Meta-analysis of the efficacy of Da Vinci robotic or laparoscopic distal subtotal gastrectomy in patients with gastric cancer

Abstract Background: Robotic-assisted gastrectomy has been used for treating gastric cancer since 2002. This meta-analysis was conducted to systematically evaluate the efficacy of Da Vinci robotic distal subtotal gastrectomy (RDG) or laparoscopic distal subtotal gastrectomy (LDG) in patients with gastric cancer. Methods: We conducted searches in domestic and foreign databases, and collected literature in Chinese and English on the efficacy of RDG and LDG for gastric cancer that have been published since the inception of the database. RevMan 5.4.1 was used for meta-analysis and drawing and Stata14.0 was used for publication bias analysis. Results: A total of 3293 patients in 15 studies were included, including 1193 patients in the RDG group and 2100 patients in the LDG groups respectively. The meta-analysis showed that intraoperative blood loss was significantly lower and the number of resected lymph nodes was higher in the RDG group compared to that in the LDG group. In addition, the times to first postoperative food intake and postoperative hospital stay were shortened, and there was a longer length of distal resection margin and prolonged duration of operation. No significant differences were found between the 2 groups with respect to the first postoperative anal exhaust time, length of proximal resection margin, total postoperative complication rate, postoperative anastomotic leakage rate, incidence of postoperative gastric emptying disorder, pancreatic fistula rate, recurrence rate, and mortality rate. Conclusion: RDG is a safe and feasible treatment option for gastric cancer, and it is non-inferior or even superior to LDG with respect to therapeutic efficacy and radical treatment.


Introduction
Gastric cancer, a malignant tumor of the digestive tract, is the fourth most common cancer worldwide and is more common in developing countries. [1] Surgery and chemotherapy are the mainstay of treatment for gastric cancer, and surgical resection is the predominant approach. The first laparoscopic-assisted gastrectomy (LG) was reported by Kitano et al [2] in 1994. With the development of minimally invasive surgical approach for gastric cancer, laparoscopic instruments exhibited several disadvantages, such as limited motion in a confined space, two-dimensional view, and obscure vision during surgery. Therefore, based on laparoscopic technique, Intuitive Surgical has developed the Da Vinci robotic surgical system, which was first applied for gastrectomy in 2002 by Hashizume et al. [3] Subsequently, robotic-assisted radical gastrectomy (RG) has been widely performed in various countries, and a large number of clinical studies have verified its safety and prognostic effects. Currently, there have been multiple meta-analyses on the comparison of short-term effects of LG and RG for gastric cancer, and, most of the articles included patients undergoing total gastrectomy. However, robotic distal subtotal gastrectomy (RDG) mainly refers to a surgical method of distal gastrectomy using a Da Vinci robotic device. Despite several articles that included patients undergoing distal gastrectomy, some disparities are present in some results and conclusions. In addition, with the continuous development of Da Vinci robotic surgical system, research data on RG are constantly being updated. Therefore, in this study, we performed a meta-analysis of available comparative studies to compare the efficacy of laparoscopic distal subtotal gastrectomy (LDG) and RDG for gastric cancer.

Materials and methods
Ethical approval and patient consent were not required because this was a meta-analysis of previously published studies. This meta-analysis is conducted and reported in adherence to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses).

Search strategies
We systematically searched the databases of PubMed, Embase, Web of Science, the Cochrane library, full-text database of digital journals (Wanfang data), full-text database of Chinese journals (CNKI), and other databases for the comparison of the efficacy of LDG and RDG for gastric cancer. For each database, we used the following 5 keywords: "robotic", "laparoscopy", "gastrectomy", "gastric cancer", "gastric neoplasms" in English as well as "robots", "laparoscopy", and "gastric cancer" in Chinese from the inception of each database. In addition, we expanded the scope of retrieval databases through the "extension in Chinese and English" function to avoid missed retrieval. Finally, in PubMed, 336 documents were retrieved using the keywords "robotic," "gastrectomy," and "laparoscopy;" 315 documents were retrieved using the keywords "robotic," "gastric cancer," and "laparoscopy;" 282 documents were retrieved using the keywords "robotic," "gastric neoplasms," and "laparoscopy." In Embase, 452 documents were retrieved using the keywords "robotic," "gastrectomy," and "laparoscopy;" 379 documents were retrieved using the keywords "robotic," "gastric cancer," and "laparoscopy;" and 363 documents were retrieved using the keywords "robotic," "gastric neoplasms," and "laparoscopy." In Web of science, 91 documents were retrieved using the keywords "robotic," "gastrectomy," and "laparoscopy;" 82 documents were retrieved using the keywords "robotic," "gastric cancer," and "laparoscopy;" and 17 documents were retrieved using the keywords "robotic," "gastric neoplasms," and "laparoscopy".

