We assessed the potential of dual-energy computed tomography (CT) for the differentiation between uric acid (UA)-containing and non-UA-containing urinary stones. Forty urinary stones of 16 different compositions in two sizes (<and>or=5 mm) were examined in an ex vivo model. Thirty stones consisted of pure calcium oxalate (whewellite or wheddellite), calcium phosphate (apatite, brushite, or vaterite), ammonium magnesium phosphate (struvite), UA, ammonium acid urate, ammonium phosphate, sodium hydrogen urate, or cystine, and ten stones were of mixed composition (UA-sodium hydrogen urate, whewellite-urate, wheddellite-urate, whewellite-brushite, or whewellite-brushite-struvite). Scans were performed using dual-source CT in a dual-energy mode with the tubes simultaneously operating at 80 and 140 kV. Two readers analysed the data with respect to stone attenuation at each energy level. The stones were classified as UA- or non-UA-containing using manual attenuation measurements and software analysis results. Sensitivity, specificity, PPV, and NPV were calculated using crystallographic stone analysis as the gold standard. Twenty-six out of 40 stones (65%) contained no UA; 14 stones (35%) contained UA. When compared with UA-containing stones, the differences in attenuation values at 80 and 140 kV were significantly (P<0.001) higher in stones containing no UA. The software automatically mapped 39/40 stones (98%). Only one (2%) 2 mm UA-stone was missed. The software correctly classified all detected stones as UA- or non-UA-containing. The attenuation values of the missed stone were manually plotted into the analysis sheet which allowed for the correct classification of the stone (containing UA). Therefore, the sensitivity, specificity, PPV, and NPV for the detection of UA-containing stones was 100%. Ex vivo experience indicates that differentiation between UA- and non-UA-containing stones can be accurately performed using dual-source dual-energy CT.
Read full abstract