In vivo evaluation of the chemical composition of urinary stones using dual-energy CT.

Abstract

The purpose of this article is to evaluate in vivo the chemical composition of urinary stones using dual-source and dual-energy CT, with crystallography as the reference standard. Forty patients (mean [± SD] age, 49 ± 17 years) with known or suspected nephrolithiasis underwent unenhanced abdominal CT for urinary tract evaluation using a dual-energy technique (tube voltages, 140 and 80 kVp). For each stone 5 mm or larger in diameter, we evaluated the site, diameter, CT density, surface (smooth vs rough), and stone composition. Patients were treated with extracorporeal shock wave lithotripsy (n = 34), percutaneous nephrolithotomy (n = 4), or therapeutic ureterorenoscopy (n = 2). Collected stones underwent crystallography, and the agreement with the results of dual-energy CT was calculated with the Cohen kappa coefficient. The correlation among stone composition, diameter, and CT density was estimated using the Kruskal-Wallis test. Thirty-one patients had a single stone and nine had multiple stones, for a total of 49 stones. Forty-five stones were in the kidneys, and four were in the ureters; 23 had a smooth surface and 26 had a rough surface. The mean stone diameter was 12 ± 6 mm; mean CT density was 783 ± 274 HU. According to crystallography, stone composition was as follows: 33 were calcium oxalate, seven were cystine, four were uric acid, and five were of mixed composition. Dual-energy CT failed to identify four stones with mixed composition, resulting in substantial agreement between dual-energy CT and crystallography (Cohen κ = 0.684). Stone composition was not correlated with either stone diameter (p = 0.920) or stone CT density (p = 0.185). CT showed excellent accuracy in classifying urinary stone chemical composition, except for uric acid-hydroxyapatite mixed stones.