT) scan of the thorax and upper abdomen were done for reassessment. In some patients, where positron emission tomography-CT scans were done for restaging if it was done for initial staging.[8] The aspects of the operating technique are described below. Post-operative events, complications and follow-up visits records were noted.

The continuous variables are expressed as mean or median, with standard deviation or range. Categorical data are represented in frequencies. The analysis is performed using Microsoft™ Office™ Excel, version 2015.

Operating technique

The choice of the procedure and subsequently the initial area of dissection was based on the location of the tumour along with the histopathological diagnosis. Ivor Lewis oesophagectomy (abdomen followed by thoracic part) was performed in patients with AC of distal 1/3rd while McKeown's oesophagectomy (thoracic followed by abdominal portion followed by the neck) was offered for SCC involving distal 2/3rd of the oesophagus. Neoadjuvant therapy was given for resectable, bulky lesions or suspected node-positive disease, irrespective of the histology. Gastric conduit was used for anastomosis all the cases.

All patients underwent operation under general anaesthesia with endotracheal (single lumen) tube intubation. Thoracic dissection was done placing the patient in a left semi-prone position [Figure 1]. The abdominal part was performed with the patient supine having lower limbs abducted and 15° reverse Trendelenburg's tilt. Details of procedural steps are as follows.

Figure 1

Robotic port positions for thoracic dissections C: 12-mm Camera port at 6th ICS (Intercostal Space), R1: 8-mm first arm at 3rd ICS, R2: 8-mm second arm at 8th ICS, R3: 8 mm third arm (optional) at 10th ICS, A: Thoracoscopic assisting port position

Thoracic part

The access is gained using 5-mm Visiport™ (Medtronic, MN, United States) through right 6th intercostal space (ICS). Pneumothorax is created using carbon dioxide having the pressure at 8–10 mmHg, facilitating the collapse of the right lung. Two 8-mm working ports are placed in 3rd ICS and 10th ICS [Figure 1]. Under vision, 5-mm visiport changed to 12-mm robotic camera port. During initial cases, an additional third robotic working port was used for retraction. One 12-mm port is used for the assistant surgeon to handle suction and firing endo-stapler. Procedure proceeds with the docking of robotic arms. As a routine, monopolar hook or Maryland forceps and bipolar forceps are used in the right and left hand working port, respectively.

The first step here is to divide inferior pulmonary ligaments, followed by incision over mediastinal pleura to expose the ventral aspect of the entire thoracic oesophagus, which is mobilized subsequently from the hiatus to the carinal level. By keeping constant positive pressure, this ventral dissection pushes the pericardium down, lung parenchyma to collapse more and the oesophagus is driven away, thereby increasing the working space. The azygos vein is dissected and divided between the ligatures [Figure 2].

Figure 2

Ligation of Azygos vein A: Divided Azygos vein, L: Lung

After finishing ventral dissection, a gauze piece is kept in between oesophagus and the left bronchus to prevent iatrogenic thermal injury to later. The incision then extends toward pleura dorsal to the oesophagus. Complete meso-oesophageal dissection is facilitated by mobilizing all the peri-oesophageal tissue including pre-aortic fascia from hiatus to thoracic inlet [Figure 3]. For ease of dissection, the oesophagus is looped with umbilical tape to maintain adequate traction by the assistant surgeon. Subcarinal, right and left parabronchial nodes are taken out as a routine. As the dissection approached the superior mediastinal portion, the focus is kept removing all the lymph nodes from aorto-pulmonary window safeguarding left recurrent laryngeal nerve (RLN) [Figure 4]. For complete mediastinal lymphadenectomy, lymph nodes around right RLN are taken out [Figure 5]. Use of energy sources is avoided near RLNs. In McKeown's procedure, dissection ends at this point. In case of Ivor Lewis procedure, where abdominal dissection and conduit creation were done before thoracic dissection, specimen along with gastric conduit is pulled into the thoracic cavity. The oesophagus divided by the subcarinal region using linear cutting endostapler, and the specimen is taken out in a bag through 12-mm port incision. Then, oesophagogastric end-to-end anastomosis is done using circular endostapler, by inserting anvil through the mouth through railroad technique. Haemostasis is achieved and chest drain tube placed in the right pleural cavity. Pneumothorax is deflated and port sites closed.

