To date, existing studies showed that laparoscopic D2 radical gastrectomy is safe and feasible[14–15], and the long-term efficacy is comparable to that of open gastrectomy. Compared with traditional open gastrectomy, laparoscopic gastrectomy have some advantages including minimally invasiveness, less bleeding, less pain, faster recovery and shorter hospital stay[17–19]. However, with the wide application of laparoscopic technology in clinical practice, limitations were exposed gradually, including two-dimensional views lacking hierarchical sense, non-articulated instruments that have difficulties to operate in deep narrow space, and longer learning curve[20, 21]. These factors hinder the application of laparoscopy in complex surgeries.
With three-dimensional views, wristed instruments and tremor filtration technology, robotic surgery fully shows its advantages of minimally invasiveness, accuracy and flexibility in the operation of advanced gastric cancer, and overcomes some limitation of laparoscopic surgery. In this study, we compared the short-term clinical efficacy of robotic versus laparoscopic gastrectomy, and found that there were two advantages in robotic gastrectomy.
First, more lymph nodes were dissected in the robotic group than in the laparoscopic group, especially the suprapancreatic nodes.
We collected 210 consecutive cases of locally advanced gastric cancer that received D2 radical resection and stratified the two types of gastrectomy with distal gastrectomy and total gastrectomy. The number of lymph nodes, lymph nodes at each station, lymph nodes in each region and total number of harvested lymph nodes were analyzed. We found that whether it was in distal gastrectomy or total gastrectomy. The number of lymph nodes dissected by surgical robots in the upper pancreatic region and the station 2 was higher (P = 0.01; 0.01; 0.043; 0.03). With BMI 25kg/m2 as the boundary, further stratified analysis was conducted: whether or not the patients were overweight, the robotic group was associated with more harvested lymph nodes(P = 0.001, 0.036). We believe that in D2 radical gastrectomy, robotic surgery had significant advantage in the number of suprapancreatic lymph node dissection compared with the laparoscopic surgery.
The main reasons are: (1) Flat two-dimensional vision has no stereoscopic sense, and tremor of surgeon’s hand caused by long-term operation leads to unstable surgical view and operation. However, lymphatic drainage channels in advanced gastric cancer is extensive, and metastatic lymph nodes mainly distribute around blood vessels, while vascular anatomy is complex and is often accompanied by variations. The large volume of bleeding following improper handling of vessels anatomy will affects the operation and increases the difficulty, and such outcomes will not only challenge the surgeon's confidence seriously in the operation, but also hinder precise dissection.
(2) The non-articulated arms of laparoscopic device lacks flexibility and the operation area is limited. With the conventional straight forceps used in conventional laparoscopic surgery, the surgical field is limited and it is difficult for the surgeon to reach these deep areas during dissection of N0.8a lymph nodes, N0.11P lymph nodes and N0.12a lymph nodes. In addition, convex body of the pancreas often interfere with the laparoscopic device and obstruct surgeon from precise dissection, and affect the dissection of lymph nodes. (3) In laparoscopic surgery, it is difficult for assistants to cooperate well with the surgeon, and to achieve "proper" relative tension of tissue. Excessive traction of the lymph nodes and adipose tissue of the upper pancreas often leads to laceration, bleeding and inadequate dissection of lymph node. While robotic surgery can provide a higher-definition, three-dimensional and stable vision with better depth perception in the surgical area, which is conducive to the precise clearance of perivascular lymph nodes; the instrument arms of robots are flexible and tremor-filtering, and in the narrow operating space, the free rotation of the robot arms is undoubtedly an expansion of space, which makes it relatively easy to dissect perigastric lymph nodes and adipose tissue, especially the lymph nodes in the upper and posterior area of the pancreas which are difficult to reach by traditional laparoscopic instruments. In addition, the autonomous control the 2nd robotic arm and 3rd robotic arm can greatly reduce the surgeon’s dependence on the cooperation of assistants and reduce the incidence of damage.
Kim et al studied 375 cases of radical distal gastrectomy for gastric cancer and found that the number of lymph node dissections in the robotic group and the laparoscopic group was 37.1 and 34.1, respectively. The number of station 2 lymph node dissections in the robotic group was significantly higher than that in the laparoscopic group(16.3 vs 13.2). It was considered that the robotic group had advantage in station 2 lymph node dissection. Shen et al. compared 93 cases of robotic-assisted gastrectomy with 330 cases of laparoscopic-assisted gastrectomy for lymph node dissection (33 vs 31.3 P = 0.047). Huang K. et al. compared open, laparoscopic and robotic gastrectomy (n = 586, 64 and 39 cases), and found that the number of lymph node dissection in robotic gastrectomy and open gastrectomy was similar, and was more than that in laparoscopic gastrectomy. Huang K. et al believed that robotic gastrectomy was associated with easy operation in perigastric lymph node dissection compared with laparoscopic gastrectomy, especially in the soprapancreatic areas and subpyloric regions. Our findings were comparable to the results of Huang K. et al.
Second, there was no significant difference in the rate of pancreatic fistula between the two groups, while the robotic group showed less pancreatic injury.
