Jejunal and ileal gastrointestinal stromal tumors: differences in clinicopathologic and computed tomography features.

PURPOSE: To summarize and compare the clinicopathologic and computed tomography (CT) features of jejunal and ileal gastrointestinal stromal tumors (GISTs).

METHODS

A total of 108 patients with pathologically confirmed small bowel GISTs were retrospectively enrolled and categorized into jejunal and ileal groups according to anatomic sites. Two experienced radiologists independently reviewed CT images, with discrepancies resolved by consensus. Clinicopathologic and CT features were compared between the two groups. Multivariate logistic regression analysis was used to identify independent predictors for differentiating jejunal and ileal GISTs, and receiver operating characteristic (ROC) curve analysis was performed to evaluate predictive performance.

RESULTS

Among 108 GISTs, 63 were jejunal and 45 were ileal. Ileal GISTs more frequently presented with abdominal discomfort, high-risk stratification, increased mitotic count, and larger tumor size (all adjusted p < 0.05), whereas jejunal GISTs were more prone to present with gastrointestinal bleeding (adjusted p = 0.014). Regarding CT features, jejunal GISTs tended to exhibit mixed growth pattern, ulceration, tumor vessels, early venous return, larger draining vein diameter, and higher attenuation in the arterial phase (AP) and portal venous phase (all adjusted p < 0.05). In contrast, ileal GISTs were more likely to show exophytic growth pattern and hemorrhage (all adjusted p < 0.05). Multivariate analysis identified tumor size (p = 0.031), diameter of draining vein (p = 0.023), and AP (p = 0.008) as independent predictors. ROC curve analysis showed that AP achieved the best predictive performance (area under the curve [AUC] = 0.912), followed by tumor size (AUC = 0.710) and diameter of draining vein (AUC = 0.640).

CONCLUSION

Jejunal GISTs exhibit distinct clinicopathologic and CT features compared with ileal GISTs. These findings may aid in clinical differentiation and potentially improve diagnostic accuracy.