Purpose or Case Report: Rapid, non-contrast MRI has been reported as accurate for diagnosing acute appendicitis at specialized pediatric hospitals, but there remains a dearth of evidence-based, MRI-specific criteria upon which the diagnosis should be established. To our knowledge, our study comprises the largest study with the most independent readers of MRI-specific diagnostic criteria for pediatric appendicitis reported to date.
Methods & Materials: Our human-factors experimental design included 210 studies from children (4-18 years) with suspected acute appendicitis who underwent rapid, non-contrast MRI following equivocal US exams from January 2014 to December 2017. A positive diagnosis of acute appendicitis was defined as resolution of acute abdominal pain after surgical or medical appendicitis-specific intervention. All exams were graded independently and blinded to clinical data by four fellowship-trained pediatric radiologists with experience ranging from 5-10 years. Complete experimental block design was used to determine the performance characteristics of individual diagnostic criteria. Data were modeled using generalized mixed modeling with SAS/GLIMMIX. The study was powered equally for sensitivity and specificity.
Results: MRI had a sensitivity of 0.93 and specificity of 0.98. The mean appendix diameter of positive and negative cases was 10.45mm (CI 9.97-10.95) and 5.78mm (CI 5.59-5.99), respectively. The mean appendix wall thickness of positive and negative cases was 2.57mm (CI 2.44-2.71) and 1.69mm (CI 1.54-1.85), respectively. Appendix intraluminal fluid-fluid level had a sensitivity of 0.26 and specificity of 1.00. Peri-appendiceal fluid had a sensitivity of 0.65 and specificity of 0.89. Peri-appendiceal fatty edema had a sensitivity of 0.92 and specificity of 0.97. Presence of appendix wall hyperintensity compared to small bowel wall intensity had a sensitivity of 0.37 and specificity of 1.00.
Conclusions: Appendix diameter wall thickness is a strong quantitative predictor of acute appendicitis with MRI. Presence of peri-appendiceal fatty edema on MRI is also an accurate predictor of acute appendicitis. Other categorical diagnostic criteria including presence of appendix intraluminal fluid-fluid level, peri-appendiceal fluid, and appendiceal fatty edema have high specificity but relatively low sensitivity with MRI. Overall, this study provides valuable information regarding the MRI-specific accuracy of various imaging features associated with pediatric appendicitis.