Creating diagnostic criteria for perforated appendicitis using cross-sectional imaging.

Introduction: Preoperative determination of perforated versus acute appendicitis can be difficult. We compared CT and MRI performance in diagnosing perforated appendicitis, and created diagnostic criteria.
Methods: We retrospectively reviewed all pediatric patients who underwent appendectomy within one day of CT or MRI between 1/1/2013 and 1/16/2016. True diagnosis was determined by pathology report. Findings on CT/MRI were grouped into "hard" findings (abscess, pneumoperitoneum, extruded fecalith, appendiceal wall with visible hole) and "soft" findings (extensive/diffuse inflammation/free fluid, phlegmon). Correlation of white blood cell count (WBC), temperature, peritoneal signs, and symptom duration >72 h with perforation was assessed using logistic regression. Significant correlates were incorporated in clinical criteria.
Results: 135 patients underwent appendectomy after CT/MRI. Fifty patients underwent MRI and 85 CT. Using hard and/or soft findings, MRI was 86.7% sensitive and 74.3% specific, compared to 68.4% (p = 0.19) and 92.4% (p = 0.025) for CT. WBC > 15, temperature >38.0 °C, and peritoneal signs predicted perforation. Diagnostic accuracy of MRI was highest using imaging findings alone. Accuracy of CT was improved by mandating at least one of the previous three clinical correlates, resulting in 68.4% sensitivity and 93.9% specificity.
Conclusions: MRI trended toward more sensitive and CT was more specific for complicated appendicitis. CT specificity is improved by our algorithm.