PURPOSE: To optimize the examination protocols of multislice CT (MSCT) of the pelvis for dose reduction with regard to image quality. MATERIALS AND METHODS: MSCT of the pelvis was performed on 5 cadaver specimens with stepwise reduction of the tube current at 140 kV (250, 200, 150, 100, 75, 50, 35, 25 mAs) and 120 kV (50, 27 mAs). The remaining scan parameters like collimation (4 x 1 mm) and table speed of 4 mm per rotation remained unchanged. Axial sections and coronal reconstructions were used to evaluate cortex, trabecular structures, subjective image quality, image noise and detail detectability (pelvis and SI joint), with evaluation performed independently by four blinded experienced radiologists on a 5-point scale. Kappa coefficient, accuracy of the observers to sort the films with regard to dose reduction and mean scores of image evaluation were determined for statistical analysis. RESULTS: The deterioration of the image quality was most pronounced for all criteria between 75 mAs and 50 mAs at 140 kV. Good results with adequate image quality were obtained for detail detectability at 50 mAs and 140 kV (effective dose [E]: 3.3 mSv) and for the remaining four criteria at 75 mAs (E: 4.9 mSv). There was a moderate agreement between the four observers (kappa coefficient: 0.27). All observers were excellent in arranging the images according to the increasing dose reduction. CONCLUSIONS: Image quality of MSCT of the pelvis appears to be acceptable at 75 mAs and 140 kV with the dose reduced to 46% of the average value of the nation-wide survey of the German Roentgen Society in 1999 for this type of examination.
