Gastric cancer is the most common cancer and the fourth leading cause of cancer-related deaths in South Korea [1]. Early gastric cancer (EGC) is defined as adenocarcinoma confined to the mucosa or submucosa of the stomach with or without regional lymph node metastasis (LNM), and the prognosis is excellent even in the presence of LNM [2].

The standard treatment option for gastric cancer is gastrectomy with regional lymph node dissection. However, in Korea and Japan, endoscopic submucosal dissection (ESD) has been accepted as an optional treatment for EGC without LNM [3]. The absolute indication for ESD as recommended by the Japanese gastric cancer treatment guidelines 2014 (4th edition) includes a differentiated type of adenocarcinoma without ulceration, the depth of invasion of which is clinical T1a and the diameter is ≤2 cm [4].

For patients who underwent non-curative ESD, additional surgery is recommended to prevent residual cancer (RC) or LNM. Non-curative factors include submucosal invasion (SM1) >500 μm, lymphovascular invasion, undifferentiated histology, large tumor size, and a positive resection margin [5].

In this study, we aimed to evaluate the clinicopathological characteristics of patients who underwent an additional gastrectomy after a non-curative ESD procedure to identify the risk factors of RC and LNM.

ESD is a minimally invasive procedure performed for the treatment of EGC. Compared to surgery, ESD can avoid side-effects such as appetite loss, dysphagia, eating restrictions, and anxiety [7]. However, Tanabe et al. [8] reported a local recurrence rate of 0.22% in the absolute indication group for ESD, and local recurrence and LNM rates were 1.26% and 0.14% in the expanded indication group. They mentioned that local recurrence can occur when curative resection is not achieved at the horizontal or vertical margin, and LNM is associated with the presence of risk factors such as submucosal microinvasion and lymphovascular invasion.

Toyokawa et al. [9] reported that among patients who underwent non-curative ESD without additional gastrectomy, the incidence rates of RC and LNM were 9% and 9%, respectively. Lymphovascular invasion and the presence of undifferentiated type tumors were identified as risk factors for LNM, and a positive horizontal margin was identified as a risk factor for RC. Choi et al. [10] reported recurrence in 15% (3/20) of patients with SM2 gastric cancer who were followed up after ESD without surgery. In both studies, the authors strongly recommended additional gastrectomy for patients that underwent non-curative ESD.

In our study, the rate of RC was 9.6% (7/73). The depth of invasion (≥SM2), positive horizontal margin, and ESD margin were associated with RC in the univariate analysis (P=0.045, P<0.001, and P=0.001). However, these factors were not statistically significant in the multivariate analysis. In previous studies of patients who underwent additional gastrectomy after ESD, RC rates ranged from 7.7% to 22.2%. Positive horizontal margin, positive vertical margin, ulceration, lymphatic invasion, and neural invasion were reported as risk factors for RC [11–14]. In our cases, RC did not exist in the negative horizontal margin. However, Park et al. [15] reported the possibility of remnant malignant cells remaining on the negative resection margin because the coagulated portion of the resection margin could not be assessed accurately. In addition, RC may be considered to be cancer cells that spilled out from the vessels due to vascular stasis [9].

LNM rates for patients who underwent additional gastrectomy after non-curative ESD ranged from 7.5% to 17.8% [12–14,16]. Depth of invasion, lymphatic invasion, vascular invasion, and differentiation were the associated risk factors for LNM [12,13,16]. In our study, the LNM rate after additional gastrectomy was 8.2% (6/73), and lymphatic invasion was the only risk factor for LNM (P=0.012). Yamao et al. [17] reported that lymphatic invasion is an independent risk factor for LNM. Furthermore, Kamata et al. [18] demonstrated that matrix metalloproteinase-1 secreted from gastric cancer cells plays an important role in LNM as a sequela of lymphatic invasion.

There are some limitations to our study. This study was conducted retrospectively at a single institution. Because the number of cases included in this study was small, the evidence regarding the risk factors identified in this study is insufficient. Further multicenter data and prospective studies will be able to validate our findings and establish the risk factors definitively.

In conclusion, additional gastrectomy is necessary to prevent RC or LNM after non-curative ESD. Lymphatic invasion was also associated with LNM in patients who underwent an additional gastrectomy after a non-curative ESD, and in such cases, active treatment is required.