Bursectomy and non-bursectomy D2 gastrectomy for advanced gastric cancer, initial experience from a single institution in China

The Ethics Committee of West China Hospital, Sichuan University, approved this retrospective study. The participants’ written consent were not obtained, but the patients’ records were anonymous to analysis.

From January 2012 to December 2013, those gastric cancer patients from the Department of Gastrointestinal Surgery, West China Hospital, Sichuan University, were included in this study based on the following criteria: (1) histologically confirmed gastric adenocarcinoma; (2) pT2-4, N0-3, and M0 stages according to the Japanese Classification of Gastric Carcinoma [15]; (3) distal gastrectomy and total gastrectomy by the conventional open method; (4) D2 lymphadenectomy according to the Japanese gastric cancer treatment guidelines [13]; and (5) curative resection without residual tumors (R0 resection). And those patients with distant metastases or positive cytology examination were excluded. Those patients underwent preoperative adjuvant chemotherapy were excluded to minimize the survival influence of it. Finally, the records of the patients who fulfilled the inclusion and exclusion criteria for this study were obtained and analyzed.

All the gastric cancer patients included in this study had undergone surgical treatments by well-trained surgeons in the gastric cancer treatment team of our department. Because this is not a randomized controlled study, patients who underwent bursectomy or non-bursectomy surgery were intraoperatively decided by the chief surgeons. The surgical treatment principle was adopted by the Japanese Gastric Cancer Treatment Guidelines which was published by the JGCA [16]. For the resection patterns, those advanced gastric cancer patients with tumors located in the upper or middle third had total gastrectomy. Patients with tumors located in the lower third of the stomach and with lymph nodes metastasis in no. 1, no. 2, and no. 4sb stations during the intraoperative frozen section evaluation also underwent total gastrectomy. Only lower third gastric cancers and distal gastrectomy can secure the tumor-free radical resection and have distal gastrectomy. According to the guidelines of the JGCA, in order to attain the complete dissection of subpyloric lymph nodes, the right side of the anterior plane of the transverse mesocolon and the pancreas was routinely resected, which was partially bursectomy. Besides the right side of the bursa omentalis, total bursectomy should be with en bloc resection of the peritoneal lining of the bursa omentalis (the anterior lobe of the transverse mesocolon and the pancreas), and the anterior lobe of the transverse mesocolon and the capsule of the pancreas should be dissected as much as possible [10], which was defined as the total bursectomy. Patients who underwent partial bursectomy (right side) were included in the non-bursectomy group, and patients who underwent total bursectomy were grouped in the bursectomy group. During the surgical procedures, the exposure of the surgical field, appropriate tension is essential to the success of completely removing the anterior lobe of the transverse mesocolon and the capsule of the pancreas (Fig. 1). Therefore, when the bursectomy is finished, only the posterior layer of the transverse mesocolon remained and the anterior lobe of the transverse mesocolon and the capsule of the pancreas were entirely separated from the transverse mesocolon and the pancreas (Fig. 2). Additional surgical procedures between the two groups were similar. D2 lymphadenectomy was carried out according to treatment guidelines published by Japanese Gastric Cancer Association [13]. Roux-en-Y esophagojejunostomy reconstructions were for total gastrectomy. And for distal gastrectomy, Billroth type I/II or Roux-en-Y gastrojejunostomy reconstructions were performed according to the tumor characteristics and others.

Clinicopathological characteristics, perioperative morbidity, and mortality were analyzed. The cross-sectional location and the longitudinal location of tumors were recorded according to the standard of the Japanese Classification of Gastric Carcinoma [15]. Lymph nodes were separately examined by the anatomical definitions of lymph node stations, according to the Japanese guidelines [13]. The number of positive and examined lymph nodes of the no. 4d, no. 4sb, and no. 6 LNs were compared between the two groups. The cancers were staged according to the Union for International Cancer Control (UICC) tumor-node-metastasis (TNM) system, 7th Edition [17]. Postoperative mortality and morbidity were counted for 30 days or during the same hospitalization. Specifically, the classification of pancreatic fistula was according to the international study group of pancreatic fistula classification [18]. The grades of the postoperative complications by the Clavien–Dindo classification of surgical complications between the two groups were evaluated and compared [19].

The follow-up was performed by means of routinely outpatient visit. Mail and telephone interviews were the supplementary methods. The follow-up information was updated until January 1, 2015. The follow-up rate, median follow-up duration (months), recurrence type, and overall survival outcomes were analyzed in the study. Recurrence types include the local-regional recurrence (anastomotic recurrence) type, peritoneal seeding recurrence type, hematogenous recurrence (liver, pulmonary, bone, et al.) type, distal lymph nodes (non-regional lymph nodes) type, and multisite recurrence type. The reasons for the patients lost to follow-up were predominantly refusal of the outpatient visit or a change in the telephone number and address.

