Unsedated non-contrast MRI for suspected pediatric appendicitis: A prospective study of diagnostic performance, interobserver agreement, and clinical impact.

PURPOSE: To evaluate the diagnostic accuracy, interobserver agreement, and clinical feasibility of an ultra-rapid, unsedated, non-contrast magnetic resonance imaging (MRI) protocol for suspected appendicitis in children, emphasizing negative appendectomy rate and alternative diagnosis detection.

METHODS

This prospective STARD-2015-compliant study included 73 children (11.4 ± 3.2 years; 57.5% male) with suspected appendicitis. All underwent non-contrast MRI without sedation on a 1.5 Tesla scanner. The protocol included T2-weighted turbo spin-echo sequences and diffusion-weighted imaging (b-values 150, 800, 1000 s/mm2) with apparent diffusion coefficient maps. Mean acquisition time was 9.2 ± 1.8 min. Reference standard was histopathology or clinical follow-up. Interobserver agreement was assessed using Cohen's kappa (κ).

RESULTS

Appendicitis prevalence was 32.9% (24/73). MRI showed 95.8% sensitivity (95% CI: 79.8-99.3%), 95.9% specificity (95% CI: 86.3-98.9%), and area under the curve of 0.959. Positive and negative predictive values were 92.0% and 97.9%, respectively. Appendix visualization was 100% in appendicitis vs. 57.1% in non-appendicitis (p < 0.001). Fecalith was 100% specific for appendicitis (45.8% vs. 0%, p < 0.001). Restricted diffusion occurred in 66.7% vs. 12.2% (p < 0.001). Interobserver agreement was almost perfect (κ = 0.937; p < 0.001). All examinations were completed without sedation. Negative appendectomy rate was 4.3% (1/23). Alternative diagnoses were identified on MRI in 38.8% (19/49) of patients without appendicitis.

CONCLUSION

These findings suggest that rapid non-contrast MRI is a promising radiation-free, sedation-free option for children aged ≥ 5 years with suspected appendicitis. However, implementation into routine pediatric emergency pathways should await prospective multicenter validation to confirm reproducibility across diverse settings, scanner types, and radiologist experience levels.