The aim To assess the diagnostic performance of dual-energy computed tomography (DECT) in the evaluation of the composition of urinary stones "in vivo". Materials and methods A total of 91 patients aged from 20 to 70 years old (mean 42.7) with urinary stone disease were examined at Sechenov University, including 68 men (75%) and 23 women (25%). Prior to surgery, all patients underwent DECT (Canon, Japan) in order to predict the chemical composition of urinary stones in vivo. Extracorporeal shockwave lithotripsy (ESWL), percutaneous nephrolithotomy (PCNL) and ureteroscopy (URS) was performed in 53 (58.2%), 18 (19.7%) and 20 (22.1%) patients, respectively. Postoperatively, all stones or stone fragments (n=91; 100%) were examined using a comprehensive physical and chemical analysis (X-ray phase analysis, electron microscopy, infrared spectroscopy). Results In 6 patients (6.6%) staghorn stones were diagnosed, while in 15 (16.5%), 17 (18.7%), 22 (24.2%) and 31 (34.1%) stones were located in ureteropelvic junction, pelvis and ureter, respectively, including 24 patients with lower ureteric stones (26.4%). Prediction of the stone composition in vivo was carried out on the basis of the one indicator, the dual energy ratio (DER). The threshold values of DER for different types of stones were taken from the literature. All stones were divided into 4 groups according to the DECT results: vevellite stones (n=40, 43.9%), Ca-containing stones without vevellite (n=34, 37.3%), uric acid stones (n=10, 10.9%) and struvite stones (n=7, 7.9%). Thus, when comparing the results of DECT and physical and chemical analysis, in the first group four stones were incorrectly assigned by DECT to the group of Ca-containing stones without vevellite and three stones were incorrectly assigned to the group of struvite stones; in the second group four stones were incorrectly assigned to the group of vevellite stones; in the third group one stone was incorrectly assigned to the group of struvite stones; in the fourth group one stone was incorrectly assigned to the group of vevellite stones and one stone in the group of uric acid stones. In order to increase the diagnostic efficiency of DECT, we performed a comprehensive analysis of five specific DECT indicators (stone density at 135 kV, Z eff of the stone, DER, DEI, DED) using discriminant analysis. Thus, the sensitivity, specificity and overall accuracy of DECT with the use of just one indicator (DER) were 83.3%, 89.8%, 86.8% for vevellite, 88.2%, 92.9%, 91.2% for Ca-containing stones without vevellite, 90%, 98.8%, 97.8% for uric acid stones and 60%, 95.3%, 93.4% for struvite stones, respectively. When using discriminant analysis with five specific DECT indicators, higher values of sensitivity, specificity and overall accuracy were seen: 95.2%, 89.8%, 92.3% for a vevellite, 85,3%, 96,4%, 92,3% for Ca-containing stones without a vevellite and 100%, 100% and 100% for both uric acid and struvite stones, respectively. Conclusions Dual-energy computed tomography is a highly informative method which allows to perform preoperatively the reliable assessment of the chemical composition. DECT in patients with urinary stone disease allows to optimize the treatment strategy and carry out preventive measures on individual basis, taking into account the stone type.