Endoscopic resection versus surgery for early gastric cancer and precancerous lesions: a meta-analysis

Aim To compare the efficacy and safety of endoscopic resection (ER) and surgery for the treatment of early gastric cancer and precancerous lesions. Methods Databases, such as PubMed, EMBASE, Cochrane Library, and Science Citation Index, from 2000 to 2016, were searched for eligible articles. In this meta-analysis, the main outcome measurements were local recurrence, complications, metachronous lesions, hospital stay, and 5-year overall survival. Results Nine trials were identified and a total of 2748 patients were included. The rate of complication was higher in the surgery group compared with the ER group (OR 0.41; 95 % CI 0.30–0.55). The rates of local recurrence and metachronous lesions were lower in the surgery group (OR 0.03; 95 % CI 0.00–0.06; OR 8.76; 95 % CI 4.17–18.41). The hospital stay was shorter in the ER group (mean difference −6.96; 95 % CI −7.94 to −5.99). The 5-year overall survival rate did not significantly differ between the two groups (OR 1.23; 95 % CI 1.03–1.47). Conclusions We provided evidence that, ER was comparable to surgery in terms of the 5-year overall survival. In addition, ER had a lower rate of complications and shorter hospital stay, but a higher rate of local recurrence and metachronous lesions for the treatment of early gastric cancer and precancerous lesions.


Background
Gastric cancer is the second most common cause of cancer-related death in the world, and it remains difficult to cure, primarily because most patients present with advanced disease. Early gastric cancer is defined as gastric carcinoma confined to the mucosa and submucosa of the stomach, with or without regional lymph node metastasis (Japanese Gastric Cancer Association 1998. Gastrectomy with lymph node dissection remains the standard therapy for early gastric cancer. A variety of complications can develop after surgery, and the patient's quality of life can be greatly impacted because of the modification of the stomach (Okamura et al. 1988).
Endoscopic resection (ER) methods, such as endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD), are widely accepted for the treatment of superficial gastric cancer without lymph node metastasis, because of the minimal invasion, low cost, good patient tolerance, and better quality of life after the operation, with low adverse event rates (El-Sedfy et al. 2014;Shen et al. 2013).
There is no current consensus on the optimal method for the treatment of early gastric cancer and precancerous lesions. We conducted a systematic literature review to compare the efficacy and safety of ER and surgical methods for the treatment of early gastric cancer and precancerous lesions.

Data sources and searches
We searched databases including PubMed, EMBASE, the Cochrane Library, and Science Citation Index from January 2000 to March 2016 to identify related articles, without language restriction, which compared ER and surgery. All bibliographies were indentified in the reference lists. The search terms were "gastric cancer or gastric neoplasia" and "endoscopic mucosal resection or endoscopic submucosal dissection". Major proceedings of international meetings (such as Digestive Disease Week and Asian Pacific Digestive Week) were also hand-searched.

Study selection
The inclusion and exclusion criteria are shown in Table 1.

Data extraction and quality assessment
Data were extracted by one investigator and confirmed by the other according to a predefined data extraction form. Disagreements were resolved by consultation with a third investigator. The following data were collected: year of publication, first author, country, duration, age, sex, depth of invasion, differentiation and follow up period, the local recurrence rate, procedure-related complications, metachronous lesions, hospital stay, and 5-year overall survival. The Newcastle-Ottawa Scale was used to assess the quality of the included non-randomized studies (Wells et al. 2015).

Statistical analysis
All extracted data were entered in the freeware program Review Manager (Version 5.0 for Windows, Cochrane Collaboration). The weighted mean difference was recommended for continuous data, and the odds ratio (OR) with 95 % confidence intervals (CI) was recommended for dichotomous data. Statistical heterogeneity between trials was evaluated by the Chi square test and was considered to be present when P was less than 0.1. We also used I 2 to assess the heterogeneity. An I 2 of more than 50 % was considered to be statistically significant. In the presence of statistical heterogeneity, heterogeneity was explored by subgroup analysis or a random-effects model. Publication bias was detected by a funnel plot, and then the symmetry of the funnel plot was confirmed by the Egger's test, with a P value of 0.05.

Study selection
A total of 2251 potential studies were retrieved for the meta-analysis. 1675 were excluded for not including the surgical treatment and 565 were excluded because ER and surgery were not compared. Of the 11 remaining articles, 2 were excluded because they had not compared the main outcomes. The remaining 9 eligible studies Song et al. 2015;Choi et al. 2011Choi et al. , 2015Kim et al. 2014Kim et al. , 2015Park et al. 2014;Chung et al. 2014;Chiu et al. 2012) were chosen for further analysis (Fig. 1). A total of 2748 patients were included in the meta-analysis, including 1339 patients in the ER group and 1409 patients in the surgery group. Of these studies, 6 included only patients with ESD Song et al. 2015;Park et al. 2014;Kim et al. 2014;Chung et al. 2014;Chiu et al. 2012), 1 included only patients with EMR (Choi et al. 2011), and 2 included patients with both ESD and EMR (Choi et al. 2015;Kim et al. 2015). Open surgery or laparoscopic surgery were mentioned in 5 studies (Choi et al. 2015;Kim et al. 2014Kim et al. , 2015Chung et al. 2014; Chiu  Song et al. 2015;Park et al. 2014;Choi et al. 2011). All of the studies were retrospective case-control studies, not randomized controlled trials (RCTs). The key characteristics of the studies are listed in Table 2.

