In this study, we retrospectively evaluated 22 patients who underwent LPG (DT: 16 patients, JIP: 6 patients) and 30 patients who underwent LTG from April 2014 to March 2019. All of these cases were followed up at our outpatient clinic, and upper gastrointestinal endoscopy (UGE) was performed within one year after surgery. If the patient complained of dysphagia, UGE was performed as quickly as possible using a 9.9-mm-diameter endoscope (GIF-Q260J; OLYMPUS Co., Ltd., Tokyo, Japan). If the endoscope could not be passed through the esophagojejunostomy, endoscopic balloon dilatation (EBD) was performed. Briefly, a balloon catheter (CRE wire-guided balloon dilation catheter; Boston Scientific, Natick, MA, USA) was inserted through the endoscope, and the 15/18-mm-diameter balloon was gradually filled with water and maintained at an adequate pressure for 1-2 min. After the dilatation, the endoscope was confirmed to easily pass through the anastomotic site without force. If restenosis was recognized, balloon dilatation was repeated until the stenosis was adequately treated.
Since we opined that the anastomotic stricture is caused by tension on the esophageal and jejunal walls, in July 2015 we decided to perform an additional intraoperative procedure to release this tension. The period before we started performing our tension-relieving procedure was considered as the early phase, and the subsequent period was regarded as the later phase. In addition, as described below, since JIP only crimps the jejunum on the caudal side of the gastrojejunostomy during DT, with the procedure for esophageal jejunostomy being the same, we compared cases of LPG reconstructed using both JIP or DT versus cases of LTG alone. Anastomotic leakage as a complication related to anastomosis and pancreatic fistula as a complication related to lymph node dissection were also investigated. Our surgical procedures and the study protocol were approved by the Human Ethics Review Committees of Tokai University School of Medicine (Institutional Review Board numbers 14R043). Written, informed consent was obtained from each enrolled patient before the surgery, in accordance with the Declaration of Helsinki.
Laparoscopic procedure
Patients are placed in the reverse Trendelenburg position with their legs apart. The surgeon stands between the patient's legs, the first assistant (standing on the left side) handles the laparoscope with the left hand and assists the operator with his right hand. The second assistant (standing on the right side) assists the operator with his right hand. A vertical incision about 2 cm in length is made at the umbilicus for insertion of a 12 mm port, and the laparoscope is inserted into the port after creation of the pneumoperitoneum. Next, a 5 mm port is inserted on the right and a 12 mm port is inserted on the left, at a position about 6-8 cm lateral to the navel. In addition, a 5 mm port is inserted on the right and a 12 mm port on the left slightly above and 12-14 cm lateral to the navel. Nathanson's retractor inserted just caudal to the xiphoid process is used to elevate the round ligament and the lateral segment of the liver. In all our cases, LPG and LTG were performed with D1+ lymph node dissection according to Japanese treatment guidelines [17].
LPG
DT method
In order to resect half to one-third of the proximal stomach, gastric transection was performed using a linear stapler, with the line of dissection extending orally from the pylorus to 10-12 cm along the lesser curvature and 15-17 cm along the greater curvature. The esophagus was transected slightly oblique from the right to the left side with a linear stapler. At this time, proximal gastrectomy was completed. The umbilical wound was extended longitudinally to a length of 4 cm, and a wound retractor (Alexis Wound Retractor M, Applied Medical, Rancho Santa Margarita, CA, USA) was inserted. The gastrectomy specimen was pulled out extracorporeally through the minilaparotomy and the proximal margin was evaluated. A surgical glove was attached to the wound protector, and pneumoperitoneum was re-established. Next, the OrVilTM system (DST EEA 25; Covidien) was inserted orally toward the right edge of the esophageal stump along with its anvil. Further, the mesentery supporting a section of the jejunum between 20-30 cm from the ligament of Treitz was dissected along with the corresponding jejunum (Fig. 1a), using a linear stapler for the distal jejunal dissection. The jejunum was then pulled out through the minilaparotomy. Next, the head of the shaft of the CS (DST EEA XLTM shaft (Covidien)) passing through a surgical glove was inserted orally through the incised jejunum that had been pulled out from the abdomen, 10-15 cm from the oral end of the jejunal stump, and the central rod was introduced from just at the caudal side of the jejunal stump (Fig. 1b). At this time, a rubber band was used to secure the jejunum to the head of the shaft, to prevent slippage of the jejunal stump from the shaft. A surgical glove was attached to the wound protector, and pneumoperitoneum was re-established. The jejunal stump and the head of the shaft were simultaneously re-inserted into the abdominal cavity (at this time, these were observed using a laparoscope inserted via the 12 mm port on the left flank.), and anastomosis was performed under a good visual field (HDST, Fig. 1c). Further, anastomosis was performed using Albert-Lembert sutures between the insertion hole of the CS and the oral edge of the remnant stomach to complete jejunogastrostomy under direct vision. The resultant anastomotic diameter was approximately 6 cm. Finally, a side-to-side jejunojejunostomy 20 cm from the caudal end of the jejunogastrostomy was performed extracorporeally (Fig. 1d). Petersen’s defect and the mesenteric gap were closed intracorporeally.
