Introduction

The Japanese Gastric Cancer Association (JGCA) reported the results of a nationwide gastric cancer registry in 2006 [1], 2011 [2], and 2013 [3]. The patients were restricted to those who underwent surgical therapy. Since then, remarkable progress has been made in endoscopic treatment of early gastric cancer, which is now performed in increasing numbers of patients. The conventional procedure of endoscopic mucosal resection (EMR) was used as the basis for the development of a procedure for endoscopic submucosal dissection (ESD) [4, 5], and the indications for endoscopic treatment have been expanded [6]. We report the short-term and long-term outcomes of registered patients with gastric cancer who underwent endoscopic treatment from 2004 through 2006.

Materials and methods

The Registration Committee of the JGCA requested hospitals affiliated with JGCA members to enroll consecutive patients with early gastric cancer who underwent EMR or ESD from January 2004 through December 2006. As shown in Table 1, the numbers of participating hospitals and enrolled patients have increased year by year. One hundred twenty-six hospitals participated, and 12,647 patients were enrolled during the 3 years. The Institutional Review Boards of each participating hospital approved the study protocol. In patients with multiple lesions, the main lesion was registered. The main lesion was defined by tumor depth. If multiple lesions had the same tumor depth, the main lesion was defined by tumor size. The tumor size was based on the histology results. Curability criteria were in accordance with the gastric cancer treatment guidelines described later. All tumors that were resected in a piecemeal fashion were classified as noncurative resection.

Table 1 Registered institutions and patients by year
Full size table

This registry used a computerized database with terminology in accordance with the Japanese Classification of Gastric Carcinoma (13th and 14th editions) and the Japanese Gastric Cancer Treatment Guidelines from 2010.

This nationwide registration program was approved by the Ethics Committee of the JGCA.

Results

Table 2 shows the characteristics of the patients and lesions. The mean age was 71 years. Men accounted for 76% of the registered patients. The endoscopic treatment procedure was ESD in 10,259 patients (81.1%), EMR in 2355 patients (18.6%), and other procedures in 30 patients (0.2%). The gastric status was an intact stomach in 96% of the patients, but some patients had a remnant stomach after gastrectomy or a reconstructed gastric tube after esophagectomy. More than 80% of the lesions were located in the region from the middle to the lower part of the stomach. The largest proportion of lesions was located along the lesser curvature of the stomach. With regard to the macroscopic type, 50% of the lesions were depressed type, 40% were elevated type, and 9% were mixed type. Histologically, 95% of the lesions were differentiated-type cancer. The median tumor diameter was 15 mm and the tumor diameter ranged from 0.5 to 169 mm. The depth of tumor invasion extended to the mucosa in 84.1% of the patients, and was less than 500 μm from the muscularis mucosae in 7.9% of the patients and 500 μm or more from the muscularis mucosae in 7.2% of the patients. With regard to lymphovascular invasion, 4.5% of the patients had lymphatic invasion and 2.3% had venous invasion. Ulcers (including open ulcers and ulcer scars) were found in 10.5% of the patients.

Table 2 Patient (n = 12,647) and lesion characteristics
Full size table

Short-term outcomes of endoscopic resection

With regard to short-term outcomes (Table 3), the en bloc resection rate was 89.3%. Horizontal margins were negative in 81.6% of patients, and vertical margins were negative in 94.5%. The rate of en bloc resection with negative margins was 78.9%. The total proportion of patients who underwent curative resection was 69.2%; 43.8% of patients underwent curative resection for absolute indication lesions, and 25.4% underwent curative resection for expanded indication lesions. Noncurative resection was performed in 3704 patients (29.3%). Of the patients who underwent noncurative resection, 63.2% received no additional treatment, 25.2% underwent surgical gastrectomy, and 3.2% underwent additional endoscopic resection. Endoscopic ablation was performed in 155 patients (4.2%). Emergency surgery was performed to treat bleeding in 38 patients (0.3%) and perforation in 45 patients (0.4%). The short-term outcomes are shown according to the method of endoscopic treatment in Table 4. The rate of en bloc resection was significantly higher for ESD (94.5%) than for EMR (66.8%; p < 0.01). The rate of complete en bloc resection was significantly higher for ESD (86.0%) than for EMR (48.2%; p < 0.01). The curative resection rate for absolute indication lesions and expanded indication lesions combined was significantly higher for ESD (75.1%) than for EMR (43.7%; p < 0.01). The rate of emergency surgery for complications did not differ significantly between ESD and EMR (0.3% vs 0.4% for bleeding and 0.3% vs 0.4% for perforation).

Table 3 Short-term outcomes (n = 12,647 patients)
Full size table
Table 4 Short-term outcomes according to the method of endoscopic resection
Full size table

Long-term outcomes of endoscopic resection

The 5-year follow-up rate after endoscopic resection was 70%. Figure 1 shows the 5-year overall survival rates according to curability. The 5-year overall survival rate was 91.6% after curative resection of absolute indication lesions and 90.3% after curative resection of expanded indication lesions, as compared with 86.5% after noncurative resection, which was slightly lower. Figure 2 shows the disease-specific survival rates according to curability. The 5-year disease-specific survival rate was 99.9% after curative resection of absolute indication lesions and 99.7% after curative resection of expanded indication lesions, indicating good outcomes. In patients who underwent noncurative resection, the 5-year disease-specific survival rate was 98.7%, indicating good outcomes.