Inclusion and exclusion criteria
Studies were included if they fulfilled the following criteria: randomized or non-randomized controlled studies published on RDG and LDG for the treatment of gastric cancer; studies focusing only on patients with gastric cancer undergoing distal radical gastrectomy; at least 1 item of data provided on clinical efficacy comparison of RDG and LDG; studies with original data; for continuous variables, mean and standard deviation or mean and extreme values provided; for count data, the number of incidents and total number of samples provided; for binary variables, a combined odds ratio (OR) value and 95% confidence interval (CI) provided or a regression coefficient that could be converted into an OR value and 95% CI and its standard deviation provided. The following studies were excluded: comparative study of non-LDG and RDG; studies on patients only undergoing palliative subtotal gastrectomy, cytoreductive surgery, gastric volume reduction surgery, or short circuit surgery; studies without necessary comparison data; published literature with duplicated data.

Data extraction
The data were extracted independently by 2 researchers. Any disagreement was resolved through discussion or judged by the third researcher. The extracted data included the following: general information, including author, year of literature publication, research type, sample size, body mass index; outcome indicators, including duration of operation and intraoperative blood loss, the number of resected lymph nodes, length of proximal and distal resection margins, postoperative hospital stay, first postoperative anal exhaust time and time to first postoperative food intake, postoperative total complication rate, postoperative anastomotic leakage rate, postoperative gastrointestinal emptying disorder rate, postoperative pancreatic fistula rate, recurrence rate, and mortality rate.

Literature quality evaluation
In this study, the modified Newcastle Ottawa Scale recommended by Ji et al [4] was used to evaluate the quality of the selected studies. Quality evaluation was performed in 3 aspects, including the research design (I: presence or absence of randomized controlled study; II: presence or absence of inclusion criteria; III: presence or absence of a total sample size of over 100 cases), comparability (a: sex; b: age; c: body mass index; d: tumor location; e: tumor staging), and result evaluation (lymph node dissection). Each item that met the criteria was marked " * ", with a total of 9 " * "; the score of 6 " * " points and above suggested high-quality of the included studies.

Statistical analysis
RevMan 5.4.1 (The Cochrane Collaboration, Oxford, England) was used for meta-analysis and drawing. The Software Stata14.0 (StataCorp LP, Texas, United States) was used for publication bias analysis. Count data uses OR value, continuous variables were evaluated to obtain weighted mean difference (WMD) for comparison and meta-analysis, and the effect size was calculated with 95%CIs for each outcome. Statistical heterogeneity among studies was evaluated by using the I 2 and the Q tests. If the I 2 value was >50% and P < .1, it indicated heterogeneity, and a random-effects model was used for meta-analysis. We used a fixed-effects model if no heterogeneity was found between studies when I 2 < 50% and P > .1. The original data, which were represented by the median and interquartile range, were converted to the mean and standard deviation. Sensitivity analysis was performed with a one-by-one exclusion method.

