Intrathoracic side‐to‐side esophagogastrostomy with a linear stapler and barbed suture in robot‐assisted Ivor Lewis esophagectomy

Abstract Background The side‐to‐side anastomosis was considered a promising approach to create an intrathoracic esophagogastrostomy in the minimally invasive esophagectomy, with advantages over the side‐to‐end anastomosis with aspects of no need for additional mini‐thoracotomy and lower occurrence of stenosis. The hand‐sewing anterior aspect of the anastomosis is technically challenging in the thoracoscopic Ivor Lewis esophagectomy. Here we introduced our initial experience to facilitate this approach by using the surgical robot and barbed suture. Methods A retrospective study of all patients underwent robot‐assisted Ivor Lewis esophagectomy with side‐to‐side esophagogastrostomy from February 2016 to September 2018 was performed. The technical details are described in this paper. Results A total of 37 patients (35 male and 2 female, median age of 62.7 years) were successfully treated with completely robot‐assisted Ivor Lewis esophagectomy. The median total surgical time was 340 minutes including 65 minutes to perform the anastomosis. Median estimated blood loss was 120 mL and the length of hospital stay was 10 days. There was no 90‐day mortality. Three patients suffered anastomotic leakage (8.1%,3/37), who were successfully treated without reoperation. Conclusion Our initial results imply that it is technically feasible to perform intrathoracic gastroesophageal anastomosis by taking advantage of a robotic system and knotless suturing.


| INTRODUCTION
Being considered associated with reduced postoperative morbidity, the minimally invasive esophagectomy (MIE) has been widely adopted for patients with esophageal carcinoma in recent decades. 1 However, thoracoscopic minimally invasive esophagectomy has some identified weakness in the aspect of manipulation capabilities, which become troublesome when the procedure requires fine motions such as performing anastomosis via an Ivor Lewis approach. 2 The robotic system can offer high-definition 3-d vision, tremor filtration and a 7-degree articulation of the instruments, which appears useful for precise manifestation in tiny spaces. 3 advantage of these features, surgical robots seems to become the preferred surgical instrument in esophagogastric anastomosis by the means of hand-sewn, circular stapled, or linear stapled approaches. [4][5][6][7][8][9] Whereas, owing to the 3-d vision and articulated instruments, the surgical robot may provide an alternative with which to overcome these problems. Since February 2016, we have been performing minimally invasive Ivor Lewis esophagectomy (MILE) with the da Vince Si System. Here, we introduce our initial experience of intrathoracic side-to-side esophagogastrostomy with a linear stapler and report the short-term outcomes.

| Data collection
During February 2016 to September 2018, a total of 37 patients diagnosed with esophageal carcinoma in the lower third and gastroesophageal junction underwent robot-assisted Ivor-Lewis esophagectomy by one medical team experienced in both minimally invasive esophagectomy and robotic surgeries. The patient demographics and perioperative parameters were collected from the institutional database, and the short-term follow-up was completed by return visits or telephone calls. Descriptive statistics for continuous variables are reported as median (range) and categoric variable as percentage and frequency. The study protocol was approved by the applicable institution review board of West China Hospital, Sichuan University. Written informed consents were obtained from all 37 patients.

| Technical details
The operation was performed in two stages during the same period of anesthesia using a single-lumen endotracheal tube. First, in the reverse Trendelenburg position, the stomach mobilization and lymph nodes dissection were completed with a four-arm robotic platform (da Vinci Si robotic system, Intuitive Surgical, Inc.) as previously reported. 10 After stomach mobilization, a 4-cm wide gastric tube was fashioned extracorporeally through a 7-cm incision in the upper abdomen. Gastric content was emptied and the gastric cavity was three times-cleaned with diluted povidone-iodine.
In the second stage, the patient was repositioned in a left semi- Hemostasis was checked, and the right hemithorax was irrigated with normal saline. A 28-F chest tube was inserted into the thoracic cavity through the camera port, and an 18-F silicone tube was placed in postmediastinum behind the gastric conduit through the port for robotic arm 2.

| Postoperative management
Total parenteral nutrition was used in the first 4 days after operation and the oral intake began at postoperative day (POD) 3 to POD5 with a liquid diet, to which was added Methylene blue to explore potential anastomotic fistula. With the absence of leakage, the drainage tube would be removed at POD7 to POD10. All patients were discharged once oral feeding could be tolerated.

| RESULTS
The demographic, intra-and peri-operative data of all 37 patients are listed (Table 1). In summary, the median age of patients was 62.7 years and the median body mass index (BMI) was 22.1 kg/m 2 .
A total of eight patients with local advanced esophageal cancer underwent neoadjuvant therapy in this case. The median total surgical time was 340 minutes, with 165 minutes for the thoracic phase and 65 minutes required to create anastomosis. The estimated blood loss ranged from 50 mL to 160 mL, with a median of 120 mL. However, technical difficulties in endoscopic suturing may hinder the extensive adoption of this procedure. The application of a surgical robot may find a way out of this dilemma by significantly improving the surgeon's ability to suture even in deep and narrow anatomical sites such as the posterior mediastinum. Actually, the robot-assisted suturing was so facilitating that a hand-sewing anastomosis was achievable even in the posterior aspect of the anastomosis for low tension in anastomosis. 9 However, according to our own experience, both exposure and operation condition are more unsatisfactory in the posterior aspect when comparing with the anterior one. We agree with Dr. Cerfolio to staple instead of hand-sew the posterior aspect. 9 We believe that it is an acceptable compromise to create an anastomosis using a hybrid technique involving stapling the posterior wall and handsewing the anterior wall.
Triangular stapling is another anastomotic technique totally using linear staplers, which is reported being associated with a decreased rate of anastomotic complications. 12 Nevertheless, the gastric conduit and remnant esophagus need to be stapled three times in three different directions in the classic triangulating stapling technique, which is a technical challenge for intrathoracic manipulation in MIE. [13][14][15] We have several lessons when performing triangular stapling anastomosis thoracoscopically. Thus, we were not attempting triangular stapling technique in robotic intrathoracic esophagogastrostomy.
The knotless barbed suture appears to be ideal for endoscopic suturing according to the literature and our own experience. [6][7][8] First, it is self-retaining, thus liberating the surgeon from the need to hold the suture. Second, it provides knotless wound closure, which reduces the time to make knots. Furthermore, the barbed suture allows multiple points of fixation along with the closure,  After creating the posterior part of anastomosis with linear stapler, we stitched the anterior defect from two corners and combined at the midpoint, which was different from the typical endto-end modus operandi. During operation, we found that the surgical viewing became worse from the midline to the corner, and this problem was aggravated when sewing up the corner. We speculated that starting at each corner and finishing the suturing at the midline could ameliorate this problem.
There are three postoperative fistulas in our cohort: one presumed cause is the compromised blood perfusion associated with the longer esophageal stump in the side-to-side approach compared with hand-sewn or end-to-side anastomosis. Hodari 16

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.