Does D3 surgery offer a better survival outcome compared to D1 surgery for gastric cancer? A result based on a hospital population of two decades as taking D2 surgery for reference

Background We conducted a retrospective study in our hospital in which we compared D1 with D3 through D2 lymphadenectomy for gastric cancer in terms of morbidity, postoperative mortality, long-term survival after surgery. Methods 567 patients who were performed curative intent between 1980 and 2003 were enrolled. 187 in the D1 group, 189 in the D2 group and 191 in the D3 group. Every procedure was verified by pathological analyses. The primary endpoints were 5-year overall survival. Results Median follow-up periods were 36 months and 60 months for D1 group and D3 group. Overall 5-year survival rate was significantly higher in patients underwent D3 surgery than in those performed D1 surgery (37.4% vs 48.7%; log-rank, p = 0.027). For the cases followed up to 120 months, the 10-year overall survival rate was 29% (95% CI, 22.1% to 35.9%) for the D1 group and 33.7% (95% CI, 26.6% to 40.8%) for the D3 group (log-rank, p = 0.005). Conclusions D1 surgery should be operated only for patients with Borrmann I disease. As D3 gastrectomy is associated with low mortality and adequate survival times when performed in selected institutions that have had sufficient experience with the operation and with postoperative management, we recommend D3 lymphadenectomy for patients with curable gastric cancer.


Background
Gastric cancer is still the most common cause of cancer related deaths worldwide, and a major clinical problem needing to be resolved because of the poor prognosis and the leak of treatment methods. Nowadays, surgical management is the major treatment method for gastric cancer. However, the efficacy of various extent of nodal dissection is still under debate. It was reported that improved prognosis was got in patients with gastric cancer who underwent D3 lymphadenectomy (first edition of Japanese classification of gastric cancer [1][2][3]. Furthermore, there were some randomised multi-institutional trials showing no survival benefits, but high morbidity and mortality, after D3 gastric dissection compared with D1 dissection [4,5]. To be mentioned, there were many participating surgeons with little experience in D3 surgery in these trials, hence, it's difficult to control the quality [4,5].
The more extended the surgery, the greater the risk of operation related morbidity and mortality is, as reported previously that nodal dissection increased morbidity [6]. It was reported that the postoperative mortality rate for gastrectomy surgery often exceeds 5% in West, even gets close to 16% in some articles, [7][8][9] only some Japanese studies reported a lower than 2% [10]. Besides the operation related morbidity, there was also a report showing that lymph node dissection did not adversely influence QOL, [11] and the operation related morbidity did not influence survival [12,13].
We conducted a single-institutional study and reported the long-term survival data for these two surgical groups of D1 and D3 taking D2 group as reference. Finally, we demonstrated that D3 surgery has overall survival benefit without significant operative complications and mortality.

Patients
We selected 567 patients who were histologically confirmed gastric cancer and underwent a radical operation at the First Affiliated Hospital of the China Medical University between 1980 and 2005. All of them, 187 were performed D1 dissection (D1 group), 189 received D2 surgery (D2 group) and 191 were treated with D3 dissection (D3 group). The inclusion criteria were as follows: 1), histologically proven, potentially curablengastric adenocarcinoma, and had physical fitness suitable for elective operation of either type of lymphadenectomy; 2), diagnosed based on the 5th UICC TNM classification system; 3), curative D1, D2 or D3 operations were performed; 4), a complete medical record was available; 5), patients of every period of diagnosis and every surgeon are roughly equal; and 6), never received neoadjunctive therapies and any kind of adjunctive therapy. Exclusion criteria were as follows: 1), older than 75 years; 2), previous or concomitant other cancer; 3), previous or concomitant gastrectomy for benign disease; 4), previous chemotherapy or radiotherapy; 5), clinical evidence of early gastric cancer on laparotomy; 6), oesophageal involvement; 7), macroscopically enlarged lymph nodes around the hepatoduodenal ligament or para-aortic regions; and 8), distant metastatic disease.
All patients were followed up by posting letters or telephone interviews. The last follow-up was December, 2008. Clinical findings, surgical findings, pathological findings and every follow-up were collected and recorded in the database. All the subjects gave written informed consent to study protocol, which was approved by the Ethics Committee of China Medical University.