Hospital stay
The hospital stay was reported in 3 studies Song et al. 2015;Kim et al. 2014). There was no heterogeneity in the studies (P = 0.95, I 2 = 0 %), and a fixed effect model was applied. The mean hospital stay was significantly shorter in the ER group, compared with the surgery group (Mean difference −6.96; 95 % CI −7.94 to −5.99) (Fig. 5).

5-Year overall survival
The 5-year overall survival was reported in 5 studies Song et al. 2015;Choi et al. 2011Choi et al. , 2015Kim et al. 2015;Park et al. 2014). There was no heterogeneity in these studies (P = 0.33, I 2 = 13 %), and a fixed effect model was applied. Analysis showed that there was no significant difference in the 5-year overall survival rate between the ER group and the surgery group (OR 1.23; 95 % CI 1.03-1.47) (Fig. 6).

Publication bias
We used the local recurrence rate as the outcome, and no publication bias was detected by funnel plot and the Egger's test (P = 0.862).

Discussion
The comparison between ER and surgery in the treatment of early gastric cancer is still controversial. Even though ER is less invasive and less expensive, and could better preserve the physiological function of the stomach, additional surgery with lymph node dissection is recommended when ER is histologically non-curative or submucosal invasion is detected regardless of margin status because of possible residual tumors or lymph node metastasis (Oda et al. 2008;Song et al. 2008). Therefore, we designed the meta-analysis to systematically evaluate the two techniques, providing evidence for the optimal treatment of early gastric cancer. In the present analysis, 9 retrospective studies were included, and the results confirmed that compared with surgery, ER had a higher rate of local recurrence and metachronous lesions for the treatment of early gastric cancer and precancerous lesions. Because the ER method preserves the whole stomach, the incidence of local recurrence and metachronous lesions after ER was higher than after surgery (Nakajima et al. 2006;Takeda et al. 1998;Hosokawa et al. 2002). In most cases, local recurrent and metachronous lesions were successfully cured with additional endoscopic treatments. That is why there was no significant difference in the 5-year overall survival rate between the two groups.
The complications of ER included bleeding and perforation, which can be managed using endoscopic treatment. The complications that occurred after gastrectomy included bleeding, duodenal leakage, ileus and hepatic dysfunction, and pancreatic leakage. 5 studies provided the data of the complication rate for EGC. The pooled analysis showed that it was higher in the surgery group than the ER group.
Five studies compared the hospital stay between the two groups, three studies provided the data of mean ± SD, and the mean hospital stay was significantly shorter in the ER group, compared with the surgery group.
Two studies (Choi et al. 2011;Kim et al. 2014) compared the medical cost between the two groups. The Table 2 The key characteristics of the included studies results showed that ER patients had lower medical costs than patients who had conventional surgeries for early gastric cancer. Choi et al. (2015) reported that endoscopic treatment for EGC provides a better quality of life, but stomach preservation might provoke cancer recurrence worries.

Year of publication
There is an alternative therapeutic approach for the treatment of early gastric cancer which combines the Fig. 2 The rate of local recurrence comparing ER and surgery Fig. 3 The rate of complications comparing ER and surgery Fig. 4 The rate of metachronous lesions comparing ER and surgery advantages of both surgical and endoscopic treatment also known as laparoscopic endoscopic cooperative surgery (LECS) which was introduced in 2009. It was reported that LECS held great promise for the future of minimally invasive oncologic procedures for the treatment of early gastric cancer (Ntourakis and Mavrogenis 2015). It can be used as an alternative to endoscopic therapy and surgical treatment.
There were certain limitations in our analysis. First, the major limitation of this meta-analysis was that none of the included studies were randomized. This certainly attenuated the evidence level and value of this meta-analysis. Second, included studies were from only 2 countries, Korea and China, so the results need further confirmation in other countries.
In conclusion, based on the findings of our meta-analysis, ER showed advantages over surgery for early gastric cancer and precancerous lesions regarding the procedure-related complication rate and hospital stay. The disadvantages of ER were the higher rate of local recurrence and metachronous lesions. Fortunately, we can treat these with additional endoscopic treatments without affecting overall survival. In the view of the present meta-analysis and all available trials, we suggest that ER is appropriate to most early gastric cancer and precancerous lesions.