ObjectivesTo determine CT features that can identify gastrointestinal stromal tumors (GISTs) among gastric sub-epithelial tumors (SETs) and to explore a practical scoring method.MethodsSixty-four patients with gastric SETs (51 GISTs and 13 non-GISTs) from hospital I were included for primary analyses, and 92 (67 GISTs and 25 non-GISTs) from hospital II constituted a validation cohort. Pre-operative CT images were reviewed for imaging features: lesion location, growth pattern, lesion margin, enhancement pattern, dynamic pattern, attenuation at each phasic images and presence of necrosis, superficial ulcer, calcification, and peri-lesion enlarged lymph node (LN). Clinical and CT features were compared between the two groups (GISTs versus non-GISTs) and a GIST-risk scoring method was developed; then, its performance for identifying GISTs was tested in the validation cohort.ResultsSeven clinical and CT features were significantly suggestive of GISTs rather than non-GISTs: older age (>49years), non-cardial location, irregular margin, lower attenuation on unenhanced images (43 HU), heterogeneous enhancement, necrosis, and absence of enlarged LN (p<0.05). At validation step, the established scoring method with cut-off score dichotomized into 4 versus <4 for identifying GISTs revealed an AUC of 0.97 with an accuracy of 92%, a sensitivity of 100% and a negative predictive value (NPV) of 100%.ConclusionsGastric GISTs have special CT and clinical features that differ from non-GISTs. With a simple and practical scoring method based on the significant features, GISTs can be accurately differentiated from non-GISTs.