In this study, the robotic group reported 7 cases of postoperative pancreatic fistula (7.4%) in comparison with 18 cases (15.7%) in the laparoscopic group. There was no statistical difference between the groups. There was no significant difference between severe POPF grade B and C. However, there was a statistical difference in biological leakage which needs no additional therapeutic intervention, with 5 cases (5.3%) in the robotic group and 16 cases (13.9%) in the laparoscopic group.
The robotic group was associated with lower levels of blood amylase and drainage fluid amylase on postoperative day 1, day 2, day 3 and day 5 when compared with the laparoscopic group. There was significant difference in amylase levels of drainage fluid on postoperative day 1 and in levels of blood amylase on postoperative day 1 and day 3 between the groups. It can be seen that the robotic and laparoscopic surgery were equally safe for lymph node dissection in the surface of the pancreas, and the robotic surgery had less trauma to the pancreas.
Considerations are as follows: (1) In laparoscopic surgery, with the absence of effective tactile feedback, it is difficult for the surgeon to identify the pancreatic boundary clearly and accurately only by the flat two-dimensional vision provided by the laparoscope. In addition, the pancreas is fragile in texture, rough in surface and irregular in shape, so it is difficult to prevent pancreatic injury. However, surgical robots have 3D views with strong depth perception, real vision and high resolution, so the surgical view of robotic surgery is more closer to the view in open surgery, and can effectively overcome the limitation of insufficient tactile feedback and improve positioning accuracy. Therefore, during the operation in the upper pancreatic region, the surgeon can better identify the local tiny anatomical gap and reduce the occurrence of iatrogenic trauma . (2) D2 radical gastrectomy for gastric cancer is indeed a anatomy conducted around the upper edge of the pancreas. The location of pancreatic tissue is fixed. In laparoscopic surgery, difficulties exposed when dissecting lymph nodes on the upper edge of the pancreas, especially No. 8p lymph nodes along the common hepatic artery, No. 11p lymph nodes along the proximal splenic artery, and No. 12a lymph nodes in hepatogastric ligament. A high tail in the pancreas, high BMI, and excess visceral fat all increase the difficulty of surgical field exposure and lymph node dissection.
In order to meet the operation requirement of the expert consensus on quality control of the laparoscopic radical resection for gastric cancer in China (2017 edition), it is often necessary to press the pancreas to the lower left to expose the deep lymphatic adipose tissue as much as possible. Even though the laparoscopic group routinely uses T-shaped gauze to protect the pancreas and reduce the superficial tear of the pancreas, it was difficult to avoid direct compression caused by the tip of forceps, which resulted in deep injury of pancreatic tissue, and rupture of pancreatic acinar, trypsin-induced bleeding, pancreatitis and even pancreatic fistula and other complications are caused by pancreatic acinar rupture and pancreatic enzyme activation. At the same time, the surgeons were worried about postoperative pancreatitis or pancreatic fistula caused by injury of pancreas due to excessive compression, therefore they had no choice but to reduce the number of lymph node dissection, which explained why the number of harvested lymph node in the laparoscopic group was significantly less than that in the robotic group. During lymph node dissection in the deep area, the flexible arm of the robot can be lifted, therefore directed compression to the pancreas can be avoided. Thus all the lymph node can be dissected with less pancreatic injury.
Kumagai et al. believed that compression caused by forceps is a risk factor for postoperative fistula in laparoscopic gastrectomy. Satoshi et al. made it clear that long-time direct compression caused by assistant’s forceps in a swine model would lead to pancreatic injury and pancreatic leakage. Tsujiura et al. reached similar conclusions.
In addition, the heat conduction of ultrasound scalpel was also a major factor of pancreatic injury. During lymph node dissection on the upper edge of the pancreas, it is more convenient for robot to grasp the tissue and reduce the direct thermal damage of the functional surface of ultrasound scalpel to the pancreas. Postoperative peritoneal drainage in the robotic group was also less than that in the laparoscopic group, and there was a significant difference in the amount of peritoneal drainage fluid on the first day, suggesting that the robotic group had a better treatment for the exudation of surgical field wounds and lymphatic leakage.
Seo et al compared 40 cases of robot-assisted distal gastrectomy(RADG) with 40 cases of laparoscopy-assisted distal gastrectomy(LADG). They found that the incidence of postoperative fistula was lower in the RADG than that in the LADG group (10% vs 22%), and the difference was statistically significant according to the univariate analysis of risk factors for postoperative fistula. Suda et al. performed radical gastrectomy on 520 patients (robotic group n = 88, laparoscopic group n = 438), and found that there was no pancreatic fistula in the robotic group in their study (robotic 0% vs conventional lap 4.3%, p = 0.029).
This study had some limitations:
This was a single-center study without matching research, and the sample size of selected cases was relatively small, which may have caused selection bias.
There was no further stratified analysis of the data according to T-staging, therefore the result might be a little biased.
This was a retrospective but not a prospective randomized controlled study, so the argument strength was comparatively weak.
There was no further analysis of the long-term clinical outcome of robotic surgery in oncology.
In conclusion, in D2 radical gastrectomy for locally advanced gastric cancer, robotic gastrectomy was associated with lower intraoperative blood loss compared with traditional laparoscopic gastrectomy, and it revealed more harvested lymph nodes especially in the suprapancreatic area. Meanwhile, robotic gastrectomy showed less damage to the pancreas.