The registration committee of the Japanese Gastric Cancer Association (JGCA) started a new registration program in 2008 after a 10-year blank period, and we reported the 5-year follow-up data of the patients treated in 2001 [1]. The registration has been continuing, and here we report the results of those treated in 2002.

Materials and methods

Leading hospitals in Japan voluntarily downloaded and fulfilled the database provided by the JGCA and sent the anonymized data to the JGCA data center. The collected data were analyzed according to the previously reported methods [1].


Data of 14,394 patients were collected from 208 hospitals; 126 (60.6 %) hospitals participated in both years, but 82 hospitals were new, which was a 10 % increase as compared to the previous year (13,067 patients from 187 hospitals). The geographic distribution of the registered patients among the 47 prefectures is illustrated in Fig. 1. In Tokyo, 2,332 patients per year were registered, followed by 1,464 in Osaka. Four other prefectures registered more than 500 patients. On the other hand, the number of registered patients was fewer than 100 in 10 prefectures, and there were no registered patients in 2 prefectures.

Fig. 1
figure 1

Geographic distribution of registered patients by prefecture

Patients with remnant stomach cancer, non-epithelial malignant tumor, and gastric cancer combined with malignant tumor of other organs were excluded. Patients who were treated by endoscopic mucosal resection were also excluded. Data of 768 patients lacked essential items. Consequently, data of the remaining 13,002 patients were used for the final analysis.

The results are shown in Tables 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, and 28. Data given for each category of patients are: total number of patients, survival rates by year, standard error of 5YEARS, the number of direct death within 30 postoperative days, the number of patients lost to follow-up within 5 years, the number of 5-year survivors, and main cause of death, such as local and/or lymph node metastasis, peritoneal metastasis, liver metastasis, distant metastasis, recurrence at unknown site, other cancer, and other disease. Figures 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, and 17 provide cumulative survival curves of patients stratified by essential categories.

Table 1 Primary cancer
Table 2 Resected cases and unresected cases and other surgeries
Table 3 Sex (resected cases)
Table 4 Age (resected cases)
Table 5 Tumor location (resected cases)
Table 6 Macroscopic type (resected cases)
Table 7 Histological type (resected cases)
Table 8 Histological findings (resected cases)
Table 9 Lymphatic invasion(ly) (resected cases)
Table 10 Venous invasion(v) (resected cases)
Table 11 Depth of invasion (resected cases)
Table 12 pT classification (resected cases)
Table 13 Lymph node metastasis (resected cases)
Table 14 Peritoneal cytology (resected cases)
Table 15 Peritoneal metastasis (P) (resected cases)
Table 16 Liver metastasis (H) (resected cases)
Table 17 Distant metastasis including peritoneal and liver metastasis (M) (resected cases)
Table 18 Japanese stage (resected cases)
Table 19 Japanese stage (resected cases)
Table 20 TNM stage (resected cases)
Table 21 TNM stage (resected cases)
Table 22 Surgical approach (resected cases)
Table 23 Surgical procedures (resected cases)
Table 24 Lymph node dissection (D) (resected cases)
Table 25 Resection margins (resected cases)
Table 26 Combined resection of neighboring organs (resected cases)
Table 27 Combined resected organs (resected cases)
Table 28 Curative potential (resected cases)
Fig. 2
figure 2

Kaplan–Meier survival for all patients with primary gastric cancer. 5YEARS 5-year survival rate

Fig. 3
figure 3

Kaplan–Meier survival for resected cases and unresected cases

Fig. 4
figure 4

Kaplan–Meier survival of resected cases stratified by sex

Fig. 5
figure 5

Kaplan–Meier survival of resected cases stratified by age

Fig. 6
figure 6

Kaplan–Meier survival of resected cases stratified by tumor location. W whole stomach

Fig. 7
figure 7

Kaplan–Meier survival of resected cases stratified by macroscopic type

Fig. 8
figure 8

Kaplan–Meier survival of resected cases stratified by histological findings

Fig. 9
figure 9

Kaplan–Meier survival of resected cases stratified by lymphatic invasion

Fig. 10
figure 10

Kaplan–Meier survival of resected cases stratified by depth of tumor invasion

Fig. 11
figure 11

Kaplan–Meier survival of resected cases stratified by pT classification

Fig. 12
figure 12

Kaplan–Meier survival of resected cases stratified by lymph node metastasis

Fig. 13
figure 13

Kaplan–Meier survival of resected cases stratified by peritoneal cytology

Fig. 14
figure 14

Kaplan–Meier survival of resected cases stratified by peritoneal metastasis

Fig. 15
figure 15

Kaplan–Meier survival of resected cases stratified by Japanese Gastric Cancer Association (JGCA) stage

