Prospective comparison of MR imaging and US for the diagnosis of pediatric appendicitis

Abstract

Purpose: To prospectively compare nonenhanced magnetic resonance (MR) imaging and ultrasonography (US) for the diagnosis of pediatric appendicitis.

Materials and methods: This HIPAA-compliant study was approved by the institutional review board, and written informed consent was obtained from the patient's parent or guardian. Eighty-one patients (34 male, 47 female; mean age, 12.3 years ± 3.5 [standard deviation]; range, 4-17 years) were enrolled in this prospective study. All patients underwent right lower quadrant US and nonenhanced, nonsedated abdominopelvic MR imaging examinations. Two pediatric radiologists blinded to US results independently reviewed the MR images. MR imaging and US findings were designated positive, negative, or equivocal for acute appendicitis. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for MR imaging and US and compared by using two-sided McNemar test or the score statistics specified by Leisenring. Kappa statistics were generated to determine intertechnique agreement between MR imaging and US and interobserver agreement between the two primary MR imaging readers.

Results: Thirty (37%) patients had pathologically proved acute appendicitis. When equivocal interpretations were designated positive, sensitivity was 93.3% for MR imaging (95% confidence interval [CI]: 77.9%, 99.2%) and 90.0% for US (95% CI: 73.5%, 97.9%), P > .99; specificity was 98% for MR imaging (95% CI: 89.6%, 100%) and 86.3% for US (95% CI:73.7%, 94.3%), P = .03; PPV was 96.5% for MR imaging (95% CI: 82.2%, 99.9%) and 79.4% for US (95% CI: 62.1%, 91.3%), P = .007; and NPV was 96.2% for MR imaging (95% CI: 86.8%, 99.5%) and 93.6% for US (95% CI: 82.4%, 98.7%), P = .45, with substantial intertechnique (κ = 0.77; 95% CI: 0.63, 0.90) and interobserver (κ = 0.76; 95% CI: 0.61, 0.91) agreement. When equivocal interpretations were designated negative, MR imaging sensitivity, specificity, PPV, and NPV were unchanged. For US, sensitivity was 86.7% (95% CI: 69.3%, 96.2%), P = .5; specificity was 100% (95% CI: 93.0%, 100%), P > .99; PPV was 100% (95% CI: 86.8%, 100%), P = .31; and NPV was 92.7% (95% CI: 82.4%, 98.0%), P = .16, with almost perfect intertechnique (κ = 0.92; 95% CI: 0.83, 1.00) and substantial interobserver (κ = 0.72; 95% CI: 0.58, 0.87) agreement.

Conclusion: Nonenhanced MR imaging demonstrates high diagnostic performance similar to that of US for suspected pediatric appendicitis.