Article Text

Download PDFPDF

Original article
Scheduled endoscopic surveillance controls secondary cancer after curative endoscopic resection for early gastric cancer: a multicentre retrospective cohort study by Osaka University ESD study group
  1. Motohiko Kato1,
  2. Tsutomu Nishida1,
  3. Katsumi Yamamoto2,
  4. Shiro Hayashi2,
  5. Shinji Kitamura3,
  6. Takamasa Yabuta3,
  7. Toshiyuki Yoshio4,
  8. Takeshi Nakamura5,
  9. Masato Komori6,
  10. Naoki Kawai7,
  11. Akihiro Nishihara8,
  12. Fumihiko Nakanishi9,
  13. Masanori Nakahara10,
  14. Hideharu Ogiyama11,
  15. Kazuo Kinoshita12,
  16. Takuya Yamada1,
  17. Hideki Iijima1,
  18. Masahiko Tsujii1,
  19. Tetsuo Takehara1
  1. 1Department of Gastroenterology and Hepatology, Osaka University Graduate School of Medicine, Suita, Osaka, Japan
  2. 2Department of Gastroenterology, Toyonaka Municipal Hospital, Toyonaka, Osaka, Japan
  3. 3Department of Gastroenterology, Sakai City Hospital, Sakai, Osaka, Japan
  4. 4Department of Gastroenterology, Osaka National Hospital, National Hospital Organization, Osaka, Japan
  5. 5Department of Gastroenterology, Kansai Rosai Hospital, Amagasaki, Hyogo, Japan
  6. 6Department of Gastroenterology, Osaka Rosai Hospital, Sakai, Osaka, Japan
  7. 7Department of Gastroenterology, Osaka Police Hospital, Osaka, Osaka, Japan
  8. 8Department of Gastroenterology, Minoh City Hospital, Minoh, Osaka, Japan
  9. 9Department of Gastroenterology, Osaka Minami Medical Centre, National Hospital Organization, Kawachi-Nagano, Osaka, Japan
  10. 10Department of Gastroenterology, Ikeda Municipal Hospital, Ikeda, Osaka, Japan
  11. 11Department of Gastroenterology, Itami City Hospital, Itami, Japan
  12. 12Department of Gastroenterology, Sumitomo Hospital, Osaka, Japan
  1. Correspondence to Tetsuo Takehara, Department of Gastroenterology and Hepatology, Osaka University Graduate School of Medicine, 2-2 Yamadaoka, Suita, Osaka 565-0871 Japan; takehara{at}

Statistics from

Significance of this study

What is already known on this subject?

  • After endoscopic resection (ER) of early gastric cancer, patients are at high risk for developing synchronous or metachronous multiple gastric cancers.

What are the new findings?

  • Nineteen per cent of synchronous cancers were missed at the initial endoscopic submucosal dissection (ESD).

  • The incidence rate of multiple cancers after ESD was constant.

  • The incidence rate of massively invading cancer was extremely low under scheduled endoscopic surveillance.

How might it impact on clinical practice in the foreseeable future?

  • Scheduled endoscopic surveillance showed that almost all recurrent lesions were treatable by ER.


Gastric cancer is one of the most common neoplasms worldwide, accounting for over 870 000 new cases and more than 650 000 deaths annually.1 Early gastric cancer (EGC) has a good prognosis. If there is no concomitant lymph node metastasis, the 5-year gastric cancer-specific survival rate is 99% in Japan.2 Therefore, endoscopic resection (ER) is widely accepted as a local treatment for these lesions, because it preserves the entire stomach and the patient's quality of life (QoL) postoperatively.3 ,4 In particular, endoscopic submucosal dissection (ESD) achieves secure resection with an acceptable complication rate and the indication for its use is expanding.5–12

Gastric cancer generally arises from Helicobacter pylori (H pylori)-associated chronic gastritis and about 10–20% of patients with gastric cancer develop synchronous and metachronous multiple cancers.3 ,4 ,13–15 Preserving the stomach contributes to maintaining postoperative QoL but could increase the risk of multiple cancer development from the remnant gastric mucosa. Therefore, follow-up is important for detecting multiple cancers after ER. Although several studies have evaluated the long-term prognosis of ER,3 ,5 ,7 most lack important information, especially for the early phase after ER, because synchronous lesions were not included in the analyses. Although in some cases distinguishing between synchronous and metachronous cancers is difficult, it is clinically important to analyse all events to consider appropriate follow-up strategies.