Figure 3

Meso-oesophageal dissection A: Aorta, E: Oesophagus

Figure 4

Lymph nodal clearance at aorto-pulmonary window A: Aorta, LRLN: Left recurrent laryngeal nerve, LB: Left bronchus, E: Oesophagus, LBA: Left bronchial artery

Figure 5

Right paratracheal lymph nodal clearance T: Trachea, LN: Lymph nodes, RRLN: Right recurrent laryngeaal nerve, RSCA: Right subclavian artery, RV: Right vagus nerve

Abdominal part

After creating pneumoperitoneum, robotic ports are placed as depicted in Figure 6 along with one extra 12-mm port for assisting surgeon for suction, retraction, clip application and for endostapler use. For retraction of the left lobe of liver and falciform, No-1 loop polydioxanone suture is used passing it transfascially twice on either side of the left lobe of liver and anchoring it to crus of the diaphragm [Figure 7]. Robotic arms are docked. Greater curvature of the stomach is mobilized from fundus to antrum by dividing gastrocolic omentum safeguarding gastroepiploic arcade. Pars flaccida is divided and the left gastric pedicle is identified. Left gastric artery and vein are ligated separately at the root after clearing the lymph nodes. The stomach is divided 5 cm away from the GE junction, and the conduit is created using linear endostapler [Figure 7]. Nodes around the celiac axis are dissected out [Figure 8] safeguarding the right gastric artery.

Figure 6

Robotic port positions for abdominal dissections C: 12-mm robotic camera port, R1: 8-mm robotic first arm, R2: 8-mm robotic second arm, R3: 8-mm robotic third arm, A: Laparoscopic 12-mm assistant port

Figure 7

Liver retraction technique and gastric conduit preparation LLL: Left lobe of liver, FL: Falciform ligament, S: Stomach, GC: Gastric conduit

Figure 8

Celiac axis clearance CT: Celiac trunk, LGA: Ligated left gastric artery, CHA: Common hepatic artery, HAP: Hepatic artery proper, GDA: Gastro-duodenal artery, SA: Splenic artery

Neck part

This part is exclusive for McKeown's oesophagectomy. Left supraclavicular skin crease incision is made. The platysma is divided, strap muscles are separated and oesophagus identified. After mobilizing all around, the oesophagus is transected and the specimen is taken out in a plastic bag through neck incision by railroad technique. Gastric conduit is also pulled up in the neck without any undue traction or twist [Figure 9]. Oesophagogastricconduit anastomosis is done in quadrangular fashion by three firings linear cutting stapler [Figure 10]. Neck incision closed after keeping a small corrugated drain in the peri-anastomotic area.

Figure 9

Specimen delivery and gastric conduit delivery by railroad technique in a polythene bag LLL: Left lobe of liver, E: Oesophagus, GC: Gastric conduit

Figure 10

Qudrangular neck anastomosis after McKeown's Oesophagectomy a: Posterior side-to-side anastomosis E: Oesophagus, GC: Gastric conduit, b and c: Transverse firing of stapler to complete the anastomosis

RESULTS

From February 2017 to November 2017, a total of 15 patients underwent RAMIE. The details of demographic characteristics are mentioned in Table 1.

Table 1

Demographic variable (n=15)

Variable	n
Age, year (mean)	62.9 (36-78)
Sex (male:female)	2:1
BMI (kg/m2)	24.4 (15.0-32.8)
ASA (%)
II	8 (53.3)
III	7 (46.7)
Addiction (%)
None	12 (80.0)
Smoker	3 (20.0)
Comorbidities (%)
No	4 (26.7)
Yes	11 (73.3)
Presenting symptoms (%)
Dysphagia	11 (73.3)
Weight loss	3 (20.0)
Anorexia	1 (6.7)
Vomiting	1 (6.7)
Previous surgery
Abdominal	2 (CS, cholecystectomy)
Thoracic	1 (CABG)
Haemoglobin (g/dl)	11.8 (9.8-13.6)
Serum albumin (g/dl)	3.7 (3.0-5.1)
Aetiology (%)
Squamous cell carcinoma	12 (80.0) (mid-6, lower-6)
Adenocarcinoma	3 (20.0) (lower + EGJ-3)
Initial staging (%)
cT2N0	5 (33.3)
cT2N1	2 (13.3)
cT3N0	3 (20.0)
cT3N1	5 (33.3)
Neoadjuvant therapy (%)
Chemoradiotherapy	3 (20.0) (CR-1, MR-2)
Chemotherapy	2 (13.3) (MR-2)