Fig. 16
figure 16

Kaplan–Meier survival of resected cases stratified by TNM stage

The 5YEARS in 13,626 patients with primary gastric cancer was 68.9 % (Table 1; Fig. 2). During the 5-year follow-up, 2,233 patients were lost; the follow-up rate was 83.6 %. Of the 13,626 patients, 13,002 underwent gastric resection. Accordingly, the resection rate was 95.4 %, and the 5YEARS of the resected patients was 70.7 % (Table 2; Fig. 3). Sixty-three of 13,002 resected cases died within 30 days postoperatively. The direct death rate was 0.48 %. The frequent causes of death in patients who had undergone gastrectomy were peritoneal metastasis (n = 1,283), followed by other diseases (n = 539), local recurrence including node metastasis (n = 410), liver metastasis (n = 357), recurrence at unknown site (n = 278), and other cancer (n = 158).

The proportion of male patients was 68.4 % with 5YEARS of 70.0 %; for female patients 5YEARS was 72.3 %, which was better statistically (Table 3; Fig. 4). Patients more than 80 years old were 7.8 % of all patients, and their 5YEARS was 51.4 % (Table 4; Fig. 5). On the other hand, 5YEARS of the patients under 39 years old was 79.4 % (P < 0.001). Cancer was located in the upper-third of the stomach in 21.1 % of the cases, and its 5YEARS was relatively low at 64.3 % (Table 5; Fig. 6). Patients with type 4 cancer amounted to 7.2 %, and their 5YEARS was markedly low at 17.7 % (Table 6; Fig. 7). The 5YEARS of type 3 was 46.0 % and that of type 2 was 60.4 %. For histological type, frequency of the undifferentiated type including poorly differentiated adenocarcinoma, signet-ring cell carcinoma, and mucinous adenocarcinoma was 46.8 % and its 5YEARS was 65.5 %, which was inferior to that of the differentiated type (75.7 %, P < 0.001; Tables 7, 8; Fig. 8). The grade of lymphatic invasion (ly0–ly3) and venous invasion (v0–v3) showed significant correlations with the prognosis (Tables 9, 10; Fig. 9).

A high incidence of early-stage cancer remained characteristic in 2002, as shown in Tables 11 and 12. The proportion of pathological M and SM (pT1) cancer was 49.7 %, and its primary cause of death was not cancer recurrence (17.9 %, n = 87) or other cancer (18.7 %), but other diseases (49.0 %, n = 238). The proportion of pathological MP and SS (pT2) was 26.8 %, SE (pT3) 19.9 %, and SI (pT4) 3.6 %. The 5YEARS of these subsets were 67.9 %, 30.3 %, and 20.6 %, respectively (Figs. 10, 11). The primary cause of death in advanced cancer was cancer recurrence, and the peritoneal recurrence rate was remarkably high in the pT3 and pT4 subsets. For the lymph node metastasis, the proportion of pN0 was 59.5 %, pN1 20.4 %, pN2 15.9 %, and pN3 4.1 %, and the 5YEARS of each subset was 88.9 %, 58.9 %, 34.6 %, and 14.3 %, respectively (Table 13; Fig. 12).

Peritoneal washing cytology was carried out in 5,836 patients with advanced gastric cancer; the positive rate was 13.0 %. The 5YEARS of cytology-positive (CY1) patients was 12.3 %, which was almost as dismal as the 5YEARS of the P1 patients (8.3 %; Tables 14, 15; Figs. 13, 14). The 5YEARS of patients with liver metastasis (H1) was 11.4 %, and of those with other types of distant metastasis was 12.4 % (Tables 16, 17).

The 5YEARS of the patients stratified by JGCA staging system was 92.2 % for stage IA, 85.3 % for stage IB, 72.1 % for stage II, 52.8 % for stage IIIA, 31.0 % for stage IIIB, and 14.9 % for stage IV. These JGCA 5YEARSs seemed to correlate well with TNM 5YEARSs, which were 92.3 % for stage IA, 84.7 % for stage IB, 70.0 % for stage II, 46.8 % for stage IIIA, 28.8 % for stage IIIB, and 15.2 % for stage IV (Table 18, 19, 20, 21; Figs. 15, 16).

