Impact on survival of the number of lymph nodes resected in patients with lymph node-negative gastric cancer

Our study was based upon rigorous analyses of the previous studies about lymph node-negative gastric cancer patients, thus excluding patients who died within 3 months after operation, patients who had an esophagogastric or cardia cancer (site code C160), patients who had distant metastasis, or patients who had no lymph node resected. The reasons for this criteria setup are as follows: (1) patients died within 3 months likely because of surgical complications; (2) patients who had adenocarcinomas of esophagogastric junction were staged as the esophageal scheme according to the AJCC seventh edition; (3) if the gastric cancer had distant metastasis, it would be a stage IV cancer regardless of the depth and lymph node invasion; and (4) if a patient did not receive the lymphadenectomy, it would be difficult to determine whether the patient had nodal metastases.

Prognostic indicators for high recurrence rate at univariate analysis were ages that were greater than 65 years, the white race, large tumors, deep depth of invasion, and little lymph nodes removed. The association between age and prognosis was hypothesized that elder patients suffer from host immunodeficiency or malnutrition [11, 12]. In our study, tumor location is not a prognostic indicator for lymph node-negative gastric cancer. But this factor is still in controversial. Xu et al. [9] and Baiocchi et al. [11] reported that distal tumors were associated with better prognosis due to a relatively late presentation of the symptoms in proximal gastric cancer [13]. Nevertheless, Lee et al. [14] did not find this phenomenon. Although we divided the tumor location into four groups: upper third, middle third, lower third, and overlapping, we did not find its value to investigate the prognosis. We further combined them into two subtypes: distal gastric cancer and others, and we did not find prognostic value either.

The SEER database is consisted of several registries with different races. Our research indicated that in the univariate analysis, race is a prognostic factor of lymph node-negative gastric patients. The overall survival rate of East Asian patients is better than others. This may be due to the fact that East Asian surgeons routinely perform the D2 lymph node dissection while many Western surgeons perform D1 lymph node dissection [15]. This is supported by the number of lymph nodes resected, combined with the fact that although all the patients underwent D1 or D2 lymph node dissection, the pathologists and surgeons searched for lymph nodes in different efforts and techniques, which could lead to omitted lymph nodes in the specimen [16]. We can see from the results that 40.5 % of East Asian patients received the lymphadenectomy with more than 15 lymph nodes removed, while only 24.2 % of the patients in the Western world received the same type lymphadenectomy.

Tumor size failed to demonstrate a significant association with overall rate in our series in the multivariate analysis. However, the value of these prognostic factors is still in controversial. Baiocchi et al. [11] and Lee et al. [14] found that tumor size (cutoff 40 and 63 mm, respectively) could not independently affect the overall survival. By contrast, Xu et al. [7] indicated that tumor size (cutoff 50 mm) was a strong prognostic factor affecting the survival of lymph node-negative gastric cancer patients. But in this study, tumor size range is analyzed to homogenize the difference by choosing multiple cutoff points which is distinct from other studies using only one cutoff point.

LNN groups, depth of tumor invasion, and ages are independent prognostic indicators for lymph node-negative gastric cancer. The four cutoff points for LNN subgroups indicate that the more lymph nodes resected, the better prognosis the patients can have. After extracting the patients who received the lymphadenectomy with more than 15 lymph nodes removed, however, we found that a more extensive lymphadenectomy (more than 25 lymph nodes resected) did not lead to better survival. For different tumor size subgroups, the overall survival rate among them has statistically significant difference when different lymph nodes were resected, even in the less-than-1-cm group. This means that we cannot judge the extent of lymph node dissection by the tumor size. But for the different depth of invasive subgroups, our study indicated that the overall survival rate for T1 + T2 stage lymph node-negative gastric cancers has no statistically significant difference, while it is statistically significant for the T3 + T4 subgroup. An alternative explanation for this observation may be that patients with T1 or T2 subcategories seldom spread to regional lymph nodes. The incidence of lymph node metastasis in T1 and T2 patients are less than 35 % [17‐19]. However, it is difficult to determine the depth of invasion and lymph node metastasis before surgery. Recent studies demonstrated that endoscopic ultrasonography has improved the local accuracy in estimating the depth of tumor invasion and lymph node involvement [20]. The accuracy of endoscopic ultrasonography is better than that of CT scan in determining the extent of infiltration of the tumor. The accuracy ranges from 67 to 92 % [21]. Although the development of technology improves the accuracy in determining the extent of infiltration of gastric cancer, difficulty still remains in differentiating the T2 stage from T3 stage [22]. A high-frequency (up to 30 MHz) miniprobe-endoscopic ultrasonography, which is able to demonstrate gastric wall up to nine different layers, can reach 100 % accuracy in identifying the T1 gastric cancer. But the accuracy of endoscopic ultrasonography is highly dependent on the experience of the operators [22]. Therefore, we strongly suggest that at least 15 lymph nodes be resected during the radical operation.