Gastric cancer treated in 2002 in Japan: 2009 annual report of the JGCA nationwide registry

Background The Japanese Gastric Cancer Association (JGCA) started a new nationwide gastric cancer registration in 2008. Methods From 208 participating hospitals, 53 items including surgical procedures, pathological diagnosis, and survival outcomes of 13,626 patients with primary gastric cancer treated in 2002 were collected retrospectively. Data were entered into the JGCA database according to the JGCA classification (13th edition) and UICC TNM classification (5th edition) using an electronic data collecting system. Finally, data of 13,002 patients who underwent laparotomy were analyzed. Results The 5-year follow-up rate was 83.3 %. The direct death rate was 0.48 %. UICC 5-year survival rates (5YEARSs)/JGCA 5YEARSs were 92.2 %/92.3 % for stage IA, 85.3 %/84.7 % for stage IB, 72.1 %/70.0 % for stage II, 52.8 %/46.8 % for stage IIIA, 31.0 %/28.8 % for stage IIIB, and 14.9 %/15.3 % for stage IV, respectively. The proportion of patients more than 80 years old was 7.8 %, and their 5YEARS was 51.6 %. Postoperative outcome of the patients with primary gastric carcinoma in Japan have apparently improved in advanced cases and among the aged population when compared with the archival data. Further efforts to improve the follow-up rate are needed. Conclusions Postoperative outcome of the patients with primary gastric carcinoma in Japan have apparently improved in advanced cases and among the aged population when compared with the archival data. Further efforts to improve the follow-up rate are needed.


Introduction
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].

Results
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.
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 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).
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   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 ? a, and D1 ? b 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).

Discussion
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 5YE-ARSs in our program. Although, the correlation between follow-up rate and survival rate is complicated, our followup system needs to be improved if we are to evaluate the survival rates more accurately.
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.