Abstract
OBJECTIVE: The purpose of our study was to assess computed tomographic (CT) findings of adnexal torsion through a matched case-control analysis. MATERIALS AND METHODS: This retrospective, single-institution case-control study included 43 women with adnexal torsion and 43 age- and ovarian mass-matched control women. CT images were evaluated independently by two readers for the following: prominent peripheral follicles, uterine deviation, thickened pedicles, a whirl sign, and a navel sign. Comparisons of CT findings were performed using the Chi square test and receiver operating characteristic (ROC) curves were obtained to assess the diagnostic performance. Differences between the areas under the ROC curves (AUCs) were compared by using a Delong test. RESULTS: The CT findings significant for adnexal torsion were uterine deviation toward the side of the affected ovary (P = < .01 for reader 1 and P = .02 for reader 2) and thickened pedicles with ancillary findings including a whirl sign, a navel sign, and uterine deviation facing thickened pedicles (P < .01 for both readers). Thickened pedicles with ancillary findings had the highest diagnostic accuracy, as measured with ROC curves (AUC, 0.86 in reader 1 and 0.85 in reader 2). Combining uterine deviation toward the side of the affected ovary with thickened pedicles with ancillary findings did not increase the performance relative to that of thickened pedicles with ancillary findings alone. CONCLUSIONS: Thickened pedicles with ancillary findings including a whirl sign, a navel sign, and uterine deviation facing thickened pedicles could be helpful for the diagnosis of adnexal torsion.