Study objectiveThe study objective was to apply a previously created composite score for the prediction of adnexal torsion in children and adolescents to a larger group of heterogeneous patients to test its validity and to refine this scoring system to more accurately predict adnexal torsion. MethodsThis was a retrospective chart review at a tertiary care children's hospital and its affiliates. Participants were female patients 0–20 years of age who underwent surgery for suspected adnexal torsion from 2016 to 2019. Data were collected from outpatient, emergency department, and inpatient notes, radiographic data, and operative reports. The primary outcome was the accuracy and predictive value of the total composite score (TCS) to identify or exclude adnexal torsion. ResultsOf the 291 patients with suspected adnexal torsion who went to the operating room during the study period, 168 (57.8 %) had confirmed adnexal torsion. The accuracy of the TCS, which included presence of vomiting, adnexal volume, and adnexal ratio, in predicting or excluding torsion was 83.6 % for all patients, 92.1 % for premenarchal patients, and 81.3 % for menarchal patients. A variation of the TCS that included only vomiting and adnexal volume had an accuracy of 85.8 %. ConclusionsOur study demonstrates the accuracy and reproducibility of a previously published composite score to predict adnexal torsion in children and adolescents. It also further refines this score to a potentially more clinically useful tool. Future studies are needed to prospectively evaluate these composite scores and their implementation in clinical settings. Level of EvidenceIII.