The aim of this study was to elucidate the patterns of multiple cancer development from remnant gastric mucosa and to determine whether scheduled endoscopic surveillance can control multiple cancers after successful ER.

Patients and methods

Participating hospitals

This study is a multicentre retrospective cohort study of patients from 12 hospitals participating in the Osaka University ESD Study Group, which comprises one university hospital, five secondary care hospitals and six tertiary care hospitals (table 1). This study was approved by the Osaka University institutional review board and was performed in accordance with the Declaration of Helsinki (2008).

Table 1

Participating hospitals


Consecutive patients with gastric cancer who underwent curative ESD in the 12 hospitals between April 1999 and December 2010 were included in the study. The curability of the initial ESD was classified into the following three groups proposed by Gotoda et al16 based on the characteristics of the initially detected tumour: ‘guideline group’, ‘expanded guideline group’ and ‘non-curative group’. The guideline group was defined as mucosal differentiated cancer with the largest diameter measuring <20 mm. In Japan, ER is definitely indicated for this group. The expanded guideline group was defined as the following: (1) mucosal differentiated cancer measuring >20 mm in diameter, (2) mucosal differentiated cancer with ulceration and measuring <30 mm in the largest diameter and (3) differentiated cancer measuring <30 mm in the largest diameter with a submucosal invasion depth of <500 μm. If the lesions did not meet these criteria, they were classified as the non-curative group. The non-curative group was advised to undergo additional gastrectomy with lymph node dissection and was excluded from the data analysis, whereas both the guideline and expanded guideline groups were enrolled in the study. Moreover, the patients whose initial ESD was incomplete (piecemeal, margin-positive or unclear) were excluded from the study.

Definition of multiple cancer development

In this study, multiple gastric cancer development was classified as synchronous (within 1 year) or metachronous cancer (after 1 year) according to the time at which the multiple cancers developed. Moreover, synchronous cancer was classified as ‘concomitant cancer’ or ‘missed cancer’. Concomitant cancer was defined as multiple cancers that had already been detected and diagnosed before the initial ESD. Missed cancer was defined as cancer being detected within 1 year, but not detected at the time of the initial ESD. The ‘miss rate’ of the preoperative screening was defined as the proportion of missed cancers out of all synchronous cancers.

Preoperative screening

We investigated the methods of preoperative screening among the 12 hospitals by questionnaire. We asked two questions: ‘How often was each endoscopic modality (small-calibre endoscopy, high-vision endoscopy, magnified endoscopy, image-enhanced endoscopy and chromo endoscopy) used for preoperative screening?’ and ‘How much experience (years and number of cases) does an endoscopist need to be allowed to perform preoperative screening in your hospital?’

The outlines for preoperative screening at the participating hospitals are shown in figure 1.

Figure 1

The outlines of the preoperative evaluation at participating hospitals. Endoscopic modality (chromo endoscopy, image-enhanced endoscopy, high-vision endoscopy, magnified endoscopy, or small calibre endoscopy) was used. Data are presented as the number of hospitals (A). And the qualification required for performing preoperative evaluation (experience years and experienced oesophagogastroduodenoscopy cases) was investigated by questionnaire. Data are presented as the number of hospitals (B).


The follow-up protocols after ESD among the participating hospitals are shown in table 1. Oesophagogastroduodenoscopy (OGD) was started within 1, 3 and 6 months after the initial ESD in 30%, 41% and 100% of the subjects, respectively. Surveillance OGD was performed every 6–12 months. Abdominal CT was added for a final pathological diagnosis in the expanded guideline group.

Infection status and eradication of H pylori

Information on the H pylori infection status and eradication was collected from the medical records of each patient. A positive result on at least one rapid urease test, urease breath test, serum H pylori IgG antibody or H pylori stool antigen test was deemed to be indicative of H pylori infection.


The miss rate under various conditions of preoperative screening (volume of ESD cases in each hospital, endoscopic modality and the quantity of experience of the endoscopists) was compared using the χ2 test. A multivariate logistic regression analysis was performed by adjusting the variables with a value <0.10 by univariate analysis.

