Ultrasound versus computed tomography for the detection of ureteral calculi in the pediatric population: a clinical effectiveness study.

OBJECTIVE

To determine the diagnostic performance of ultrasound for diagnosing ureteral calculi in children using a clinical effectiveness approach.

METHODS

Billing records and imaging reports were used to identify children (≤ 18 years old) evaluated for suspected urolithiasis using ultrasound between March 2012 and March 2017. Patients without unenhanced CT within 24 h (reference standard) were excluded. Imaging (ultrasound and CT) reports were reviewed for presence, number, size, and location of calculi. Diagnostic performance of ultrasound (versus CT) was calculated on an individual ureter basis both by direct calculus visualization as well as direct visualization combined with suspected presence of ureteral stone based on indirect ultrasound findings.

RESULTS

41 ureteral calculi were present in 38 of 69 (55.1%) patients. Mean patient age was 14.7 ± 3.6 years, and 35 of 69 (51%) patients were boys. Based on direct calculus visualization, ultrasound had a sensitivity of 12.8% (95% CI 5.6-26.7%), specificity of 100% (95% CI 96.3-100%), positive predictive value (PPV) of 100% (95% CI 56.6-100%), and negative predictive value (NPV) of 74.4% (95% CI 66.4-81.1%). When ultrasound examinations reported as suspicious for ureteral calculi based on indirect findings also were considered positive, ultrasound had a sensitivity of 41.0% (95% CI 27.1-56.6%), specificity of 95.0% (95% CI 88.7-97.8%), PPV of 76.2% (95% CI 54.9-89.4%), and NPV of 80.3% (95% CI 72.2-86.5%).

CONCLUSIONS

In clinical practice, ultrasound has low sensitivity for directly visualizing ureteral calculi subsequently identified by CT, although sensitivity improves when considering suspicious examinations as positive.