Background. Differing survival rates have been reported between patients having undergone surgical intervention for the treatment of gastric carcinoma in Japan and Western industrialized countries. Through the actual availability of the data compiled at a major Japanese medical center (National Cancer Center, Tokyo), it was possible, for the first time, to compare the patients and therapeutic results of a Japanese center (n = 1475) with that of a German center (Department of Surgery, Technical University of Munich, Munich; n = 453). Methods. The prognostic factors involving both groups were compared. Survival rates were analyzed in univariate and multivariate fashions. Results. Some of the examined prognostic factors, such as sex, histologic type, tumor size, and Borrmann classification, were similarly distributed. Differences in frequency were discovered concerning pathologic tumor (pT), node (pN), and metastasis (pM) categories, localization, and age groups. Univariate analysis showed a 2‐year survival rate of 88% for all Japanese patients with gastric cancer compared with 58% for German patients. The 5‐year survival rates were 77% and 44%, respectively. The difference in the 2‐year and 5‐year survival rates for both departments may be related to differences in frequencies of several characteristics. In performing the same analysis in a multivariate fashion for the patient populations at both centers, it became clear that an important prognostic factor was the center itself. The survival curves of patients from Tokyo and Munich with the same prognostic factors demonstrate this difference. These differences, however, were small in comparison with those of univariate analysis. Conclusions. Using a similar classification of the tumor stage and similar prognostic characteristics, the prognosis for gastric cancer in Japan and Germany may be the same.

The probability of survival of patients with gastric cancer depends upon depth of wall penetration by the primary tumour and metastatic lymph node burden. Radical lymphadenectomy may lead to prolonged survival but with increased morbidity. A computer program from Maruyama, National Cancer Centre, Tokyo enables evaluation of individual survival time and infiltration of lymph nodes. This analysis was applied to a German population. Computer-aided predictions were determined retrospectively using the prognostic factors of sex, age, Borrmann classification, histology, depth of wall penetration, location and diameter of the tumour. Data were collected from 222 patients at the Technical University of Munich (median age 66 years, sex ratio (M:F) 2:1), who had been operated on (72 per cent total gastrectomy, 28 per cent subtotal gastrectomy) for gastric cancer. The predicted results were compared with the postoperative pathological findings. The prediction of node metastases was highly accurate (lymph nodes 13-16, 96 per cent; 7-12, 89 per cent; 1-6, 82 per cent). These computer predictions may provide perioperative information of therapeutic value.

The frequency and prognostic relevance of sarcoid-like lesions and microcarcinosis in regional lymph nodes in gastric cancer (n = 113; pT1-3, pN0-1, pM0, R0) were investigated; the prognostic value was compared with pT and pN stage, grading and Laurén's tumor classification with Cox's multivariate regression-model. Sarcoid like lesions were found in 34% of the cases (n = 113). Statistical analysis did not indicate that they had any prognostic value or showed whether or not microcarcinosis or metastasis was present; they were independent of pT stage, histological tumor type, tumor grading, and the clinical course of disease. Microcarcinosis (defined as scattered carcinoma cells within lymph node sinuses or pulp without adjacent stromal reaction) was revealed by immunohistochemistry in 90% of pN0 cases; the presence of 3 or more tumor cells per lymph node section in over 10% of sampled lymph nodes per case carried a significant prognostic value. In microcarcinosis without evidence of metastasis, the number of tumor cells and the number of involved lymph nodes are of prognostic value. In pN1 cases microcarcinosis was found as well as the metastases in 97% and had no additional prognostic value. Microcarcinosis alone has a different significance from lymph node metastasis for prognosis.