The aim of the study was to assess the value of quantitative attenuation values (Hounsfield units) and of gallstone pattern by computerized tomography in predicting response to bile acid therapy. We carried out a prospective study in a multicenter setting on 90 consecutive outpatients with radiolucent gallstones. All received bile acid therapy (UDCA 10 mg/kg/day or UDCA+CDCA 5 mg/kg/day of each) up to two years. Hounsfield units for gallstones were recorded using standardized criteria and six categories of patterns were defined: hypodense, isodense, homogenously dense, laminated, rimmed and speckled. We assessed gallstone dissolution rate (percent reduction in volume), response to therapy (>25% reduction in volume), and final outcome of therapy. Eighty-one percent of patients with hypodense/isodense and all four patients with speckled stone pattern responded to therapy, whereas none of the 10 patients with laminated/rimmed and only 45% of patients with homogenously dense stone pattern did. Complete dissolution was achieved by 68%, 50%, 35%, 0% of the hypodense/isodense, speckled, homeogenously dense, rimmed/laminated gallstones, respectively. The use of Hounsfield units did not show an advantage over gallstone pattern for predicting either response or final outcome to bile acid therapy. We conclude that computerized tomography analysis of gallstones is of value in predicting response to bile acid therapy and that gallstone pattern alone predicts response in most cases without the need for quantitative assessment.