Table 3
Quality scores of included articles.
Reference    studies (I 2 = 92%; P < .1); therefore, a random-effects model was used. The findings suggested that the first postoperative anal exhaust time in the RDG group was not significantly different from that in the LDG group [WMD, -0.276 (95%CI, -0.590 to 0.037); P = .08] (Fig. 8).
3.1.8. Length of proximal resection margin. Five articles [5,6,9,10,18] reported on the length of proximal resection margin, and there was heterogeneity between 2 studies (I 2 = 66%; P < .1); therefore, a random effects model was used. The results indicated that there was no difference in the length of the proximal resection margin between the RDG and the LDG groups [WMD, -0.245 (95%CI, -0.680 to 0.191); P = .27] (Fig. 9).
3.1.11. Mortality. In 6 articles, [8][9][10]12,13,18] patient quality of life was followed up for a long period of time and the mortality of patients was reported. The heterogeneity between studies was significant (I 2 = 0.0%; P > .1), and a fixed-effects model was used.
The results suggested that no significant difference was found in the mortality between the RDG group and the LDG group [OR, 0.689; 95%CI, 0.374-1.270; P = .23] (Fig. 15).

Sensitivity analysis
The one-by-one exclusion method was used to analyze the sensitivity of all outcome indicators, and no sensitivity study was shown, indicating that the results of this analysis were stable (Fig. 16).

Publishment bias analysis
The Egger test was used to analyze all the outcome indicators included in the study, and the results suggested that there was no significant publication bias in the outcome indicators of each study (Fig. 17).

