Laparoscopic versus open approach in gastrectomy for advanced gastric cancer: a systematic review

Background Additional studies comparing laparoscopic gastrectomy (LG) versus open gastrectomy (OG) for advanced gastric cancer (AGC) have been published, and it is necessary to update the systematic review of this subject. Objective We conducted the meta-analysis to find some proof for the use of LG in AGC and evaluate whether LG is an alternative treatment for AGC. Method Randomized controlled trials (RCT) and high-quality retrospective studies (NRCT) compared LG and OG for AGC, which were published in English between January 2010 and May 2019, were search in PubMed, Embase, and Web of Knowledge by three authors independently and thoroughly. Some primary endpoints were compared between the two groups, including intraoperative time, intraoperative blood loss, harvested lymph nodes, first flatus, first oral intake, first out of bed, post-operative hospital stay, postoperative morbidity and mortality, rate of disease recurrence, and 5-year over survival (5-y OS). Besides, considering for this 10-year dramatical surgical material development between 2010 and 2019, we furtherly make the same analysis based on recent studies published between 2016 and 2019. Result Thirty-six studies were enrolled in this systematic review and meta-analysis, including 5714 cases in LAG and 6094 cases in OG. LG showed longer intraoperative time, less intraoperative blood loss, and quicker recovery after operations. The number of harvested lymph nodes, hospital mortality, and tumor recurrence were similar. Postoperative morbidity and 5-y OS favored LG. Furthermore, the systemic analysis of recent studies published between 2016 and 2019 revealed similar result. Conclusion A positive trend was indicated towards LG. LG can be performed as an alternative to OG for AGC.


Introduction
Gastric cancer is one of the most common malignant diseases worldwide; the incidence and mortality for GC is still high [1]. Surgical resection with lymph node dissection is the most effective treatment for gastric cancer [2]. In the past, conventional open gastrectomy (OG) has been the mainstay of treatment for gastric cancer. Since laparoscopy-assisted gastrectomy was first described in 1994 [3], endoscopic and laparoscopic procedures for early gastric cancer have been increasingly used because of many advantages over OG, including less blood loss, fewer postoperative complications, faster bowel function recovery, shorter hospital stay, and an equivalent longterm outcomes [4][5][6]. The application of laparoscopic gastrectomy (LG) for advanced gastric cancer (AGC) was first reported [7]. Although several clinical trials have reported the effectiveness of LG [8,9], considering of lacking of long-term oncological outcomes and the technical difficulties, there is no enough evidence to support LG for treating AGC; thus, the use of LG for AGC has been still controversial. Many previous meta-analyses have compared the short-term postoperative outcomes and longterm outcomes [10,11], whereas they have analyzed the results without enough clinical randomized trials and/or with low-quality studies. Many high-quality RCTs related with LG treating AGC have been published recently, especially between 2018 and 2019. Therefore, we conducted this systematic review and meta-analysis to find some proofs for the use of LG in AGC.

Search strategy
The comprehensive publications were identified by searching medical electronic databases PubMed, EMBASE, and Web of Science, which published from July 2010 to May 2019. The following MeSH terms and free-text terms were used: "laparoscopy-assisted gastrectomy", "laparoscopicassisted gastrectomy", "laparoscopy surgery", "laparoscopies", "laparoscopic surgery", "open gastrectomy", "conventional gastrectomy", "stomach neoplasms", "gastric cancer", "gastric neoplasm", and "stomach cancer"; the Boolean operators "AND" and "OR" were used to combined these terms. The references of the relevant articles and previous meta-analysis studies were identified as additional articles. Title and abstracts of each identified article were screened, and the full text of the screened articles was assessed for eligibility. Three authors researched and reviewed independently and thoroughly through the above-mentioned search strategy; the search strategy was provided as supplementary file.

Criteria of inclusion and exclusion
All included publications in this meta-analysis should meet following criteria: (1) Clinical studies containing RCTs and NRCTs (case-control study, and cohort study) (2) Clinical studies having compared LG versus OG for treatment of advanced gastric cancer (3) Detailed/available data of clinical studies have been reported, including short-or long-term data (4) Publication in English All papers containing any of the following criteria were excluded: (1) Duplicate publication or the publication that did not provide sufficient data (2) No OG as a control group (3) Robot-assisted gastrectomy (4) Abstract only (5) Gastric surgery performed on benign lesions, nonprimary gastric cancer, or recurrent gastric cancer (6) Patients in publication had non-curative factors such as distant metastasis of organs

