Clinicopathologic Characteristics according to the Type of Recurrence in Curatively-resected Gastric Cancer Patients

Ha, Tae Kyung; Kwon, Sung Joon

Abstract

PURPOSE: Evaluating the clinicopathologic characteristics of recurred gastric cancer is essential for early detection of the recurrence and for better clinical outcome. The aim of this study was to establish the patterns of the recurrence, the timing of the recurrence, and the clinical result after recurrence according to the clinicopathologic characteristics of gastric cancer.
MATERIALS AND METHODS
From June 1992 to December 2005, of the 1338 gastric cancer patients who underwent a curative gastric resection, 241 patients who recurred during the follow-up period were selected and their cases were analyzed. The clinicopathologic characteristics of the patients, the time to recurrence after operation and survival were determined retrospectively according to the type of recurrence.
RESULTS
For the recurrent group, the numbers of total gastrectomies, advanced stages, lymphatic and/or venous infiltrations, whole stomach cancer cases, large tumors, undifferentiated tumors, and Borrmann type 4 tumors were higher than they were for the non-recurrent group, and the differences were statistically significant. When the recurrence types were classified as peritoneal seeding, hematogenous recurrence, and locoregional recurrence, independent risk factors were female gender, stage III, upper third, and whlole stomach cancer, and undifferentiated-type, diffuse-type, and Borrmann type 4 tumors for peritoneal seeding; early gastric cancer, stage I for hematogenous recurrence; and Borrmann type 1, 2, and 3 tumors for locoregional recurrence. Survival duration after detection of the recurrence was shorter for peritoneal seeding than for hematogenous or locoregional recurrence (7.0 months vs. 9.5 months and 12.5 months).
CONCLUSION
For early detection of the recurrence after curative surgery for gastric cancer, it is important to recognize that the high risk factors for recurrence vary with the clinicopathologic data for the patients.