영상진단을 이용한 담석의 성분예측에 관한 연구 - 초음파, 전산화단층촬영 및 자기공명영상을 중심으로

Abstract

Background/Aims: To predict the gallstone composition for nonsurgical treatment, various imaging techniques have been used. One such technique that several researchers have recently showed interest in magnetic resonance imaging(MRI). However, the usefulness of this technique has yet to be studied. Thus, the purpose of this study is to compare the effectiveness of ultrasonography(LJS), computed tomography(CT), and MRI for the prediction of gallstone composition. Methods: Thirty- nine cases of gallstones removed surgically were used. The findings of each imaging studies in vitro were compared with the type of stone classified according to the cholesterol content measured by infrared spectrophotometry. Results: Infrared spectrophotometry showed that seventeen cases were cholesterol, 8 mixed, and 14 pigment stones. When US findings were classified into shooting star, crescent, half-moon, full-moon and star-dust pattern, a single pattern was noted in l9 cases, and two or three patterns were noted in the remaining 20 cases. There was no correlation between the US patterns and gallstone types. CT findings were classified into dense(7 cases), faint(4 cases), hypodense(4 cases), rimmed(lO cases) and laminated(4 cases) patterns. Higher incidence of cholestero1 stones was found in hypodense(4/4) and rimmed(6/lO) patterns, while the occurrence of pigment stones was higher in the dense(l 1/l7) pattern. In MRI, while Tl-weighted images showed visibly incrcased signal in 34 of the 39 cases, 17 cases were signal void on T2-weighted images. There was no correlation between signa1 intensity and cholesterol or calcium content. MRI findings were categorized into dark(5 cases), faint(l case), bright(2 cases), rimmed(7 cases), laminated(13 cases) and central sunburst(l3 cases) patterns. There was higher incidence of Cholesterol stones in the central sunburst pattern(lO/l l), while pigment stones occurred more in rimmed(5/7), laminated(6/l3), and bright(2/2) patterns. The sensitivity, specificity, positive and negative predictive value for the prediction of cholesterol stone were 58.8%, 8l.8%, 71.4% and 72.0% by CT, and 64.7/o, 95.5%, 9I.7% and 77.8% by MRI, respectively. Conclusions: US has no value in predicting the composition of gallstones, while CT and MRI seem to have some predictive power. However, MRI has higher specificity and negative predictive value than CT, and therefore MRI is expected to have more clinical usefulness.