Data extraction
Clinical data was extracted independently and evaluated critically by two authors. Relevant data included characteristics of included study (author, year of publication, country of publication, study design, study period, male/ female, age, tumor size, BMI, ASA (1:2:3)); summary of laparoscopic technique of included study (type of dissection, type of gastrectomy, retrieved LN, proximal margin, distal margin); systematic review of OS outcomes (follow-up (months), 5-y OS with relevant P); systematic review of recurrence pattern and sites; surgical outcomes including operative time, intraoperative blood loss, and harvested lymph nodes; recovery outcomes including time to first flatus, time to first oral intake, hospital stay, and mortality (defined as 30-day operative mortality); long-term outcomes including tumor recurrence and 5year OS; and postoperative complications were classified as morbidity, overall complications, specific complications, and general complications. General complications included pneumonia, wound problems, postoperative ileus, and pancreatitis or pancreatic leakage; specific complications included intra-abdominal bleeding, anastomotic bleed, anastomotic stenosis, anastomotic leakage, duodenal stump leakage, abdominal infection, and lymphatic fistula.

Quality assessment
To assess the quality of included studies, we used the Newcastle-Ottawa Quality Assessment Scale (NOS) for non-RCT [12]. NOS contains 3 categories including selection, comparability, and outcome, which were scaled by eight elements; high-quality elements are awarded by adding a star, no more than one stars could be added into the elements of selection and outcome, and no more than two stars could be added into the elements of comparability; then, studies were compared according to the number of stars, total score was 9 stars, 0-5 stars was considered as low-quality and 6-9 stars was considered as high quality. The risk of bias and quality of RCTs were determined by the Jadad scale (JCS) [13]. The high-quality trials should score ≥ 3 of a maximum possible score of 5.
Statistical analysis STATA 12.0 for Windows was performed for this study. Dichotomous data was calculated by relative risks (RR) with 95% confidence intervals, and continuous variables were calculated by weighted mean differences (WMD) with 95% (CI); 5-year OS was evaluated by pooled hazard ratios (HR) and their 95% CI. When the HR and 95% CI were not provided in the studies, some published formula were performed to calculate HR with 95% Cl [14]. A random effect model was used for studies with high heterogeneity, while a fixed-effect model was used for low heterogeneity. χ 2 test was used to assess heterogeneity. Funnel plots and Egger's linear regression test were used to assess the publication bias. P ≤ 0.05 was considered to indicate statistical significance.

Results of the search and quality assessment
The study selection process is summarized in the flowchart (Fig. 1). A total of 1220 publications were researched according to the search strategy, eighteen articles were excluded after duplication, and after titles, abstracts, and language were retrieved to assess further, 1140 publications were excluded. Of these studies, twenty-six studies were excluded because they included early gastric cancer only or almost early gastric cancer, were protocols of ongoing studies, were review, and were no control group. In the end, eighteen case-control studies, ten cohort studies, and eight RCT were enrolled in the studies . In terms of non-RCT studies, four studies scored 6 (moderate-quality study) on the NOS, and twenty-four studies scored 8-9 (high-quality study) (Table 1). With regard to RCT, two studies scored 4 (high-quality study) on the JCS, and six studies scored 2 (moderate-quality) ( Table 2).

Characteristics of included study
According to the search strategy and criteria of inclusion and exclusion, a total of thirty-six studies published from 2010 to 2019 were eligible for the meta-analysis. A total of 11,808 cases (5714 cases in LAG and 6094 cases in OG) were involved in the study. Among the thirty-six studies, twenty-five studies originated from China, five originated from Japan, four originated from Korea, one originated from Italy, and one originated from Chile. Detailed information for characteristics of included study is shown in Table 3.