Fig. 1
figure 1

Overall survival according to curability

Full size image
Fig. 2
figure 2

Disease-specific survival according to curability

Full size image

Discussion

EMR [7] was developed in the 1980s as an endoscopic treatment for early gastric cancer. The development of EMR made possible the endoscopic resection of small differentiated mucosal carcinomas measuring 2 cm or less in diameter. However, EMR often had to be performed in a piecemeal fashion because of technical limitations [8], and histopathological evaluation was difficult in an appreciable number of patients. In the late 1990s, ESD was developed to make possible the en bloc resection of larger lesions [4] and rapidly became popular. Gotoda et al. [6] attempted to expand the clinical indications for ESD in accordance with the positioning of ESD in clinical trials. However, the long-term outcomes of ESD for expanded indication lesions remain unclear. Although ESD has several advantages, such as a high rate of en bloc resection and accurate histopathological assessment, it also has drawbacks when compared with EMR, including a higher incidence of complications such as posttreatment bleeding and perforation [9]; moreover, a longer time is required to master the technique for ESD as compared with that for EMR.

Oda et al. [10] retrospectively studied 714 patients who underwent endoscopic resection of early gastric cancer in 11 Japanese hospitals in 2001. During the study, EMR was performed in more than half of the patients (EMR in 411 patients and ESD in 303 patients). In the present study, we compiled data on patients who underwent endoscopic treatment of early gastric cancer in more than 100 Japanese hospitals from 2004 through 2006. To our knowledge, studies of a similar size have not been reported previously. ESD for treatment of early gastric cancer was approved for coverage by the Japanese National Health Insurance in April 2006 and was soon performed in more than 80% of patients, indicating its rapid acceptance. ESD is thus considered a very therapeutically useful procedure.

With regard to short-term outcomes, the rates of en bloc resection and complete en bloc resection achieved by ESD (95% and 86% respectively) were significantly higher than those achieved by EMR (67% and 48% respectively). The rate of conversion to open surgery because of complications such as bleeding and perforation did not differ between ESD and EMR. The incidences of late bleeding and perforation have been reported to be higher for ESD than for EMR [9], but some studies found no difference in complication rates between ESD and EMR after the procedure for ESD had been technically mastered [11, 12]. In our study, the proportions of patients who had serious complications requiring surgery did not differ significantly between ESD and EMR.

With regard to long-term outcomes according to curability, several studies have reported highly favorable 5-year overall and disease-specific survival rates in patients with early gastric cancer who underwent curative ESD [13, 14]. In our study, the 5-year disease-specific survival rate was nearly 100% in patients who underwent curative resection for absolute indication lesions or expanded indication lesions. In studies comparing ESD with surgery, the overall survival rate after ESD was similar to that after surgery, and the benefit of ESD as compared with surgery included fewer late complications and a shorter hospital stay [15, 16].

Outcomes in patients with expanded indication lesions have been sporadically reported by various groups of investigators, and many studies showed no difference in outcomes in comparison with patients who had absolute indication lesions [14, 17, 18]. These studies were conducted in relatively small numbers of patients at single centers. In our study, 5 (0.16%) of 3056 patients with expanded indication lesions that were curatively resected by ESD died of gastric cancer. In another multicenter collaborative study that we conducted, 6 (0.14%) of 4202 patients with expanded indication lesions had metastatic recurrence [19]. Prospective studies of expanded indication lesions, including the JCOG0607 study [20], the JCOG1009 study [21], and the J-WEB/EGC study [22], have been conducted, and the results of these studies are awaited.

The results of previous studies suggest that expanded indication lesions might be associated with extremely low risks of metastatic recurrence and death from gastric cancer. Kikuchi et al. [23] reported that 3 (0.35%) of 851 patients with early gastric cancer died after surgical therapy, indicating that surgery is also associated with a certain level of risk. Given this background, endoscopic therapy is considered one of the treatment options for expanded indication lesions provided that informed consent is obtained from patients after they are given an adequate explanation about the risk of metastatic recurrence, albeit the incidence is less than 1%.

This is the first time the results of a nationwide study performed by the JGCA that analyzed the outcomes of endoscopic treatment of early gastric cancer have been reported. We compiled and analyzed data on more than 12,000 patients from 126 Japanese hospitals. Our results indicate that endoscopic treatment had good short-term and long-term outcomes in patients with early gastric cancer. The present study was performed during the shift from EMR to ESD and compared the treatment outcomes of these procedures. We therefore consider our results to be valuable. Future participation of more hospitals in nationwide registration may allow various new findings, including the long-term outcomes of patients with expanded indication lesions, to be obtained. On the other hand, the proportion of patients who could be followed up for at least 5 years was only 70%, which was not satisfactory. We will attempt to increase the follow-up rate in future studies and thereby report more accurate results.