Various clinicopathological factors at the initial ESD were compared between patients with and without multiple gastric cancers. The χ2 test was used to analyse categorical variables and Student's t test was used to analyse continuous variables.

The cumulative incidence rate of metachronous cancers and overall survival rate were plotted using the Kaplan–Meier method. The observation period was measured from the date of the initial ESD to the date of detection of metachronous multiple cancers or the date of the latest endoscopic examination for patients in whom multiple cancers were not found. Local recurrence was defined as a multiple cancers that developed at, or adjacent to, the site of the initial ESD.

A p value <0.05 was considered statistically significant. Statistical analyses were performed using JMP software (V.8.0.1, SAS Institute Inc, Cary, North Carolina, USA).


Initial ESD

Between April 1999 and December 2010, ESD was performed on 1836 patients with a diagnosis of gastric epithelial neoplasm in 12 hospitals. Among them, 1471 patients had a final diagnosis of gastric cancer. Of these, 131 patients were placed in the non-curative group and 55 patients had a positive or unclear margin; these 186 patients were excluded from the study. Among the 1285 patients who were curatively treated by ESD, 27 were lost during the observation period and the remaining 1258 patients were analysed (figure 2). In our study, the prognosis-chasing rate was 97.9% and the mean observation period was 26.8 months.

Figure 2

Flow chart of patients. GL, guideline group; Ex-GL, extended guideline group.

Patient characteristics

During the observation period, 175 (13.9%) multiple cancers were detected. Of the 852 patients who were followed up for longer than 12 months, 385 (45%) and 163 (19%) were H pylori positive and H pylori negative, respectively. In the remaining 304 patients (36%), the H pylori infection status was not determined. The main reason for not determining the infection status was lack of health insurance coverage in Japan at that time. The characteristics of the initially resected lesions (location, size, macroscopic type, histological type, invasion depth and guideline classification) did not significantly differ between patients with and without multiple cancers (table 2).

Table 2

Characteristics of initially resected lesions

Synchronous cancers and miss rate of preoperative screening

Synchronous multiple cancers were seen in 110 patients (9%). Concomitant multiple cancer was detected in 89 patients and missed cancers accounted for the remaining 21. The miss rate was 19% (figure 3).

Figure 3

Flow chart of included patients. ESD, endoscopic submucosal dissection.

We further investigated the 21 missed cancers. The median time to detection of the missed cancers was 6.3 months (range 1.6–11.6). The mean diameter of the lesions was 21 mm (range 4–43). Seven lesions (33%) developed in the upper third of the stomach and 17 lesions (81%) developed in the lesser curvature or posterior wall. Three lesions invaded the submucosal layer massively and one lesion invaded the muscularis propria (see supplementary table S1).

We investigated the association between the methods of preoperative screening and miss rate by questionnaire. Univariate analysis showed that a lower volume of ESD cases at the hospital and endoscopists with fewer years of experience and fewer numbers of cases were significantly associated with the higher miss rate of preoperative screening (table 3). Multivariate analysis showed that only the endoscopist's experience (<500 cases) was an independent predictor of oversight for preoperative screening (table 4).

Table 3

Univariate analysis of the association between preoperative screening and the miss rate

Table 4

Multivariate analysis of the association between preoperative evaluation and the miss rate

Cumulative incidence of metachronous cancers and overall survival rate

The cumulative incidence curve of metachronous cancers is plotted in figure 4. The cumulative incidence curve revealed a linear increase. The cumulative incidence rate of metachronous cancers at 2, 3, 4 and 5 years was 3.7%, 6.9%, 10% and 16%, respectively. Local recurrence was seen in five cases with an incidence rate of 0.40%.

Figure 4

Cumulative incidence of metachronous cancers after endoscopic submucosal dissection.

To examine whether H pylori eradication could decrease the incidence of multiple cancers, we compared the incidence of multiple cancers between patients with and without H pylori eradication among H pylori-positive patients. Among the 385 H pylori-positive patients, H pylori eradication treatment was given to 322 (84%) and was not given to the remaining 63 (16%). Successful eradication was confirmed in 263/322 (82%) patients. The incidence of multiple cancers in patients who underwent successful H pylori eradication was not decreased compared with that of those who did not receive eradication or for whom eradication failed (figure 5). Moreover, the incidence rate did not differ between the patients in hospitals with more precocious endoscopy (1 month after the initial ESD) from the others (see supplementary figure S1).