Summary of laparoscopic technique
All the included studies have reported the laparoscopic technique. Thirty-four studies have demonstrated the details on the level of lymphadenectomy, D2 lymphadenectomy was performed in 29 studies, D1, D1+, D2, and D2 + lymphadenectomy were used in three studies, D1 + α/β and D2 were used in one study, and D0, D1, D1 + α/β, D2, and D2+ were performed in one study. All the studies have reported the type of gastrectomy; compared with proximal and subtotal gastrectomy, distal and total gastrectomy were frequently used for advanced gastric cancer. Twelve studies reported the resection margin in both LADG and ODG groups, only one article showed there was significant difference in proximal margin between the two groups [21], and all the rest indicated no significant difference between the two groups for proximal margin and distal margins. Thirty-five studies have shown detailed data of retrieved LN between the two groups, thirty-two studies indicated no significant difference, while three studies showed P < 0.05. Detailed information for characteristics of included studies is shown in Table 4. Relevant pathological characteristics of included studies are shown in Table 5.
In terms of the studies published between 2016 and 2019, eleven studies indicated significantly shorter hospital stay in LG than OG [34-38, 40-42, 44-47, 49, 50]; three studies reported the similar result between LG and OG [37,40,44]. Compared with OG, the first flatus returned earlier with statistical significance in recent ten studies for LG [34-38, 40, 41, 45, 47, 49], and three studies indicated no significant difference between LG and OG [42,44,50]. Three significant studies showed that first out of bed returned earlier in LG compared with OG [34,35,45], and one study revealed no statistical difference between the two groups [49]. For the first oral intake, six studies demonstrated a significant result for LG [34-36, 42, 45, 49], and two studies indicated LG was similar to OG [37,50].

