LAUSR

PURPOSE The purpose of this study was to assess the performance of magnetic resonance imaging (MRI) in children and adolescents with suspected adnexal torsion (AT) after inconclusive initial ultrasound examination. MATERIALS AND METHODS Twenty-eight girls with a mean age of 12±4 (SD) years (range: 1 month to 18years) were included. All had clinically suspected AT and inconclusive initial ultrasound findings followed by pelvic MRI as a second-line imaging modality. The final diagnosis was obtained by surgery or follow-up. Two radiologists blinded to the clinical, ultrasound and surgical data, retrospectively and independently reviewed MRI examinations. Clinical and MRI features associated with AT were searched for using univariate analyses. RESULT Among the 28 patients, 10/28 patients (36%) had AT and 22/28 (79%) had an ovarian or tubal mass. AT was associated with an age<13years (OR: 10.7; 95% CI: 1.3-148.2) (P=0.022) and a whirlpool sign at MRI (OR: 61.0; median unbiased estimate, 7.2) (P<0.0001). When a mass was present, the best quantitative MRI criteria for AT were mass volume and ovary-corrected volume≥30cm3 (κ=0.72 and 0.61, respectively), mass axis length≥5cm (κ=0.90), and mass surface area≥14 cm2 (κ=0.58), with moderate to almost perfect interobserver agreement. The overall sensitivity, specificity and accuracy of MRI for the diagnosis of AT were 100% (10/10; 95% CI: 69-100), 94% (17/18; 95% CI: 73-100) and 96% (27/28; 95% CI: 82-100) respectively, with perfect interobserver agreement (κ=1). CONCLUSION In pediatric patients with suspected AT and inconclusive initial ultrasound examination, a strategy including MRI as a second-line imaging modality should be considered if MRI does not delay a potential surgery.