EP23.06: External validation of the ovarian‐adnexal reporting and data system lexicon in a pediatric population

To evaluate the diagnostic performance of the ovarian-adnexal reporting and data system (O-RADS) lexicon in a pediatric population. Retrospective external diagnostic validation study in a single tertiary referral centre for pediatric gynecology in Australia. Experienced examiners retrospectively reviewed all images and assigned classification based on the ORADS Lexicon. All images reviewed had been performed transabdominally. Total of 313 patients diagnosed with an adnexal mass who underwent surgical management at our centre between 1982 - 2018 were included in this study. Age range of the patients was 0 – 18 years. Of the 313 adnexal masses 165 were excluded as the imaging was not available or quality did satisfy the criteria to apply the ORADS lexicon. The reference standard for all masses was histology. Of 148 masses reviewed 144 were benign, 1 borderline and 3 were malignant. The benign cysts comprised of 46 functional cysts, 53 germ cell, 16 epithelial, 1 sex cord stromal, 18 simple cysts,10 non-classifiable. There was 1 borderline epithelial. The malignant tumours comprised of 2 germ cell and 1 sex cord stromal. The relatively high rates of coexistence torsion and adnexal pathology in this pediatric population provided an equivocal ultrasound picture resulting in an up grading to higher ORADS classification. 149 masses received an ORADS classification. The malignancy rates for ORADS 0 were 0% (0 of 2) ORADS 1 were 0% (0 of 6), ORADS 2 were 0% (0 of 78), ORADS 3 3.6% (1 of 28), ORADS 4 0% (0 of 27) and ORADS 5 42.9% (3 of 7). The area under the receiver operator curve was 0.91 95%CI 0.85-0.95. The optimal cut-off value for predicting ovarian malignancy was ORADS > 3 with a sensitivity of 100%, specificity of 61.1%, and negative predictive value of 100%. The ORADS lexicon performed well as a diagnostic test utilised in transabdominal imaging and in predicting malignancy risk in a pediatric population.