Jian Wang1, Xiaoxuan Zhou2, Fangyi Xu2, Weiqun Ao1, Hongjie Hu2. 1. Department of Radiology, Tongde Hospital of Zhejiang Province, Hangzhou, Zhejiang, China. 2. Department of Radiology, Sir Run Run Shaw Hospital, 12377Zhejiang University, Hangzhou, Zhejiang, China.
Abstract
PURPOSE: To discuss significant computed tomography (CT) findings that differentiate gastric leiomyomas (GLs) from small gastric stromal tumors (GSTs). METHODS: One hundred sixty cases with pathologically proven GLs (n = 50) and GSTs (n = 110) with comprehensive CT images were enrolled in this retrospective study. Computed tomography findings (ie, size, location, contour, growth pattern, enhancement degree, necrosis, ulceration, calcification, and lymph nodes) were analyzed through the χ2 or Fisher exact test, independent T test, and multivariate (logistic regression) analysis. Sensitivity and specificity were also calculated. RESULTS: Features of cardia location, endophytic growth, homogeneous gradual enhancement, absent of necrosis, long diameter less than 24 mm, short diameter less than 20 mm, unenhanced CT value larger than 35.2 Hounsfield units (HU), portal venous phase CT value larger than 67.4 HU, and enhancement degree of arterial and venous phase less than 16.2 HU and 32.4 HU were found to be statistically significant between GLs and small GSTs (P < .05). On multivariate analysis, cardia location, endophytic growth, and homogeneous gradual enhancement were independent predictive factors for GLs and small GSTs. CONCLUSION: These 10 CT criteria are very helpful to differentiate GLs from small GSTs. Especially cardia location, endophytic growth, and homogeneous gradual enhancement are of high value in differential diagnosis.
PURPOSE: To discuss significant computed tomography (CT) findings that differentiate gastric leiomyomas (GLs) from small gastric stromal tumors (GSTs). METHODS: One hundred sixty cases with pathologically proven GLs (n = 50) and GSTs (n = 110) with comprehensive CT images were enrolled in this retrospective study. Computed tomography findings (ie, size, location, contour, growth pattern, enhancement degree, necrosis, ulceration, calcification, and lymph nodes) were analyzed through the χ2 or Fisher exact test, independent T test, and multivariate (logistic regression) analysis. Sensitivity and specificity were also calculated. RESULTS: Features of cardia location, endophytic growth, homogeneous gradual enhancement, absent of necrosis, long diameter less than 24 mm, short diameter less than 20 mm, unenhanced CT value larger than 35.2 Hounsfield units (HU), portal venous phase CT value larger than 67.4 HU, and enhancement degree of arterial and venous phase less than 16.2 HU and 32.4 HU were found to be statistically significant between GLs and small GSTs (P < .05). On multivariate analysis, cardia location, endophytic growth, and homogeneous gradual enhancement were independent predictive factors for GLs and small GSTs. CONCLUSION: These 10 CT criteria are very helpful to differentiate GLs from small GSTs. Especially cardia location, endophytic growth, and homogeneous gradual enhancement are of high value in differential diagnosis.