For operative procedures, the proportion of patients who underwent laparoscopic gastrectomy was only 5.1 % in 2002, and their 5YEARS was 93.3 % (Table 22). Eligibility for laparoscopic surgery was strictly limited at that time, and the laparoscopic approach was selected almost exclusively in patients with the preoperative diagnosis of early gastric cancer. Only 1.2 % of the patients were treated by thoracolaparotomy, and their 5YEARS was 35.4 %. Thoracolaparotomy was usually carried out in patients with advanced gastric cancer with esophageal invasion more than 3 cm in length. Total gastrectomy was performed for 30.5 % of the patients, and their 5YEARS was 51.9 % (Table 23). D2 lymph node dissection, a standard procedure for resectable advanced gastric cancer according to the JGCA treatment guidelines, was performed in 49.2 % of the patients (Table 24) [2, 3]. The risk of direct death among those who underwent D2 gastrectomy was only 0.3 %. The proportion of patients treated with less invasive surgery such as proximal gastrectomy, pylorus-preserving gastrectomy, segmental gastrectomy, and local resection of the stomach was 9.8 %. D0, D1, D1 + α, and D1 + β dissection were carried out in 6.5 %, 20.7 %, 13.7 %, and 7.2 % of the patients, respectively. D0 and D1 dissection were carried out mainly in patients with noncurative factors or poor surgical risks. The incidence of positive resection margin (PM+ and/or DM+) was 3.1 % (Table 25). Combined resection of other organs was performed in 35.1 % (Table 26). The frequent combined resected organs in patients who underwent gastrectomy were gallbladder (n = 2121), spleen (n = 1444), caudal pancreas (n = 313), transverse colon (n = 101), liver (n = 96), and so on in descending order (Table 27).

The curative potential of gastric resection was an important prognostic factor. The proportion of patients with no residual tumors with high probability of cure (resection A) was 63.9 %, and their 5YEARS was 88.6 %. On the other hand, patients with definite residual tumors (resection C) amounted to 11.9 % of all patients who underwent laparotomy, and their 5YEARS was 9.9 % (Table 28; Fig. 17).

Fig. 17
figure 17

Kaplan–Meier survival of resected cases stratified by curative potential of gastric resection


Estimates of the worldwide incidence, mortality, and prevalence of 26 cancers in the year 2002 were available in the GLOBOCAN series of the International Agency for Research on Cancer [4]. With an estimated 934,000 new cases per year in 2002 (8.6 % of new cancer cases), the incidence of stomach cancer is in fourth place, after cancers of the lung, breast, and colon and rectum. It is the second most common cause of death from cancer (700,000 deaths annually).

The data presented in this report were collected from 208 hospitals in Japan. Cancer incidence rate (annual number of newly diagnosed cases per 100,000 population) in Japan in 2002 was approximately 520 for males and 370 for females. The incidences of various cancers in Japan are estimated from data collected by the cancer registry system in a dozen prefectures. According to these statistics, the number of cancer incidences in 2002 was approximately 589,000. The stomach was the leading site (21 %) for males and the second highest site (14 %) for females. The number of new patients who were diagnosed as gastric cancer in 2002 was estimated to be 106,760 [5]. Accordingly, 13,626 patients registered by this program corresponded to approximately 13 % of the whole population affected by gastric cancer in Japan. Even though these patients may not represent the average features of gastric cancer found in this country, this report is considered to have analyzed the largest number of patients for the past 10 years, clarifying the trends of gastric cancer in Japan. Just for reference, the proportion of patients registered in the nationwide registry of other organs of all patients diagnosed were 6 % in colon cancer, 24 % in esophageal cancer, 25 % in liver cancer, and 26 % in lung cancer, respectively [6].

The reliability of the results in this report depends on the quality of data accumulated in the JGCA database. Because of the complexity of the JGCA staging system, the error checking system on the data entry screen did not function completely. In several categories such as lymph node metastasis (N), the JGCA system could not be converted to the TNM system automatically. Therefore, the registration committee had to make great efforts to confirm raw data sent to the data center from the participating hospitals.

