ABSTRACT
OBJECTIVE: To determine factors predicting if a radiologists... report of a .. stone... on ultrasound (US) was not actually a clinically significant stone, based on subsequent computed tomogram (CT). US often overestimates stone size and various pathologic entities are also hyperechoic;.ßthus, a subsequent CT without a clinically significant stone may represent unnecessary radiation exposure. A decision-tree and nomogram were developed to predict when stones are unlikely on subsequent CT. METHODS: Retrospective analysis of patients, of any age, receiving CT within 24.ßhours of a sonographic report documenting a single renal stone, during 2019...2020, in any phase of care, at one institution. Novel stone-likelihood-systems for US and CT (US-SLS, CT-SLS) were devised and validated to classify stones as clinically significant or insignificant, with CT as the gold standard. Binomial logistic regression predicting clinically significant stones was performed with sonographic and patient characteristics. RESULTS: Eight hundred twenty patients had US followed by CT, 228 (27.8%) reported documented stones, 140 (17.1%) reported a single stone. Clinically significant stones were associated with larger stone size (P: .002), location (P: .002), hydronephrosis (P: .04), shadowing-artifact (P: .02) depth.ßto.ßstone (P: .008), and Body mass Index (BMI) (P: .01). US-SLS had higher sensitivity (95.4%) and negative-predictive-value (81.8%) compared to a multivariate model of significant variables. CONCLUSION: US-SLS appears to exclude clinically irrelevant .. stones... better than established criteria including twinkle or shadow in this retrospective analysis. A diagnostic algorithm and nomogram are presented. US-SLS and the associated decision tree can assist providers in avoiding unnecessary radiation when clinically significant stones are unlikely.