Figure 5

Cumulative incidence curve of metachronous cancers between patients with successfully eradicated Helicobacter pylori (HP) and patients without eradication.

The overall survival is plotted in figure 6. Treatment-related deaths did not occur in this study. The 3-year overall survival rate was 96.8% and death due to gastric cancer did not occur during the follow-up period.

Figure 6

Cumulative overall survival curve. There was no treatment-related death in this study. The 3-year overall survival and cause-specific survival rates were 96.8% and 100%, respectively.

Secondary intervention

The secondary intervention protocols are described in figure 7. Of the 175 multiple cancers, 169 were treated by re-ESD. Among these, 164 lesions were diagnosed as belonging to the guideline or expanded guideline group and were followed up without additional treatment. Of the remaining five lesions, two were diagnosed as mucosal undifferentiated adenocarcinoma and three were diagnosed as submucosal cancers after ESD; these patients underwent additional gastrectomy. In addition, six lesions were treated by gastrectomy. Of these, four were pathologically diagnosed as belonging to the guideline or expanded guideline group after gastrectomy and the remaining two were pathologically diagnosed as non-curative. Taken together, seven lesions were diagnosed as non-curative: three were intramucosal undifferentiated cancers and four were massively invading cancers.

Figure 7

Flow chart of the secondary intervention. ESD, endoscopic submucosal dissection; Ex-GL, extended guideline group; GL, guideline group; M, mucosal cancer; MP, cancer invading the muscularis propria; PD, poorly differentiated adenocarcinoma; SM, cancer invading the submucosa massively (>500 μm); WD, well-differentiated adenocarcinoma.


To the best of our knowledge, this is the first multicentre study of long-term outcomes after ESD for EGC. In this study, 1258 patients with EGC successfully treated by ESD were analysed, making this study one of the largest to date.

The postoperative 5-year gastric cancer-specific survival rate of EGC without concomitant lymph node metastasis is approximately 99% in Japan.2 ,17 ,18 ER is therefore widely accepted and applied as a standard treatment for EGC with a negligible risk for lymph node metastasis.3–8 ER can preserve the stomach and contribute to maintaining postoperative QoL. However, EGC is often accompanied by synchronous or metachronous multiple gastric cancers.3 ,4 ,13–15 Therefore, preserving the stomach increases the risk of multiple cancer development from the remnant gastric mucosa. If secondary lesions are not detected at an early stage, gastrectomy is required even though it was avoided for the initial lesion. Therefore, establishing an optimal surveillance strategy is important to reduce the gastric cancer death rate but also to detect multiple cancers during the early stage and avoid surgery for the secondary cancer.

There are three important factors in the development of multiple gastric cancers. First, the appropriate method of preoperative screening should be established. Lee et al reported that about 15% of concomitant lesions are missed by endoscopic screening.19 To establish a surveillance strategy, the details of these missed lesions should be investigated. Second, local recurrence should be distinguished from multicentre carcinogenesis. Local recurrence is strongly influenced by whether the initial lesion was completely resected. Third, the analysis should be based on findings of a population that underwent scheduled surveillance OGD after ER because EGC is generally asymptomatic and endoscopic confirmation is essential.

This large-scale multicentre retrospective cohort study has three important strengths. First, we clearly distinguished missed cancers from clinically diagnosed concomitant cancers (ie, those detected before the initial ESD) among the synchronous cancers. This distinction makes it possible to discuss the miss rate of the preoperative screening. Second, we included only subjects whose initial ESD was pathologically confirmed as an en bloc margin-negative curative resection. In our study, there were five local recurrences even though the initial ESD was a pathologically confirmed R0 resection. Although we could not distinguish whether these lesions coincidentally developed near the ESD scar or were residual lesions, the incidence rate of local recurrence was extremely low at only 0.40%. Third, the follow-up rate was approximately 98%. All the subjects underwent follow-up OGDs and the recurrence or absence of recurrence was endoscopically confirmed. In view of these strengths, we can discuss the details of multiple cancer development.

