OBJECTIVE To analyze the pathological features and prognosis factors of gastrointestinal stromal tumor (GIST) after primary resection. METHODS Medical records of the diagnosis, surgery, and follow-up of 327 patients with GISTs who underwent surgery between 1988 and 2007 were retrospectively reviewed. The predic-tive factors for the survival of these patients were identi. ed using multivariate analysis. RESULTS In the 327 tumors, 152 (46.5%) were located in the stomach, 89 (27.2%) in the small intestine, 33 (10.1%) in the colon and rectum, and 43 (13.1%) in other sites including the omentum and mesentery. The 3-year and 5-year overall survival rates of the 327 GIST patients were 74.4% and 62.7%, respectively, and univariate survival analysis demonstrated that factors, such as tumor size, mitotic index, NIH categories, Ki-67 index, tumor location, surgical margins, tumor bleeding, and tumor necrosis have significant effect on survival of the patients (P < 0.05). Multivariate analysis demonstrated that the NIH categories, surgical margins, and Ki-67 index were independent prognostic factors for the survival rate. In the group of patients with postoperative recurrence or metastasis, the median survival time of patients who did not receive imatinib treatment was 30 months and that of patients who received imatinib treatment was 59 months. Their 5-year survival rates were 16.4% and 39.4%, respectively, and the difference was statistically significant (P = 0.017). CONCLUSION Complete resection is the .rst choice of treat-ment for GISTs. It is reasonable to evaluate the prognosis of resect-able GISTs and guide the adjunctive therapy with NIH categories and Ki-67 index. Imatinib treatment can signi.cantly increase the survival rate of patients with recurrent and metastatic GISTs.