Background: We evaluated the influence of several clinicopathologic variables on 5-year survival of patients with gastric cancer associated with esophageal or duodenal invasion, and determined the significance of resection line involvement.
Patients and methods: A review of the database for gastric adenocarcinoma at Sendai National Hospital between January 1985 and December 1995 identified 923 patients who underwent gastric cancer resection. Of these patients, 37 were reported to have tumour infiltration of the esophagus or duodenum on histological examination of the resected specimens. Univariate and multivariate analyses of patients with esophageal or duodenal invasion were performed to evaluate the prognostic significance of clinicopathologic features. Then the patients were divided into two groups based on the results of microscopic examination: a tumour wedge-positive group for resection margins of less than 5 mm in width and a tumour wedge-negative group for resection margins of more than 5 mm in width. There were 8 patients in the narrow (margin-positive) group and 29 patients in the wide margin (margin-negative) group, respectively.
Results: Univariate analysis revealed that the significant prognostic factors were nodal involvement (p=0.0004) and gross type (p=0.0031). Multivariate analysis of the esophagus or duodenum-invaded cancer cases, however, revealed that only nodal involvement was a significant prognostic factor. There were statistical correlations between these groups (margin-positive and margin-negative groups) and the Borrmann type of tumour and tumour size. The survival rate was worse in patients with tumour line involvement.
Conclusions: Multivariate analysis revealed that the prognosis of patients with esophageal or duodenal invasion was affected only by nodal involvement independently. The risk of surgical margin involvement was high in cases of a large Borrmann type-4 tumour and infiltrative carcinoma.