Ovarian volume ratio is a reliable predictor of ovarian torsion in girls without an adnexal mass

Purpose The aims of this study were to identify ultrasound-based predictors of ovarian torsion in girls without an adnexal mass and establish a set of normal values for ovarian volume ratio (OVR). Methods A retrospective review was performed of all premenarchal patients ≥3 years of age with a normal pelvic ultrasound between January 2016 and January 2019. A comparison group of premenarchal girls presenting between 2011 and 2019 with torsion in the absence of an adnexal mass was utilized. Results Five-hundred and four premenarchal girls underwent pelvic ultrasound evaluation with a normal examination. The mean OVR was 1.6 ± 0.7 (range 1.0–6.5). OVR did not vary with age (r = −0.06) as compared to ovarian width which increased steadily with age (r = 0.53, p < 0.001). OVR was increased in girls with torsion (7.6 vs 1.4, p < 0.0001), and by receiver operating characteristic (ROC) analysis a cutoff value of >2.5 demonstrated the best diagnostic accuracy of any predictive variable (sensitivity 100%, specificity 94%, AUC 0.991, p < 0.001). Conclusions OVR is an excellent predictor of ovarian torsion in premenarchal girls without an adnexal mass. Unlike ovarian width, OVR does not increase with age, and a cutoff OVR > 2.5 demonstrates high sensitivity and specificity for identifying ovarian torsion in this population. Type of study Study of diagnostic test. Level of evidence Level III.