Feasibility of computed tomography for differentiating calcium oxalate and struvite calculi using mean Hounsfield unit and Hounsfield density

Abstract

Accurate preoperative prediction of cystic calculi composition is essential for effective treatment planning in dogs. This prospective study aims to investigate the utility of non-contrast multi-detector computed tomography (MDCT) in differentiating calcium oxalate (CaOx) and struvite calculi based on mean Hounsfield unit (HU) and HU density values. Nineteen client-owned dogs with confirmed cystic calculi underwent non-contrast MDCT imaging at 80 and 120 kilovoltage peak (kVp) settings prior to surgical removal. Calculi composition was determined via laboratory analysis. Mean HU and HU density were measured and compared between calculi types. The results showed CaOx calculi demonstrated significantly higher mean HU (1317.53 ± 84.38) and HU density (191.89 ± 18.11) compared to struvite calculi (959.38 ± 202.92 and 101.27 ± 22.17, respectively) at 120 kVp. The optimal cut-off value for HU density at 120 kVp is greater than 137.295 HU/mm, indicating CaOx (84.6% of sensitivity and 83.3% of specificity). No significant differences were observed between groups at 80 kVp. These findings suggest that both mean HU and HU density at 120 kVp may be used as parameters for distinguishing between CaOx and struvite calculi. In conclusion, this study supports the clinical application of non-contrast MDCT, particularly using mean HU and HU density at higher kVp settings, as a non-invasive modality for predicting calculi composition in dogs.