Statistical analysis was conducted using SPSS statistics software, version 19.0 (SPSS, Chicago, Illinois, USA). For the continuous variables with normal distribution, they were analyzed by the one-way ANOVA test. Both the continuous variables without normal distribution and the ranked variables were weighed by the Mann–Whitney U test. The categorical variables were adopted with the Pearson’s chi-square test (or likelihood ratio). Survival outcomes were reported to Kaplan–Merrier method, log-rank test. Multivariate adjusted factor survival analysis was performed using Cox proportional hazard modeling. Hazard ratio (HR) and 95 % confidential interval (CI) were used to present the results of the univariate and multivariate survival analysis. Two-tailed p value of less than 0.05 was regarded as statistically significant.

From January 2012 to December 2013, 406 gastric cancer patients were included in the final analysis, 159 of whom in bursectomy group and 247 in non-bursectomy group. Most of the clinicopathological characteristics were comparable between the two groups, such as age, gender, tumor size, and tumor locations, and as well as tumor stages (Tables 1 and 2). Surgical duration was significantly increased in the bursectomy group when compared with the non-bursectomy group (260.1 ± 43.4 vs. 227.9 ± 48.6 min, p < 0.001, respectively). However, the amount of blood loss was comparable between bursectomy group than the non-bursectomy group (198.9 ± 43.4 vs. 201.1 ± 53.7 ml, p = 0.729). Meanwhile, we found out that more proportion of patients underwent total gastrectomy in bursectomy group than non-bursectomy group (54.1 vs. 25.5 %, p < 0.001). Also, there existed difference in the longitudinal location of tumors (upper, middle, lower) between the bursectomy and non-bursectomy groups (39.0, 13.8, 47.2 % vs. 11.2, 18.2, 70.4 %, respectively, p < 0.001).

Total number of examined LNs was significantly higher in bursectomy group than non-bursectomy group (40.6 ± 17.5 vs. 25.4 ± 9.9, p < 0.001). Besides, the number of positive lymph nodes was also higher in the bursectomy group than the non-bursectomy group (7.5 ± 8.7 vs. 5.9 ± 6.4, p = 0.045). Comparison of the number of the positive LNs and the number of examined LNs between the two groups in no. 4d LNs, no. 4sb LNs, and no. 6 LNs were listed in Table 2. The number of examined LNs was 4.9 ± 4.0 in no. 4d LNs of bursectomy group compared to 3.7 ± 2.7 of non-bursectomy group, p < 0.001. The number of examined LNs in no. 4SB LN and no. 6 LNs was similar between the two groups, p = 0.743 and p = 0.362, respectively.

There was no incidence of intraoperative mortality of patients in this study. The mean postoperative hospital stay was 11.4 ± 4.4 days in bursectomy group and 11.4 ± 4.4 days in non-bursectomy group (p = 0.850). Postoperative morbidity rate was comparable between the bursectomy and the non-bursectomy groups (26.4 vs. 17.8 %, p = 0.179). Details of postoperative complications were listed in Table 3. Reoperation was needed in five patients within 30 days of operation: three patients for abdominal cavity hemorrhage and two patients for intra-abdominal abscesses in bursectomy group and two patients for abdominal cavity hemorrhage and another patient of intestinal obstruction in non-bursectomy group. There was just one hospital death; a patient in non-bursectomy group died of postoperative acute myocardial infarction. All patients recovered well and successfully discharged from the hospital, including one patient in bursectomy group and one patient in non-bursectomy group who had a duodenal stump fistula with full drainage and effective nutrition support with more than 30-day postoperative hospital stay. We also compared the grades of postoperative complications according to the Clavien–Dindo classification and found that it was comparable between the two groups, p = 0.783 [19].

For the postoperative follow-up, there were 382 patients with fully postoperative follow-up information and a 94.1 % follow-up rate, 20 (2–35) months median follow-up duration. Survival benefits can be observed in bursectomy group compared to non-bursectomy group, although without significant difference, p = 0.055 (Fig. 3). In the univariate survival analysis, the longitudinal location (p = 0.030), macroscopic type (p = 0.027), tumor size (p = 0.002), resection patterns (p = 0.012), pT stage (p < 0.001), and pN stage (p < 0.001) were prognostic risk factors for the overall survival. And in the multivariate analysis, bursectomy (without vs. with, p < 0.001), pT stage (pT2-3 stages vs. pT4 stage, p < 0.001), and pN stage (N0 vs. N3, p = 0.002) were independent prognostic risk factors for the overall survival (Table 4). Subgroup analyses were conducted in the resection patterns and pT stages. For those patients who underwent distal gastrectomy, the survival outcome was comparable between the two groups (p = 0.129), whereas the bursectomy group had better prognosis of patients who underwent total gastrectomy than patients in non-bursectomy group (p = 0.016) (Fig. 4a, b). There is no significant difference between the two groups for patients with pT2-3 stages (p = 0.117) and pT4 stages (p = 0.128) (Fig. 5a, b). And for patients with pT4 stage, although there exist differences in the survival curves, no significant differences are shown for patients who underwent distal gastrectomy (p = 0.154) and total gastrectomy (p = 0.160) (Fig. 6a, b)