Recently, the survival rate of gastric cancer has improved due to advancements in diagnosis and surgical treatment, particularly due to the increase in the diagnosis of asymptomatic early gastric cancer (EGC) through improvements in diagnostic technology.[1,2] Consequently, the number of gastric cancer surgery survivors has increased.[1,3] Therefore, many physicians are interested in improving quality of life through limited lymphadenectomy, endoscopic submucosal dissection, and sentinel lymph node navigation surgery. However, the basic principles of gastric cancer treatment are extensive dissection of the lymph nodes (LNs) and complete removal of the primary tumor in patients with resectable cancer.[4–7] The prognostic risk factors for EGC include LN metastasis,[8,9] tumor size,[9] histologic type,[10] lymphovascular invasion,[10] and age.[9] Lymph node metastasis is the most important prognostic risk factor for recurrence in early gastric cancer,[8,9] but the prognostic risk factors for EGC other than LN metastasis remain debatable. Lymph node metastasis is an important prognostic factor in EGC. Therefore, gastrectomy with LN dissection is the primary treatment.[11] Based on the AJCC 5th staging system, more than 15 retrieved LNs are required for accurate staging.[12] Many studies have reported that the number of retrieved LNs is correlated with better prognosis.[13–15] The number of resected LNs for each specimen was determined by the extent of lymphadenectomy and thoroughness of LN retrieval.[16] Lymph node dissection is important because LN metastasis is a significant prognostic factor for early gastric cancer. The optimal number of retrieved LNs in gastric cancer has been shown to be greater than 15.[12,17,18] The aim of our study was to evaluate the relationship between the number of retrieved LNs and prognosis in early gastric cancer with submucosal invasion and to analyze the factors affecting the number of retrieved LNs.

Among the 1953 patients who were pathologically diagnosed with gastric cancer and underwent gastrectomy between January 1995 and December 2014, 443 who were diagnosed with submucosal invasion after gastrectomy were enrolled in this study. We excluded patients who underwent non-curative surgery, had a history of other cancers, or had remnant stomach cancer. Follow-up was conducted until October 2019, and the mean follow-up was 58.0 ± 40.9 months, with a median of 55.7 months. Recurrence patterns were classified into 4 categories: locoregional recurrence, peritoneal dissemination, hematogenous metastasis, and distant LNs. Gastric cancer surgery was performed according to the Japanese Gastric Cancer Treatment Guidelines.[19] Pathological staging was conducted according to the AJCC 7th edition.[20] The patients were followed up every 6 months after surgery for 5 years. The follow-up evaluation consisted of history taking, physical examination, laboratory findings, endoscopy, chest radiography, and computed tomography. Magnetic resonance imaging, positron emission tomography, and bone scans were performed when required. Recurrence was determined using a retrospective review of medical records.

We retrospectively analyzed the risk factors for recurrence in patients with submucosal invasion of the gastric cancer. All statistical analyses were performed using IBM SPSS software package (version 21.0; IBM Co., Armonk, NY, USA). The Kaplan–Meier method was used to analyze the risk factors for recurrence, and the log-rank test was used to analyze statistical significance. Chi-square tests were used to analyze clinicopathological correlations, and Cox proportional hazards models and logistic regression analyses were used for multivariate analyses. The relationship between retrieved LNs and metastatic LNs was analyzed using Spearman’s rank correlation coefficient. A P-value < .05 was considered at significant. This study was approved by our institutional review board.

Among the 443 patients, 300 (67.7%) were male and 143 (32.3%) were female, with a mean age of 62.2 ± 10.6 years. Recurrence was observed in 22 of the 447 patients with submucosal gastric cancer. Complications occurred in 71 patients (16.0%), and the most common was wound complications (42 patients). Recurrence was observed at 47 sites in 22 patients, and the most common recurrence pattern was hematogenous (Table 1). The mean number of retrieved LNs was 31.0 ± 13.8. In the univariate analysis, retrieved LNs ≤ 25 (P-value = .004), complications (P-value = .002) and node metastasis (P-value = .017) were prognostic risk factors. The 5-years disease-specific survival (DSS) of retrieved LNs ≤ 25 and greater than 25 were 92.1% and 98.2%, respectively. According to node metastasis, the 5-years DSS rates for node negativity and positivity were 97.4% and 91.3%, respectively (Table 2, Fig. 1). In multivariate analysis, retrieved LNs ≤ 25 (hazard ratio [HR] = 5.754, P-value = .001) and node metastasis (HR = 3.031, P-value = .029) were independent prognostic risk factors (Table 3).

The recurrent graph according to the number of retrieved lymph nodes (A), and nodal status (B).

According to the number of retrieved LNs, 5-years DSS of retrieved LNs < 15, 16 to 20, 21 to 25, 26 to 30, and > 31 LNs were 93.2%, 89.7%, 93.4%, 100%, and 97.5%, respectively (P-value = .046) (Table 4).

Age, sex, tumor location, tumor size, complications, differentiation, lymphovascular invasion, and perineural invasion were not associated with the number of retrieved lymph nodes. However, BMI (body mass index) and node metastasis were related with number of retrieved LNs fewer than 25 in univariate analysis. In multivariate analysis (Table 5), a BMI greater than 25 kg/m2 (HR = 0.510, P-value = .002) and node metastasis (HR = 2.084, P-value = .006) were independent prognostic factors of retrieved LNs ≤ 25 (Table 6). There was a statistically significant correlation between the number of retrieved and metastatic LNs (r = .167, P < .001) (Fig. 2).