Our results provide important information about synchronous cancers and their oversight during preoperative screening. In this study, 9% of the patients who underwent ESD had synchronous gastric cancers and 19% of these were missed by the preoperative endoscopic evaluation. Of these 21 missed lesions, four were massively invading cancers (including one advanced cancer), which suggests that careful preoperative screening is needed and that the first follow-up OGD should be performed within 6 months of the ESD to detect missed invasive cancers. Many missed cancers were located in the upper third of the stomach and at the circumference of the lesser curvature or posterior wall. These locations tend to be in a tangential direction and constitute a blind area for endoscopic observation. The miss rate was associated with the extent of the endoscopist's experience, suggesting that preoperative screening should be performed only by endoscopists who have performed OGD for at least 500 cases. However, high vision endoscopy, image-enhanced endoscopy and magnified endoscopy did not affect the miss rate and may not be mandatory for the preoperative detection of concomitant multiple lesions.

Previous studies reported that the incidence of multiple gastric cancers after ESD for EGC was about 3%,3 ,13 whereas in our study, the incidence rate of metachronous cancers was 3.5% a year. This study analysed a larger number of subjects than previous studies and R0 resection only, to exclude local recurrence from secondary cancers, suggesting that our study added reliability to the previous studies. The incidence rate of multiple cancers in patients with atrophic gastritis or intestinal metaplasia was reported as 0.1–0.5%,20 ,21 showing that the patients after ER for EGC had a higher risk of multiple cancer development than those with atrophic gastritis or intestinal metaplasia. Therefore, careful follow-up is needed for patients who have undergone ESD for EGC.

In this study, surveillance OGD was performed at least once a year after the ESD at all participating hospitals. Massively invading cancer detected more than 1 year after ESD occurred in only one of 852 (0.12%) cases. Based on these findings, we recommend that annual surveillance be performed.

Although a multicentre randomised controlled trial showed that H pylori eradication reduces multifocal carcinogenesis of the stomach,13 we detected no difference in the cumulative incidence rate of multiple cancers between patients who underwent successful H pylori eradication and those who did not receive eradication treatment or for whom eradication failed. This discrepancy might be due to differences in the study designs (Fukase's prospective randomised trial compared with our retrospective cohort study) and different calculations of the observation period (Fukase's from eradication of H pylori compared with ours from the initial ESD).

Our study has several limitations. First, the follow-up OGD interval varied among the participating hospitals, although an annual follow-up at least was performed. Second, the mean observation period was comparatively short, being <3 years in our study. Although our study population was large and contained more than 350 cases with an observation period >3 years, the long-term outcomes should be interpreted carefully. Third, the information on H pylori infection and eradication was incomplete because these data were collected retrospectively from medical records.

In Japan, we have universal health coverage owing to the establishment of employee-based and community-based social health insurance,22 which may enable screening and surveillance endoscopies in Japan to be performed in a more strict and standardised fashion than in other countries. Nevertheless, about 20% of synchronous cancers were not detected. We would emphasise the importance of preoperative evaluation and post-ESD surveillance.

Based on the results of this study, we recommend the following: (1) preoperative screening should be performed by an endoscopist who has performed at least 500 OGD cases; (2) intensive (every 6 months) surveillance is preferred in the first year after ER to detect missed concomitant invasive cancers and (3) annual surveillance should be performed for at least 5 years after the ER. In view of the previously described limitations, the validity of our recommendations should be confirmed by a prospective study.

In conclusion, the findings of our large-scale, multicentre retrospective cohort study indicated that the incidence rate of synchronous cancer was 9%, that about 20% of synchronous cancers were missed and that the incidence of metachronous cancer increased linearly. With an annual follow-up examination, almost all multiple lesions could be cured by ER.


Study investigators; Yoshito Hayashi, Shunsuke Yamamoto, Tomofumi Akasaka (Osaka University Hospital); Tasuku Nakabori (Toyonaka Municipal Hospital); Akira Maekawa, Aya Ishimi, Fumiaki Itakura, Kanako Abe, Satoshi Egawa (Kansai Rosai Hospital); Yuka Watanabe, Mamoru Yura (Minoh City Hospital); Daisuke Utsunomiya, Rui Mizumoto, Yasushi Matsumoto, Shinichiro Zushi (Ikeda Municipal Hospital); Masako Sato (Osaka Rosai Hospital); Osamu Kishida and Takashi Fujimoto (Sumitomo Hospital).


View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Digest
    Emad El-Omar William Grady Alexander Gerbes Thomas Rösch