Diagnostic Accuracy of Hydronephrosis on Ultrasound and STONE Score in Predicting Urolithiasis in the Emergency Department

This was a pre-planned secondary analysis of a prospective, observation study of adult Emergency Department (ED) patients suspected of urolithiasis, who presented with unilateral flank pain of less than one week duration. A power analysis with a target Se of 0.70, Sp of 0.80, prevalence of 0.5 and alpha of 0.05, yielded a minimum sample size of 284. The gold standard was ureteral stone confirmation by CT during the same visit. All patients received POCUS and CT imaging. The primary objective was to quantify the diagnostic accuracies of STONE (Sex, Timing, Origin, Nausea/Vomiting, Erythrocytes) score and hydronephrosis. The secondary objective was to determine the 30-day adverse outcomes using an ultrasound-first approach.

Non-contrast or contrast-enhanced computed tomography (CT) is frequently used on patients presenting with flank pain, owing to its high sensitivity (Se) and specificity (Sp) to detect kidney stones. However, repeated imaging raises concerns for cumulative radiation exposure. This has increased interest in using point-of-care ultrasound (POCUS) and clinical decision rules to help safely reduce unnecessary radiation exposure while maintaining diagnostic accuracy.

Of the enrolled 295 patients, positive hydronephrosis demonstrated a Se of 71.2 % (95% CI, 63.6-77.7), Sp of 84.2% (95% CI, 77.2-89.3), positive predictive value (PPV) of 83.5% (95% CI, 76.2-88.8), negative predictive value (NPV) of 72.2% (95% CI, 64.9-78.5), positive likelihood ratio (LR+) of 4.5 (95% CI, 3.0-6.7) and negative likelihood ratio (LR-) of 0.34 (95% CI, 0.26-0.45). The resultant area-under-the-curve (AUC) was 0.77 (95% CI, 0.73-0.82). On the other hand, a bifurcated STONE score using >=6 as a predictor of urolithiasis yielded an overall Se of 61.5% (95% CI, 53.4-69.2), Sp of 75.5% (95% CI, 67.5-82.4), PPV of 73.9% (95% CI, 65.4-81.2), NPV of 63.6% (95% CI, 55.8-70.9), LR+ of 2.52 (95% CI, 1.81-3.50) and LR- of 0.51 (95% CI, 0.41-0.63). The AUC was 0.72 (95% CI, 0.66-0.78). Both positive (7/156 or 4.5%) and negative (9/139 or 6.5%) urolithiasis cohorts showed statistically comparable 30-day adverse outcomes (p=0.45).

POCUS demonstrated better diagnostic accuracy than a clinical decision rule (STONE score) in predicting urolithiasis. In addition, the concomitant use of CT and POCUS in this study suggested that an ultrasound-first approach to manage unilateral flank pain is safe.