Analysis of Clinical and Ultrasound Determinants of Adnexal Torsion in Children and Adolescents

Abstract Study Objective We aimed to assess the performance of transabdominal ultrasonography (US) for diagnosis of adnexal torsion (AT) in children and adolescents and evaluate its clinical and US determinants. Design, Setting, Participants, Interventions, and Main Outcome Measures We retrospectively (2004-2014) studied data for children referred for acute or subacute pelvic pain who underwent US and included children 3 months to 18 years old who underwent surgical exploration and/or clinical and radiological follow-up for at least 3 months. The evaluation of US diagnostic performance was on the basis of the proposed diagnosis after US: AT or no AT with or without mass. Clinical and US predictors of AT were identified using regression analysis. Results Among 65 girls included (mean age 11.75 ± 4.49 years), 33 (50.8%) had AT. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of US for AT were 90.9%, 68.7%, 75%, 88%, and 80%, respectively. Overall, 20 of 33 (60.6%) AT patients had a mass, including 33% with mature teratomas. Patients with AT versus no AT more frequently showed vomiting (64% vs 28%; P  = .008), ovarian edematous stroma with a peripheral distribution of follicles (45.5% vs 9.4%; P  = .002) and a mass of pluritissular aspect (40% vs 0%; P P  = .0002) and surface area 18.5 cm 2 or greater ( P  = .0003). Conclusion US has high sensitivity but low specificity for the diagnosis of AT in children. The predictive criteria are vomiting, presence of a pluritissular mass, 4.9 area ratio or greater, or surface area 18.5 cm 2 or greater.