Introduction

Gastric cancer lesions confined to the mucosa or submucosa, irrespective of the presence or absence of lymph node metastasis, are referred to as early gastric cancer (EGC) [1]. Gastrectomy with lymph node dissection results in a 5-year survival rate of approximately 99% in cases of intramucosal carcinoma, except in cases of death from intercurrent disease [2]; thus, endoscopic resection must also achieve comparable therapeutic outcomes. Therefore, in order to achieve a radical cure comparable to surgical resection, it is important to select lesions with ≤ 1% probability of lymph node metastasis in performing endoscopic resection.

In Japan, the indication for endoscopic resection is differentiated cancer ≤ 2 cm in size without an ulcer, of which the depth of invasion is clinically diagnosed as tumor confined to the mucosa [3]. Meanwhile, endoscopic resection is not indicated for undifferentiated gastric intramucosal carcinoma, which is associated with a significant rate of lymph node metastasis ranging from 2.9 to 7.3% [4, 5].

However, in cases of undifferentiated gastric intramucosal carcinoma ≤ 2 cm in size without an ulcer and lymphovascular invasion (LVI), the probability of lymph node metastasis is as low as 0% [6, 7]. Therefore, endoscopic resection could become a standard treatment for this type of cancer.

This study aimed to retrospectively analyze the factors associated with lymph node metastasis and to determine the validity of endoscopic resection in patients with undifferentiated EGC. The factors in question were identified from the clinicopathological features of patients who underwent surgical resection for undifferentiated EGC.

Patients and Method

This study included 141 patients (85 men and 56 women with a median age of 63 years, ranging from 27 to 83 years) who underwent gastrectomy with lymph node dissection for solitary undifferentiated EGC (poorly differentiated adenocarcinoma, signet-ring cell carcinoma, and mucinous adenocarcinoma) in the Department of Surgery, The Jikei University School of Medicine Kashiwa Hospital, between January 2002 and December 2014. Patients with multiple cancers involving other organs or gastric carcinoma in the remnant stomach were excluded.

Age, sex, maximum tumor size, location of primary lesion, macroscopic type, histological type, depth of tumor invasion, the presence or absence of LVI, the presence or absence of intratumoral ulcers, and the number of dissected lymph nodes were retrospectively compared between patients with and without lymph node metastasis to identify the factors associated with lymph node metastasis.

The clinicopathological findings were documented according to the Japanese Classification of Gastric Carcinoma published by the Japanese Gastric Cancer Association [1]. Histological examinations including the diagnosis of LVI were performed with routine hematoxylin and eosin staining.

Statistical Analysis

Continuous variables were expressed as the median value (range) and were compared by the Mann-Whitney U test. Categorical variables were compared using the Fisher exact test or χ2 test. Factors found to be significant by univariate analysis were included in subsequent multivariate logistic regression analysis to determine variables independently associated with risk factors for lymph node metastasis. The statistical analysis was performed with the Microsoft Excel software program. p < 0.05 was regarded as statistically significant.

Results

According to the depth of tumor invasion, lymph node metastasis was observed in 13 patients (9.2%), including three with intramucosal carcinoma (3.6%) and ten with submucosal invasive carcinoma (17.2%). Univariate analysis identified tumor size, depth of tumor invasion, and LVI as risk factors for lymph node metastasis (Table 1).

Table 1 Univariate analysis of risk factors for lymph node metastasis in patients with undifferentiated EGC
Full size table

In terms of the association of wall invasion depth and tumor size with lymph node metastasis, the tumor size exceeded 30 mm in all three intramucosal carcinoma patients with lymph node metastasis (Table 2).

Table 2 Relationship between tumor size, depth of tumor invasion, and lymph node metastasis in undifferentiated EGC
Full size table

Among the three factors showing a significant difference by univariate analysis, LVI (positive; p = 0.002) was identified by multivariate analysis as an independent risk factor for lymph node metastasis (Table 3).

Table 3 Multivariate analysis of risk factors for lymph node metastasis in patients with undifferentiated EGC
Full size table

Discussion

The rate of lymph node metastasis in undifferentiated EGC is reported to range from 8.5 to 18.3% [5, 8], and the rate in this study was 9.2%. In this study, the rate of lymph node metastasis in intramucosal carcinoma was 3.6%, which was almost comparable to previously reported rates [4, 7, 8, 10]. While many studies [4, 5, 7,8,9,10,11,12], like ours, found that univariate analysis identified tumor size, depth of tumor invasion, and LVI as risk factors for lymph node metastasis in undifferentiated EGC, many mothers found that multivariate analysis identified these variables as independent risk factors for lymph node metastasis. However, some reports indicated that tumor size, depth of tumor invasion, and ulcerative findings were not independent risk factors for lymph node metastasis, as found in our study (Table 4) [4, 5, 7,8,9,10,11,12].

Table 4 Reports of lymph node metastasis rate in undifferentiated early gastric cancer
Full size table

Gotoda et al reported that no lymph node metastasis was observed in 141 patients with undifferentiated gastric intramucosal carcinoma ≤ 20 mm in size without an ulcer [6]. Hirasawa et al also reported that no lymph node metastasis was observed in 310 patients with the same clinical conditions (95% confidence interval, 0–0.96%), suggesting that endoscopic resection might also achieve therapeutic outcomes comparable to surgical resection in these patients [7]. In addition, no lymph node metastasis was observed in any of the 13 patients in our study. However, there are also reports of lymph node metastasis in undifferentiated intramucosal carcinoma ≤ 20 mm in size without an ulcer or LVI. For example, lymph node metastasis was reported in one patient each with a 13-mm tumor by Nasu et al [13], with a 15-mm tumor by Odagaki et al [14], and with a 17-mm tumor by Park et al [15]. Early signet-ring cell carcinoma is reportedly associated with a lower rate of lymph node metastasis than other undifferentiated types of cancer [16,17,18,19]. However, our study showed no statistically significant difference according to the histological type; all three patients with intramucosal carcinoma associated with lymph node metastasis were histologically diagnosed as having signet-ring carcinoma. In conclusion, only LVI was an independent risk factor of lymph node metastasis for undifferentiated EGC in this study. Thus, in this study the undifferentiated EGC which undifferetiated early gastric cancer is a mucosal depth. Currently, the LVI can be revealed only by the histopathological examination of the excision lesion. So it is difficult to estimate the LVI before treatment. As the evidence is still insufficient for undifferentiated EGC, the following resections are regarded as non-curative for the time being [20]. Therefore, it was thought that it should perform endoscopic resection for the undifferentiated EGC as diagnostic treatment. The use of endoscopic resection for the undifferentiated EGC should be considered carefully for patients with LVI because of the risk for lymph node metastasis.