Determination of the chemical composition of urinary calculi by noncontrast spiral computerized tomography

Abstract

Various techniques for noncontrast spiral computerized tomography (NCCT) were utilized for the determination of the Hounsefield unit (HU) values of various types of urinary calculi with the aim of determining the best technique for distinguishing various stones compositions. A total of 130 urinary stones, obtained from patients who underwent open surgery, were scanned with a multidetector row scanner using 1.25 mm collimation at two energy levels of 100 and 120 kV at 240 mA. Two post-scanning protocols were used for the HU value assignment, tissue and bone windows, for both kV values. In both protocols, three transverse planes were defined in each stone, one near the top, one in the middle, and one near the bottom. Three regions of interest (ROI) were obtained in each plane. The absolute HU value was determined by three methods: the mean of the nine ROI, the mean of the central three ROI, and the central ROI in the middle plane. Determination of the stones' composition was performed using the absolute HU value measured at 120 kV, the dual CT values (HU values at 100 kV-HU values at 120 kV), and HU values/stone volume ratio (HU density). All stones were analyzed by x-ray diffraction to determine their chemical composition. After the exclusion of groups with few calculi, 47 pure stones [25 uric acid (UA), 15 calcium oxalate monohydrate (COM), seven struvite], and 60 mixed stones [15 COM 60-90%+hydroxyl apatite (HA), 14 COM 40-90%+UA, 21 UA+COM <40%, ten mixed struvite+COM+hydroxyl apatite] were included in the statistical analysis. From the least to the most dense, the pure stone types were UA, struvite, COM. Mixed UA+COM<40% calculi were less dense but insignificantly different from pure UA, while when the COM ratio was > or =40% their density became higher than and significantly different from pure UA, and less than but not significantly differentiated from pure COM. Mixed COM+HA were the most dense stones. Using the absolute HU values at 120 kV and HU density, we could distinguish, with statistical significance, all pure types from each other, pure UA from all mixed calculi except UA+COM <40%, pure COM from mixed UA+COM <40%, and pure struvite from all mixed stones except mixed struvite stones. Dual CT values were not as good as absolute HU values and HU density in the determination of stone composition. These results demonstrate that absolute HU values and HU density derived from CT scanning using a small collimation size could uncover statistically significant differences among all pure and most of the mixed urinary stones. This permits more accuracy in the prediction of stone composition. Moreover, this technique permits diagnostic conclusions on the basis of single CT evaluation.