Surgical procedures and classifications of gastric cancer
Surgical procedures and pathological assessment refered to the Japanese classification of gastric cancer [1]. All patients in the study underwent standard total or distal subtotal gastrectomy, depending on the location and macroscopic appearance of the primary tumor. The definition of lymphadenectomy was based on the Japanese Classification of Gastric Carcinoma [5]. D1-dissection of all the group 1 nodes; D2-dissection of all the group 1 and group 2 nodes; D3-dissection of all the group 1, group 2 and group 3 nodes. Group 1 consists of the perigastric lymph nodes, and group 2 consists of the lymph nodes along the left gastric artery, the common hepatic artery, and the splenic artery and around the celiac axis. However, when the tumor is located in the lower third stomach, the lymph nodes along the splenic artery are classified as group 3. Group 3 also consists of lymph nodes in the hepatoduodenal ligament at the posterior aspect of the head of the pancreas and at the root of the mesentery.
Surgeons routinely removed lymph nodes from the excised specimens as more as possible after operation, based on the Japanese Classification of Gastric Carcinoma and their experience. The specimens and retrieved lymph nodes were stained with hematoxylin and eosin and pathologically examined in the Gastric Laboratory of the First Affiliated Hospital of the China Medical University.

Endpoints and follow-up
The primary endpoints were 5-year overall survival. Overall survival was calculated from the day of surgery until death or the last follow-up contact. Data for a patient were censored at last follow-up when they were alive. Follow-up assessments were done every 6 months for the first 5 years after surgery, and then every 12 months until death.

Statistical analyses
Data from all eligible patients were analyzed for overall survival. Survival curves were estimated by the Kaplan-Meier method and treatment comparisons were made by the log-rank test.
Potential prognostic factors were entered into a Cox's regression model including age, sex, tumor size, site of tumour in stomach, gross appearance, tumour stage, clinical node status, nodal stage, joint organ removal, gastrectomy, blood transfusion and blood loss. In multivariate analysis, the prognostic factor detected in univariate analysis and treatment group were as covariates included in the Cox regression model. Two-sided P values were calculated for all tests and are reported here. P values less than 0.05 were considered to indicate statistical significance. Analyses were performed with the use of SPSS software, version 16.0.
Results D1 group was with median age of 55 years old, D2 group was with median age of 55 years old and D3 group was with median age of 54 (table 1). All patients were followed up for at least 5 years (until December 19, 2008).
The hazard ratio for death was 0.708 (95% CI, 0.560-0.894; p = 0.004) in the D3 group (table 2, univariable analyses). After adjustment of thirteen baseline variables (age, sex, tumor size, tumor location, Borrmann type, T stage, clinical node status, lymph-node stage, histological type, joint organ removal, gastrectomy, blood transfusion and blood loss) with the use of Cox regression analysis, the hazard ratio was hardly unchanged (hazard ratio, 0.771 (95% CI, 0.599-0.992); P = 0.043) (table 2, multivariable analyses). Expectedly, the multivariate analyses showed that >7 cm in tumor size, the upper third tumor and the whole stomach tumor, Borrmann III type, N3 disease, D1 and D2 dissection were significantly associated with poor survival (table 2).
As shown in table 3 and 4, D1 group got significantly more benefit than D2 or D3 group only for Borrmann I and N3 disease, the hazard ratios for death in the D1 group were 0.618 (95% CI, 0.399-0.958; P = 0.031) and 0.369 (95% CI, 0.162-0.841; P = 0.018), respectively. D3 group has significantly more benefit than D1 and D2 surgery in the subgroups of cases with ≤5 cm and >7 cm tumors, the lower third tumor, Borrmann II and III types, T3 stage, positive clinical node, N0 and N1 disease, no joint organ removal, the subtotal gastrectomy, blood transfusion and 200-400 ml blood loss. There was no evidence indicating that D2 surgery has any significant benefit for these subgroups.