As compared with our archived data of 12,004 patients treated in 2001 [1], the proportion of early cancer declined from 51.2 % to 49.7 % [pT1 (M) cancer, 27.4 % to 25.6 %, and pT1 (SM) cancer, 23.8 % to 24.1 %], suggesting that an increasing number of patients with mucosal cancer were sent for endoscopic treatment. These data suggest that we should start to register gastric patients treated with endoscopic mucosal resection (EMR) and/or endoscopic submucosal dissection (ESD) as soon as possible. The surgical mortality within 30 days significantly improved, from 0.6 % to 0.48 % (P < 0.001). Just for reference, it was 4.0 % in 1963 and 1.0 % in 1991 [7], Moreover, the nationwide database of gastrointestinal surgery in 2008 showed that was 0.2 % in gastrectomy and 0.4 % in total gastrectomy [8].

Accordingly with the rapidly aging society in Japan, the proportion of patients more than 80 years old continued increasing (Fig. 18): it was 0.7 % in 1963, 4.9 % in 1990, 7.0 % in 2001, and 7.8 % in 2002, respectively. Although the risk for surgery increases in elderly patients who have comorbidities, evaluations of risk can allow interventions that may decrease morbidity and mortality. Appropriate treatments should be offered to the elderly. However, these data have the intrinsic weakness of being retrospectively collected 7 years after surgery. Unfortunately, we in Japan continue to have a legal difficulty in registering personal information, which is essential for long-term and prospective follow-up. The overall follow-up rate in our program was 83.5 %. In other words, the outcome of 17.5 % of the patients is unknown. The proportion of patients who were lost to follow-up in the Japanese nationwide registry of colon cancer, liver cancer, and thyroid cancer was 19.6 %, 25.8 %, and 20.6 %, respectively [6]. Rules and regulations regarding handling of these data will have to change radically to overcome the issue of accuracy and reliability of the nationwide registry in Japan, and this could be out of the hands of the surgeons who have contributed to the best of their abilities to gather these data. On the other hand, the Japanese Association of Clinical Cancer Centers, consisting of 25 cancer center hospitals, reported that their follow-up rate was 98.5 %, and 5YEARS of 9,980 patients who underwent surgery from 1997 to 2000 were 90.4 % for TNM stage I, 67.8 % for stage II, 43.3 % for stage III, and 9.3 % for stage IV, respectively [9]. When the patients with gastric cancer had a medical examination in clinical cancer centers, they registered the place where their family records were registered, and office workers of the clinical cancer centers confirmed regularly their safety from the family registration; this was the reason for the extremely high follow-up rate. In the current analyses, 5YEARS in stage IV patients was 15.2 %. We might have overestimated our 5YEARS in stage IV patients, but we found that our follow-up rate increased as the stage advanced; the follow-up rate of stage IV patients was 90.4 %. These data suggest that the lower follow-up rate may not have had serious effects on 5YEARSs in our program. Although, the correlation between follow-up rate and survival rate is complicated, our follow-up system needs to be improved if we are to evaluate the survival rates more accurately.

Fig. 18
figure 18

Chronological change of gastric cancer patients older than 80 years. The nationwide registry was suspended for a decade from 1992

Cytological examination was conducted in 3,481 (59.4 %) of 5,857 patients with T2, T3, or T4 cancer. The 5YEARS of CY1 patients was 12.3 % and their 5YEARS was as poor as that of patients with peritoneal metastasis. Although this examination was not carried out commonly in the days of 2002, it could still be regarded as a significant and independent prognostic factor from the data that were available. These findings further support the need for staging laparoscopy for accurate preoperative staging in patients with advanced gastric cancer.

JGCA restarted a nationwide registration from 2008. The object of the new nationwide registry was primarily to calculate the stage-specific 5YEARSs among patients who underwent gastrectomy. Therefore, the structure of the database was required to be simple, and the number of registration items was kept to a minimum. Undoubtedly, the next objective would be to collect and analyze data of patients with inoperable disease, remnant gastric cancer, gastrointestinal stromal tumor, malignant lymphoma of the stomach, and other entities that were excluded in the current project. We also began to register patients who were treated by EMR/ESD by adding additional items and updating data entry software from 2011.

We hope that this report will be useful when surveying trends and changes in the clinical practice and treatment results of gastric cancer in Japan. Details of the individual data presented in this report will soon become available for scientific and clinical research with the permission of the registration committee. In addition, most of the surgical and pathological data could easily be transferred to the international database in the near future for various analyses. The registration committee will continue the efforts to improve the registration system, ultimately to collect meaningful annual data.