Discussion
Between 2010 and 2019, there are many high-quality RCTs that have been published. However, many previous meta-analyses have included low-quality studies or a limited number of studies [52,53]. Thus, we conducted the meta-analysis to evaluate whether LG is an alternative treatment for AGC. Some primary endpoints were compared between LG and OG, including intraoperative LG laparoscopy gastrectomy, OG open gastrectomy a P < 0.05    time, intraoperative blood loss, harvested lymph nodes, proximal and distal resection margin distance, time to first flatus, time to first oral intake, post-operative hospital stay, complication and mortality, rate of disease recurrence, and 5-year over survival.
The LG consumed significantly more time than OG, although we could get a wider operation field by applying LG. However, the operative process is more complicated and less flexible than OG; some reasons include the narrow operating field, restriction in the number of trocar [54], lacking of tactile sensation [55], insufficient training [56], the time for setting up the equipment, and the complexity of performing the esophagojejunostomy [57], while LG combined with advanced techniques for systemic lymphadenectomy may be the main reason, which needs experienced surgeons. In terms of AGC, compared with gastrectomy alone, gastrectomy combined with systemic lymphadenectomy is more complicated. Meanwhile, compared with other laparoscopic surgery including laparoscopic colectomy and cholecystectomy, LG with lymphadenectomy is also more difficult because it is necessary to identified many important vessels and clear lymph node. Recently, some studies have indicated that the operative time could significantly reduce and reach a plateau after about 40 cases, and the operative time of LG is no longer than OG for extensive    technical expertise [51,58,59]; meantime, some highquality studies reported postoperative morbidity has no significant difference between LG and OG, with LG leading to faster postoperative recovery [60]. With the development and improvement of laparoscopic techniques, the operative time will reduce and become shorter.
In spite of the operative time is longer, blood loss is significantly less in LG. For LG, by using the laparoscopic device such as ligatures and ultrasonic scalpel, we could get enlarged surgical vision to detect large and small vessels and expose vessel adequately, which contribute to small blood loss. The small amount of blood loss may contribute to a decreased blood transfusions, which could reduce the postoperative complication such as lung injury, volume overload, and pneumonia. Thus, small amount of blood loss has an impact on postoperative recovery and recurrence [61].
The postoperative complication is usually used to evaluate the surgical safety. The meta-analysis demonstrated that the overall postoperative complication rate of LG was significantly lower than OG; meantime, the wound problem and postoperative ileus were significantly less common than OG, which is in consistent with some previous meta-analysis [62,63]. For LG, the smaller surgical surface wound and less manual handling may account for less wound problem, and LG could reduce the intervention to microenvironment of abdominal cavity and intestinal serous membrane, which may decrease the rate of postoperative ileus. The rate of postoperative pneumonia was lower than OG with no significant difference. In terms of OG, some disadvantages may make it difficult to cough, which lead to respiratory complications such as pneumonia, including tension sutures, serious pain, and abdominal bandages, while patients in LG were related with less blood loss, less blood transfusion during surgery [64,65], and less wound pain after surgery [66]. For other postoperative complications including pancreatitis/pancreatic leakage, intraabdominal bleeding, anastomotic bleeding, anastomotic stenosis, anastomotic leakage, duodenal stump leakage, abdominal infection, and lymphatic fistula, there were no significant differences.
With regard to the time to first flatus, the time to first oral intake, and post-operative hospital stay, the results were favoring for LG. LG is thought to be a less invasive procedure with smaller surgical incision and minimal gastrointestinal interference, so that the postoperative pain is less during recovery with a reduced inflammatory response and better glucose tolerance [67,68], which has a direct impact on a quick recovery of bowel function, and a quick recovery represents earlier oral intake, earlier discharge, and shorter hospital stay. In other words, it is quicker to return to normal condition in LADG than ODG. The cost of laparoscopic surgery is higher for LG compared with OG, because of the costs of the disposable instruments, while Miura et al. [69] indicated that LADG was less expensive than ODG because hospital stay is shorter and additional costs can be offset by the lower charges for ward, meals, and nursing care.
In terms of oncological safety, in most previous studies, the number of HLNs is widely considered as the index of "quality" [70][71][72][73][74][75]; adequate LN dissection could reduce the possibility of recurrence and metastasis. The efficiency for lymphadenectomy is still the main concern; the efficiency represents surgical removal of fifteen lymph nodes is the minimum standard. In most previous studies and our present studies, the mean number of harvested lymph nodes was more than fifteen for LG. However, whether laparoscopy could reach the same result as open surgery is still controversial; a previous study showed that experienced surgeons could realize the radicality in lymphadenectomy if the operative time is not limited [76]. In the present meta-analysis, we discovered that the harvested lymph nodes have no significant difference, which indicated LG could retrieve as many LNs as did OG through the improvement in laparoscopy facilities and sufficient training. D1+α or β dissection is now adequate for selected patients with early gastric cancer. With regard to advanced gastric cancer, whether D2 dissection is superior to D1 dissection remains controversial [77,78]. D2 dissection could realize more radical lymphadenectomy than D1 dissection, whereas D2 dissection could increase the postoperative mobility and mortality because of the invasiveness. In East Asia, maybe the incidence is high so that Asian surgeons are familiar with gastric cancer and have a better understanding of surgical technique; D2 lymphadenectomy is generally accepted as the standard to treat AGC. The Japanese Gastric Cancer Association has presented D2 lymphadenectomy as the standard treatment of local AGC [79]. However, some western studies have reported no significant long-term advantage with higher operative morbidity and mortality rates after D2 lymphadenectomy [80][81][82], and western surgeons would like to perform D1 lymphadenectomy; many recent report have indicated that patients who underwent laparoscopic gastrectomy combined with systemic lymphadenectomy could get a good survival result by extensively trained western surgeons [83,84]. Therefore, insufficient training of the laparoscopic gastrectomy combined with systemic lymphadenectomy may be the main reason for western surgeons. D2 dissection is an appropriate treatment for patients with advanced disease. Proximal esophageal and distal duodenal margins is also used to assess oncological adequacy, and proximal and distal margin distance could reflect the radicality of surgery, which is related with recurrence and OS and DFS in GC and other cancer [85]. Meantime, surgical margin is considered as an independent prognostic factor for GC. Our analyses also showed that there was no significant difference between the two groups; it indicated that LG is oncologically acceptable for proximal or distal located tumors.
Long-term outcome is the most useful endpoint to evaluate the oncological safety and effectiveness of surgery. Tumor recurrence and 5-year OS are usually used to evaluated the long-term outcome. Many studies have showed tumor recurrence was similar between the LADG and ODG [86,87]; previous clinical studies and meta-analysis have revealed that there was no significant difference in the 5-year OS between LADG and ODG [89]. In our meta-analysis, we make subgroup analysis based on published year; there demonstrated no statistically significant difference between the two groups; however, the results of 5-year OS favor LG with significant difference between 2016 and 2020, and the tumor recurrence showed no significant difference between the two groups. In other words, at least, LG is not inferior to ODG in terms of oncologic outcomes, which is acceptable for treatment of AGC.
Some limitations exist that should not be neglected for this meta-analysis. Many studies related with the theme are non-randomized retrospective trials; therefore, we have analyzed both the RCTs and NRCTs to avoid lack of samples. Several drawbacks of methodology may lead to heterogeneity, although the study has no observed obvious heterogeneity. In many included studies, the patients with multiple tumor stages were incorporated into one group and included in a single survival curve; it will have a significant effect on the quality and results of the meta-analysis. More ongoing RCTs should be performed to resolve the problem in the future.

Conclusion
In conclusion, we make a systematic review of thirty-six studies to release LG could be considered to be expanded in treating AGC. Gastric cancer is difficult to diagnose early with a poor prognosis, and patients often do not realize until cancer progresses to middle and advanced stages. Our study has presented the safety and curability of LG, which indicated an encouraging result for LG to be widely accepted in the future. More ongoing RCTs comparing the LAG with OG should be recommended.