JIP method
After the DT reconstruction was completed, the jejunum on the caudal side of the jejunogastrostomy was crimped with a knifeless linear stapler to achieve reconstruction by the JIP method.
LTG
Roux en Y method
After the procedures on the side of the greater curvature along with D1+ lymph node dissection were performed, transection of the duodenum 1 cm away from the pyloric ring was performed using a linear stapler. After performing the procedures on the side of the lesser curvature along with D1+ lymph node dissection, the esophagus was slightly obliquely transected from the right to the left side with a linear stapler. At this time, total gastrectomy was completed. The umbilical wound was extended longitudinally to a length of 4 cm, and a wound retractor was inserted. The specimen was pulled out extracorporeally through the minilaparotomy and the proximal margin was evaluated. A surgical glove was attached to the wound protector, and pneumoperitoneum was re-established. Then, the OrVil system (DST EEA 25; Covidien) was inserted orally toward the right edge of the esophageal stump and its anvil was positioned. Further, the mesentery supporting a section of the jejunum between 20-30 cm away from the ligament of Treitz was dissected along with the corresponding jejunum (Fig. 1e), and the jejunum was pulled out through the minilaparotomy. Then, the jejunum was dissected with a linear stapler at the oral end of the jejunum of which the mesentery was dissected, to obtain the part of the jejunum that connected with the esophageal stump. The head of the shaft of the CS was inserted by incising the oral edge of the caudal side of the jejunum of which the mesentery was dissected, and the center rod of the CS was introduced from the jejunum while maintaining its blood flow (Fig. 1f). As with LPG, to prevent slippage of the jejunal stump from the shaft, a rubber band was used to secure the sacrificed jejunum to the head of the shaft. Next, a pneumoperitoneum was created using surgical gloves, and the jejunum and the head of the shaft were simultaneously inserted into the abdominal cavity (at this time, these were observed with a laparoscope inserted via the 12 mm port in the left flank of the abdomen). Anastomosis was performed under a good visual field (HDST, Fig. 1g). Then, the shaft of the CS was removed, and the jejunum in which the shaft of the CS had been inserted was transected with a linear stapler to complete the esophagojejunostomy. In addition, a side-to-side jejunojejunostomy with a linear stapler was performed extracorporeally to create a Roux-en-Y limb (Fig. 1h). Petersen’s defect and the mesenteric gap were closed intracorporeally.
Technique to avoid anastomotic stenosis
Since it is thought that the rubber band used to fix the jejunum to the shaft of the CS (DST EEA XLTM shaft) causes tension on the jejunum, and hauling the shaft of the CS toward the caudal side causes tension on the esophagus, when the esophagus and jejunum were brought close to each other, the rubber band was cut and the jejunum on the caudal side was pushed toward the cranial side to reduce jejunal tension, and the shaft of the CS was pushed toward the cranial side to reduce esophageal tension (Fig. 2a) for both LPG (Fig. 3a) and LTG (Fig. 3b). At this time, it was necessary to be careful that the surrounding tissues were not sandwiched between the esophageal stump and lifted jejunum. As a result, the anastomotic diameter became wider (Fig. 2a). If this step was not performed, it would result in a narrower anastomotic diameter (Fig. 2b). Since we believed that performing this procedure would eliminate the tension between the esophagus and jejunum and prevent anastomotic stenosis, we decided to perform this procedure with all three reconstructive methods in cases performed after July 2015.
We retrospectively compared the incidence of anastomotic stenosis, anastomotic leakage and pancreatic fistula in cases of LPG (with DT, JIP) and LTG performed before July 2015 (early phase, 30 cases) versus those performed after this period (later phase, 22 cases).
Statistical analysis
Statistical analysis was performed using the χ2 test, Fisher’s exact test, and student’s t-test. Mann-Whitney’s U test was used for multiple comparisons. All statistical analyses were performed using the JMPÒ software program, ver. 13 (SAS Institute Inc., Cary, NC, USA). A p value of less than 0.05 was considered significant.