Accuracy of Magnetic Resonance Imaging and Ultrasound for Appendicitis in Diagnostic and Nondiagnostic Studies.

OBJECTIVES: Suggestive radiographic studies with nonvisualization of the appendix can present a challenge to clinicians in the evaluation of pediatric abdominal pain. The primary objective of this study was to quantify the accuracy of magnetic resonance imaging (MRI) and of ultrasound (US) in the setting of nonvisualization of the appendix. Secondary objectives reported include sensitivity of MRI and US overall and correlation between MRI and US for diagnosis of appendicitis.

METHODS: Records of pediatric emergency department patients aged 3 to 21 years undergoing MRI and/or US for the evaluation of appendicitis were retrospectively reviewed. Radiographs were categorized as a normal appendix, neither demonstrating the appendix nor demonstrating abnormalities consistent with appendicitis; equivocal, not demonstrating the appendix but showing evidence of appendicitis; demonstrating an abnormal appendix consistent with appendicitis; or demonstrating an alternate pathology. The reading was compared with the final diagnosis for accuracy.

RESULTS: Of the 589 patients included, 146 had appendicitis. Diagnostic accuracy for studies with a nonvisualized appendix without secondary signs of appendicitis was 100% for MRI and 91.4% (95% CI = 87.3% to 94.2%) for US. Diagnostic accuracy for studies with a nonvisualized appendix with secondary signs of appendicitis was 50% (95% CI = 2.5% to 97.5%) for MRI and 38.9% (95% CI = 18.2% to 64.5%) for US. Appendicitis was ultimately diagnosed in 8.6% of patients with an otherwise negative right lower quadrant (RLQ) US that failed to directly identify the appendix. There was a moderate correlation between US and MRI (ρ = 0.573, p = 0.0001) when all studies were considered.

CONCLUSIONS: Magnetic resonance imaging without secondary signs of appendicitis is effective in excluding appendicitis regardless of whether the appendix is directly visualized, while otherwise negative RLQ US that fail to identify the appendix are less useful. Secondary signs of appendicitis without visualization of the appendix were not helpful regardless of radiographic modality. Results of MRI and US correlated moderately well.