Predicting the composition of urinary stone by non-enhanced spiral computed tomography

Purpose: We attempted to develop a more accurate set of tools for predicting the composition of urinary stone with using Hounsfield units (HUs) in non-enhanced spiral computed tomography (NESCT). Materials and Methods: We evaluated 141 stones, and each of which was determined to contain the predominant stone component more than 70% (84 calcium oxalate, 35 uric acid, 16 carbonate apatite, 4 cystine and 2 brushite). NESCT was conducted at different collimations that varied between 3.75mm, 5mm and 7mm through the stones. One region of interest (ROI) was acquired for the plane that passed through the widest transverse diameter of the stone at the bone window setting. The mean size of the ROI in which the highest attenuation area was located was determined to be 2.0±0.5mm², and we recorded the mean and maximum HU values. In order to assess the partial volume effects, we used the S/C ratio (stone size/collimation ratio). Results: As the S/C ratio increased, both mean and maximum HU values increased, as did the ability to differentiate between different stone compositions. Also, under conditions in which the S/C ratio exceeded 2, we proved to be able to differentiate uric acid stones from other stones, with no overlap in attenuation. Maximum HU values also proved to be effective tools for determining stone composition, as compared with the mean HU values. Conclusions: In situations in which the S/C ratio exceeded 2, CT HUs proved to be a useful and accurate measurement for predicting uric acid stones.