Discussion
A large number of clinical studies have confirmed LDG to be a safe and radical treatment option, and the indications for surgery for early gastric cancer have been gradually expanded to include advanced gastric cancer. As a further extension of laparoscopic technique, Da Vinci robotic surgical technique has numerous advantages over traditional laparoscopic technique. Additionally, its safety and efficacy in gastric cancer surgery have also been verified by multiple studies. However, efficacy of robotic-assisted surgery, as compared with laparoscopic surgery, has not been established in high-quality studies. This study aims to provide a basis for the choice of clinical treatment via meta-analysis. Duration of operation is one of the important indicators to measure the quality and accessibility of surgery in clinical research. The results of this meta-analysis showed that the duration of operation was significantly longer in the RDG group than in the LDG group. This might be associated with the extra time required to assemble the manipulator arms and debug the equipment, and the different learning stage the surgeon is in. Currently, in laparoscopic surgery, lymph node dissection and gastrointestinal reconstruction are time-consuming. Since all included studies in this meta-analysis focus on distal gastrectomy, the time spent on laparoscopic-assisted radical gastrectomy was shorter. In addition to the above steps that were time-consuming, the Da Vinci robotic surgical system requires an additional process of instrument assembly before RDG, and studies [20] have reported that the average duration of this procedure is (62.9 ± 24.6) minutes, thereby significantly prolonging RDG. In addition,  insufficient experience of surgeons in using robotic surgical systems also leads to the prolonged operation. Kim et al [8] believe that the surgeon will be proficient in RG after the learning curve is overcome, thus shortening the duration of RG. Huang et al [21] compared surgeons in the middle and later stages of the learning curve between RG and LG, and found that the duration of RG was shorter than that of LG. Therefore, after the learning curve is overcome, duration of RG may be significantly shortened.
In addition to the duration of operation, intraoperative blood loss is the major concerns of surgeons as one of the quality indicators. The results of this meta-analysis showed that the intraoperative blood loss of RG was significantly less than that of LG. In the included studies, the mean blood loss volume was 83.08 mL in RDG group, as compared with 111.97 mL in LDG group. This is because the robotic surgical system has inherent advantages in surgical procedure compared to laparoscopy. RDG provides surgeons with the benefit of three-dimensional operative field, which was magnified by 10 to 15 times, and helps surgeons to observe the relationship between blood vessels and surrounding tissues more directly and clearly and recognize the tissue structure. In addition, the "hands" of the robotic surgical system the manipulator arm will help avoid unintentional tremor of human hands, which improves the stability and accuracy of the operation, thereby ensuring safety of dissection and ligation of gastric blood vessels. And studies have shown that robotic surgery can better dissect blood vessels in narrow surgical areas and reduce bleeding. [22] In addition, robotic surgery will make lymph node dissection around the stomach safer and more effective. It is well-known that the focus of radical gastrectomy is whether the lymph nodes within the relevant range are thoroughly cleaned. Extensive dissection of a sufficient number of lymph nodes can not only improve the accuracy of the patient's tumor node metastasis staging, but also reduce the risk of recurrence and metastasis of patients, [23] therefore, the number of resected lymph nodes is one of the important indicators of the efficacy of radical gastrectomy. This meta-analysis showed that the number of resected lymph nodes during RDG was much greater than that during LDG (25.3-41.2 vs 21.7-39.9). Therefore, RDG is superior to LDG with respect to the mean  number of resected lymph nodes, and RDG can better reduce the risk of potential recurrence caused by lymph node metastasis. Early resumption of food intake after gastrectomy for gastric cancer can help patients recover rapidly. The results of this metaanalysis showed that the time to first food intake after RDG was shorter than that after LDG. This may be partially explained by the fact that during LDG, cooperation of 2 surgeons are generally needed. Therefore, excessive intestinal canal traction may be caused due to the different traction force of the 2 surgeons in the process of tissue dissection, resulting in aggravation of intestinal stress. Blood vessels are easily damaged in this process, causing bleeding and intestinal paralysis, thereby prolonging the recovery time of intestinal peristalsis after surgery. In RDG, 4 simulated wrist instruments are available that can rotate in multiple angles in the body, which achieve the same effect of tissue traction during the operation, avoiding excessive traction of the relevant tissues, reducing the blood loss volume, and shortening the duration of intestinal paralysis. This is conducive to the recovery of intestinal peristalsis function after surgery, thereby shortening the time for the first food intake after RDG. In addition, RDG also shortens the postoperative hospital stay. The results of this meta-analysis showed that there was no significant difference in the first anal exhaust time in RDG and LDG groups, but the postoperative hospital stay in the RDG group was shorter than that in the LDG group. This indicates that given no significant difference found in the first anal exhaust time between the 2 groups, since the gastrointestinal stress of patients undergoing RDG is mild, they can consume food orally after early anal exhaust, thus speeding up the recovery of the patients after RDG and shortening the postoperative hospital stay.
In addition to the above indicators, the results of this metaanalysis indicated that the length of proximal resection margin of the RDG group was not different from that of the LDG group, but the length of distal resection margin of the RDG group was greater than that of the LDG group. The mean length of the RDG group and the LDG group was 5.3 cm and 4.6 cm respectively, which was contrary to the results of previous meta-analyses. [24] According to the latest Japanese guidelines, [25] the safe distance from resection margin to the lesion for invasive gastric cancer is over 3 cm, which shows that RDG is non-inferior or even superior to LDG in the radical resection of primary tumors.
In order to further analyze whether there was a disparity in the short-term prognosis of patients after RDG and LDG, in addition to the total postoperative complications, this meta-analysis also included the incidence of anastomotic leakage, postoperative gastrointestinal tract emptying disorder and pancreatic fistula, which showed no significant difference. In order to compare the long-term prognostic effects of the 2 procedures, this metaanalysis included postoperative recurrence rate and mortality. The longest follow-up time for postoperative recurrence rate was 40 months. There was no significant difference between the 2 surgical options, indicating that RDG was comparable to LDG in terms of safety and survival.
This meta-analysis has several advantages: New literature was included; All patients selected had undergone distal radical gastrectomy; The recurrence rate and mortality rate of patients after radical gastrectomy were systematically analyzed. There is no publication bias in the outcome indicators, and the conclusion is stable. The meta-analysis also has several limitations: All included articles are retrospective and high-quality randomized controlled studies are lacking; Some indicators are not included in the articles, which may reduce the credibility of the conclusion; The included articles showed incomplete data, and the positive lymph node dissection rate and the 5-year survival rate after surgery were not compared.

Conclusion
RDG is a safe and feasible procedure for distal gastrectomy, which is non-inferior or even superior to conventional laparoscopy. In view of the certain limitations of this study, a systematic analysis of multiple high-quality, multicenter clinical randomized controlled studies is needed to confirm this conclusion. With the development of medicine, RDG will be a promising approach for gastric cancer.