Determination of stone composition by noncontrast spiral computed tomography in the clinical setting

Several investigators have evaluated noncontrast computed tomography (NCCT) in predicting stone composition in vitro. We assessed NCCT in predicting stone composition in patients presenting to our emergency room with flank pain and stone disease.

One hundred twenty-nine patients presenting to our university hospital with flank pain underwent renal colic protocol NCCT scans at the request of the emergency room physicians. A General Electric, high-speed advantage CT scanner was used at 120 kV, 200 mA, and 1.4:1 pitch, with collimation varying between 3 and 5 mm. Ninety-nine patients with predominantly (greater than 50%) calcium oxalate or uric acid composition after either stone passage or stone removal were identified. Each scan was analyzed by one of two radiologists, who determined the predominant attenuation for each stone. Stones once passed or retrieved were analyzed by Urocor Laboratories. The attenuation and attenuation/size ratio (peak attenuation/size in millimeters) were compared with the results of the stone analysis.

Eighty-two calculi predominantly composed of calcium oxalate and 17 calculi predominantly composed of uric acid were identified in 99 patients. The calculi ranged in size from 1 to 28 mm. A significant difference (P = 0.017, unpaired t test) was found between the Hounsfield measurement of uric acid calculi (mean 344 +/- 152 HU) and the Hounsfield measurement of calcium oxalate calculi (mean 652 +/- 490 HU). If only the Hounsfield units from stones 4 mm or larger were compared, the data were even more compelling (P = 0.002). However, using an attenuation/size ratio cutoff of greater than 80, the negative predictive value was 99% that a stone would be predominantly calcium oxalate.

Using peak attenuation measurements and the attenuation/size ratio of urinary calculi from NCCT, we were able to differentiate between uric acid and calcium oxalate stones.