T L Tsai1, C S Changchien, T H Hu, C M Hsiaw, K C Hsieh. 1. Division of Gastroenterrology, Department of Internal Medicine, Chang Gung Memorial Hospital, 123, Tao-Pei Road, Niaosung, Kaohsiung, Taiwan, R.O.C.
Abstract
BACKGROUND: Endoscopic ultrasonography (EUS) is a useful modality to show the sonographic features of gastric stromal tumors (GST) which originate from the gastric wall, but it is difficult to distinguish benign from malignant tumors. This study tried to discriminate between them by their EUS patterns. METHODS: EUS patterns were obtained from 52 histologically proved GST. The tumor size, echo patterns and irregular margin were analyzed. RESULTS: Of 52 GST, 11 were leiomyosarcoma and 41 were leiomyoma. The mean sizes of the leiomyoma and leiomyosarcoma were 3.3 +/- 1.1 cm and 7.1 +/- 2.2 cm (p < 0.01). Using a tumor size > or = 5 cm as a criterion for malignancy resulted in a sensitivity of 72.7% and a specificity of 85.4%. A sonolucent area inside the tumor, resulted in a sensitivity and specificity of 81.8% and 80.5%. With an irregular tumor margin, the sensitivity and specificity were 90.9% and 90.2%. A combination of all 3 criteria, resulted in a sensitivity and specificity of 72.7% and 90.2%. For benign tumors, the criteria of tumor < 5 cm, no sonolucent areas and no irregular margin had a sensitivity of 80.5% and a specificity of 100%. CONCLUSION: Tumor size is not a sensitive criterion of malignancy. An irregular margin and sonolucent areas in the tumor have better sensitivity and specificity for the diagnosis of malignant GST. On the contrary, a GST smaller than 5 cm in size with neither a sonolucent area nor an irregular tumor margin strongly suggests a benign lesion.
BACKGROUND: Endoscopic ultrasonography (EUS) is a useful modality to show the sonographic features of gastric stromal tumors (GST) which originate from the gastric wall, but it is difficult to distinguish benign from malignant tumors. This study tried to discriminate between them by their EUS patterns. METHODS: EUS patterns were obtained from 52 histologically proved GST. The tumor size, echo patterns and irregular margin were analyzed. RESULTS: Of 52 GST, 11 were leiomyosarcoma and 41 were leiomyoma. The mean sizes of the leiomyoma and leiomyosarcoma were 3.3 +/- 1.1 cm and 7.1 +/- 2.2 cm (p < 0.01). Using a tumor size > or = 5 cm as a criterion for malignancy resulted in a sensitivity of 72.7% and a specificity of 85.4%. A sonolucent area inside the tumor, resulted in a sensitivity and specificity of 81.8% and 80.5%. With an irregular tumor margin, the sensitivity and specificity were 90.9% and 90.2%. A combination of all 3 criteria, resulted in a sensitivity and specificity of 72.7% and 90.2%. For benign tumors, the criteria of tumor < 5 cm, no sonolucent areas and no irregular margin had a sensitivity of 80.5% and a specificity of 100%. CONCLUSION:Tumor size is not a sensitive criterion of malignancy. An irregular margin and sonolucent areas in the tumor have better sensitivity and specificity for the diagnosis of malignant GST. On the contrary, a GST smaller than 5 cm in size with neither a sonolucent area nor an irregular tumor margin strongly suggests a benign lesion.
Authors: R Matthew Walsh; Jeffrey Ponsky; Fred Brody; Brent D Matthews; B Todd Heniford Journal: J Gastrointest Surg Date: 2003 Mar-Apr Impact factor: 3.452