BMI: Body mass index, ASA: American Society of Anesthesiologists, Co-morbidities: Diabetes mellitus, hypertension, coronary artery disease, hypothyroidism, CS: Caesarean section, CABG: Coronary artery bypass grafting, EGJ: Oesophagogastric junction, CR: Complete response, MR: Moderate response

The most common presenting symptom was dysphagia (73.3%) followed by weight loss, anorexia and vomiting. Twelve patients (80%) were diagnosed with SCC (mid-oesophagus-6, lower oesophagus-6) and three patients had AC of lower oesophagus involving oesophagogastric junction. The details of pre-operative staging are mentioned in Table 1. Three patients (20.0%) with SCC oesophagus underwent NACTRT and two patients (13.3%) (SCC-1, AC-1) underwent NACT. All of them underwent RAMIE, 4–6 weeks after NACTRT and 2–3 weeks after NACT. All other patients with early growth directly went to upfront surgery (RAMIE) after optimization. Twelve (80.0%) patients underwent McKeown's operation while Ivor Lewis procedure was done on three (20.0%) patients. Extended two-field lymphadenectomies were performed in all. The median operating time was 558 (range 390–690) min and median blood was 145 (range 90–230) ml. There were no intra-operative adverse events or conversion to open surgery [Table 2].

Table 2

Intra-operative factors

Operative factors	n
Type
McKeown’s	12
Ivor Lewis	3
Operating time (min)	558 (390-690)
Blood loss (ml)	145 (90-230)
Anastomosis
Site	Neck-12, thoracic-3
Type
Stapled:hand sewn	15:0
Triangular:circular	12:3
Gastric conduit	15
Intra-operative complications	None

Patients were mobilized from the 1st post-operative day along with rigorous chest physiotherapy. Feeding is started through a nasogastric tube on day 1 of surgery. Oral clear liquids were started after removal of the nasogastric tube usually on the 3rd day. Semi-solid diet was allowed from the day 5 onwards. A chest tube was removed when the lung is fully inflated on chest X-ray finding, with cessation of output below 100 ml. Among the post-operative complications [Table 3], 2 (13.3%) patients developed pneumonitis which was managed by broad-spectrum antibiotics and chest physiotherapy. Hoarseness of voice was observed in 3 (20.0%); two had unilateral RLN paresis which recovered within 2 weeks of operation, while one patient had bilateral RLN paresis which required temporary tracheostomy for 5 weeks. Other complications such as catheter-associated urinary tract infection in one and chyle leak in one patient. All of them responded well to conservative management. No patient had complications related to bleeding, anastomotic leak or wound infection. The median hospital stay was 9 (7–33) days and there was no 90 days mortality. Final histopathological examination (HPE) reported complete response in two. Details of other HPE findings are mentioned in Table 3. The average number of lymph node retrieval was 24 (14–39) and nine (60.0%) patients were advised adjuvant chemotherapy.

Table 3

Post-operative factors

Variables	n
Hospital stay (days)	9 (7-33)
Tumour stage
T0	3
T1	0
T2	5
T3	7
Lymph node stage
N0	7
N1	6
N2	3
Perineural invasion	2
Lymphovascular invasion	1
Lymph node retrieved	26 (14-39)
Complications
Pneumonia	2
RLN paresis	3 (B/L-1, U/L-2)
Urinary tract infection	1
Chyle leak	1
Anastomotic leak	0
SSI	0
30-day mortality	Nil
Adjuvant therapy	9/15

RLN: Recurrent laryngeal nerve, B/L: Bilateral, U/L: Unilateral, SSI: Surgical site infection

DISCUSSION

Oesophagectomy is technically demanding procedure, especially when performing for malignancies. TLE is well established for oesophageal cancer, and it showed improved outcomes in terms of peri-operative complications and in-hospital mortality than open radical oesophagectomy.[9] The latest addition to minimally invasive surgery is the da Vinci® surgical system (Intuitive Surgical Inc. CA, United States). The magnified three-dimensional high-definition (3D-HD) view with 7-degree freedom of movement of instruments allows the operating surgeon to do this time-consuming surgery with ease.

Our prior familiarity with minimally invasive oesophagectomy[10] has facilitated the rapid adaptability to the robotic system. Although most of the necessary steps remain same in RAMIE approach as in standard TLE, there are few modifications. In RAMIE, ports are placed a bit more on posterior ICSs and having a gap of at least 5 cm (2–3 ICS) in between to allow free movements of heavy robotic arms [Figure 1]. An extra port is used during thoracoscopic and abdominal dissection, in addition to three robotic arms. This port facilitated easy dissection and cleared blood from operating area and thus helped in reducing operating time.