Discussion
In the study, we found significant improvement in overall survival with D3 surgery compared to D1 surgery. Figure 1 The overall rates for all enrolled patients. There were significant differences between D1 and D3 group (p = 0.004), and between D2 and D3 group (p = 0.002). However, there was no significant difference between D1 and D2 group.      Furthermore, no significant difference was found in the incident rates of major surgery-related complications between the two groups, which was similar to the results in the trials done in Hong Kong, [14] the UK, [5] and Dutch [15]. We conducted a post hoc subgroup analysis including thirteen variables.To be interest, the results indicated that D1 surgery got significantly more benefit than D3 surgery for N3 disease, while D3 surgery has significantly more benefit than D1 surgery in the subgroups of N0 and N1 disease. Since this result was from a post hoc subgroup, it might be a false positive owing to multiple testing, [16] the possible survival benefit of D3 lymphadenectomy in node-negative patients will need to be clarified in further studies.
In a review article, Sun Hu Yang et alreported that there was no difference in the 3-or 5-year survival between D1 and D2 [17]. In the study, the 5-year overall survival was 37.4% for D1 group and 48.7% for D3 group, and the period of diagnosis didn't affect survival (p = 0.084, log-rank test) [18]. It can be seen that longterm survival is lower, when comparing our results with historical report, in which the observed 5-year survival rates were 53.6% and 59.5%, respectively [13]. This result indicated that the time of diagnosis of malignant tumours as well as gastric cancer is much later in China than in other countries, especially western countries.
In this study, we found that more than 60% of both the D1 and D3 patients had lymph node metastases which was higher than the report of Bunt et al [19]. In the D3 group, which includes lymph node dissection of the N1, N2 and N3 level, there were 42% being classified as N1, 13% as N2, and 9% as N3. The extended surgery is considered to be related to the risk of operative morbidity and mortality [6]. The mortality for gastrectomy in Western countries was usually5% and even approaches 16% in some trials [7][8][9].
Robert C G et al reported that the overall 5-year survival rate for the 286 patients undergoing gastrectomy with additional organ resection was 32%, which was The P values are for hazard ratios for death in the group assigned to D1 lymphadenectomy and the group assigned to D2 lymphadenectomy, with 95% confidence intervals. The surgery is better for D1 lymphadenectomy when the HR is <1; and is better for D2 lymphadenectomy when the HR is >1. †Significantly better for D1 lymphadenectomy. *Considered significant.
significantly less than the gastrectomy-alone group [20]. Besides the surgery extent, the participating surgeons' operative skill and experience, and the workload cases are also important factors for survival rates [21,22]. There are many studies having reported a relationship between the number of cases treated in a hospital and the outcomes of cancer treatment [22][23][24][25][26][27]. Moreover, the uniformity of treatment is also important. Our study was carried out in a hospital that performs a high volume of nodal dissections for gastric cancer with low morbidity and mortality rates. In our study, all participating surgeons were of the same department, which minimizes the variation in individual operating skill and management, and did an equal number of D1 and D3 resections during the trial, which balances the comparisons between the two groups without bias to individual surgeons' skill. Therefore, the experience as a result of caseload, surgical skill, and the case selection are very important [5,28,29].

Conclusions
As D3 gastrectomy is associated with low mortality and adequate survival time when performed in selected institutions that have had sufficient experience with the operation and with post-operative management, we recommend D3 lymphadenectomy for patients with curable gastric cancer except for patients with Borrmann I disease who are more suitable for D1 surgery.