Background. Despite complete resection of gastrointestinal stromal tumors (GIST), recurrent and/or metastatic disease occurs, often depending on the grade of malignancy. As such, markers are needed that accurately predict patients at high risk for recurrence. Previously our group reported Pfetin as a prognostic biomarker for GIST. In order to create an approach for predicting risk of recurrence, we incorporated Pfetin expression with clinicopathological data to produce a predictive model. Object. Forty-five patients with localized primary GIST were treated with complete gross surgical resection surgically at our institution between 1995 and 2010 were included. The majority of tumors originated in the stomach (38 cases), as well as small intestine (6 cases) and rectum (1 case). Method. (1) We performed retrospective analysis of the connection between Pfetin expression, clinicopathological data, and incidences of recurrence, using bivariate and multivariate analyses. (2) The reactivity of the monoclonal antibody against Pfetin was examined by immunohistochemistry. Pfetin. We have reported Pfetin, identified microarray technology, and compared between statistically different GISTs for good and poor prognoses and for prognostic marker. Results. There were 7 cases of recurrences. (1) By univariate analysis, tumor size, mitoses, exposure to abdominal cavity, and complete tumor removal predicted risk of recurrence. (2) Pfetin-negative cases were significantly related to recurrence (P = 0.002). Conclusions. This analysis demonstrates that lack of Pfetin expression is an additional predictor of recurrence in resected GIST. Further study may determine the role of this variable added to the current predictive model for selection of adjuvant therapy.