The anastomotic technique used during this study (end-to-end anastomosis using the circular stapler for Ivor Lewis and triple stapled quadrangular neck anastomosis [TRIQ] for McKeown's procedure) has shown excellent results. Few other reported studies have demonstrated similar outcomes.[1112]

Operating time during this study was 558 (390–690) min, is comparable to published literature,[5] which concluded that an experienced minimally invasive surgeon can reach proficiency after 20 cases of robotic-assisted oesophagogastrectomy and can reduce operating time significantly, this is yet to be observed in present series.

The extent of lymphadenectomy is an area of controversies in oesophageal surgery, where there are ongoing debates on routine application of two-field versus three-field approach. Till utilization of robotic approach at our centre, the extended two-field lymphadenectomy was an integral part of both McKeown's and Ivor Lewis procedures.[310] With the advantage of 3D-HD view at surgeon's console and flexible robotic instruments, we opted for more aggressive total mediastinal lymphadenectomy (TML) approach, with acceptable morbidity. Even patients with a tumour stage of T3 or more, high mediastinal lesion or lymph node metastasis at presentation can be taken up for RAMIE after neoadjuvant therapy.[13] A thorough concept of the upper mediastinum is an essential prerequisite for successful TML.[14]

Optimum LN clearance along with margin negative resections through robotic approach have yielded good long-term locoregional control,[15] while few advocated, extended cervical and upper mediastinal lymphadenectomy, especially in mid-oesophageal carcinoma for transmitting benefit of reduced recurrence.[16] With the exception of superficial cancers, upper mediastinal lymphadenectomy is considered essential even for lower oesophageal SCC.[17] A study comparing thoraco-laparoscopy and robotic oesophagectomy favored latter in relation to mediastinal nodal yield without increasing morbidity due to RLN palsy or pulmonary complications.[18]

The technically challenging area of mediastinal nodal clearance lies in the removal of nodes along the RLN's. Where the overall incidence of RLN palsy, especially after McKeown's oesophagectomy varies from 0% to 58%,[19] reports suggest clearing lymph nodes from RLNs is necessary for better survival for SCC of midoesophagus.[20] However, with the introduction of minimally invasive techniques, this frequency is drastically reduced. Shen et al.[21] in his research has shown that extensive mediastinal lymphadenectomy along the RLNs during minimally invasive oesophagectomy do not increase the incidence of RLN palsy. Instead, it is oncologically safe in experienced hands. Similarly, three of our patients had neuropraxia (two unilateral, one bilateral), without any long-term sequelae. Microinvasive oesophageal resection using oesophageal suspension technique showed the minimal incidence of the left RLN palsy.[22] We too used similar oesophageal suspension technique which was an extrapolation of our method used in thoracoscopy. This technique is also useful while clearing aorto-pulmonary window.

Chyle leak is another major complication that can happen after oesophagectomy. Although a recent meta-analysis showed that routine prophylactic thoracic duct ligation could reduce post-operative chylothorax,[23] few authors have expressed concerns for routine ligation, as shown by its adverse effect on long-term survival.[2425] We do not recommend routine ligation, unless warranted, in view of involvement by the tumour or as a hindrance in radical dissection.

The authors would like to emphasize on the importance of having profound experience of TLE to transform the practice into RAMIE with ease. The detail practical knowledge of mediastinal anatomy and unavoidable difficulties during dissection while doing TLE was a great help in the learning phase of RAMIE. However, with 7-degree movements of robotic arm and 3D-HD view at surgeons console actually helped to overcome those obstacles in obtaining CML without increasing morbidities. Although a basic animal laboratory training is necessary before embarking on advanced robotic surgery like RAMIE. Above all, the importance of having an experienced dedicated team consisting of assisting surgeon, anaesthetists and scrub nurse is paramount. This study has few limitations in form of small numbers, the absence of comparator arm and short follow-up duration.

CONCLUSION

Robotic oesophagectomy (RAMIE) can facilitate complete mediastinal lymphadenectomy without increasing morbidity related to pulmonary complications or RLN injury. Adequate experience in TLE can help a minimally invasive surgeon to adopt robotic oesophagectomy comfortably with equivalent oncological and overall outcome.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.