Analysis of radiolucent gallstones by computed tomography for in vivo estimation of stone components.

Successful oral litholytic and other nonoperative therapies of gallstones require exact determination of the stone components. Since computed tomography (CT) provides highly sensitive measurement of density, we performed a study to evaluate whether CT measurement of stone density allows a prediction of the composition of radiolucent gallstones. Twenty-eight patients presenting with 29 radiolucent gallbladder (n = 17) or common bile duct stones (n = 12) were included. Prior to operative or endoscopic therapy the attenuation values (Hounsfield Units, HU) were assessed in vivo by CT under standardized conditions (Somatom II, 125 KV, 130 mAs). After surgical or endoscopic stone removal the concrements were dehydrated, homogenized and then analysed by infra-red spectroscopy. The previously measured Hounsfield units were not known to the investigator. Eighteen cholesterol and 11 pigment stones could be identified. The attenuation values (Hounsfield units) of cholesterol stones amounting to 28-98 HU (48.7 +/- 4.4 HU) differed significantly (P less than 0.001) from pigment stones (90-120 HU/105.5 +/- 2.8 HU). We conclude that computed tomography provides exact discrimination between cholesterol and non-cholesterol stones in vivo. Since only cholesterol stones can be dissolved by cheno- and ursodeoxycholic acid we recommend measurement of the radiodensity of gallstones by CT prior to any litholytic therapy. Furthermore the prediction of the stone composition facilitates the decision on extracorporeal shock wave lithotripsy and the selection of specific solvents for contact lysis via a nasobiliary probe.