The Diagnostic Accuracy of Magnetic Resonance Imaging for Maternal Acute Adnexal Torsion during Pregnancy: Single-Institution Clinical Performance Review

Background: For acute adnexal torsion of pregnant women, appropriate treatment based on an accurate diagnosis is especially important for fertility preservation and timely treatment. The 2017 American College of Obstetricians and Gynecologists (ACOG) Committee Opinion No. 723 announced its practice-changing guidelines to ensure that diagnostic magnetic resonance imaging (MRI) conducted during the first trimester and gadolinium exposure at any time during pregnancy are safe for fetal stability. Unfortunately, few studies have been performed to evaluate the usefulness of the diagnostic accuracy of MRI for acute adnexal torsion during pregnancy. Objective: We sought to determine the efficacy of diagnostic MRI modality using multiparameter for maternal adnexal torsion during pregnancy. Methods: From 1 January 2007 to 31 January 2019, 131 pregnant with MRI tests were reviewed. In this retrospective cohort study, 94 women were excluded due to conditions other than an adnexal mass, and 37 were identified through MRI analyses conducted before surgery for suspected adnexal torsion. The primary outcome was the diagnostic accuracy of sonography and MRI, and the secondary outcome was the usefulness of Apparent diffusion coefficient (ADC) values for predicting the severity of hemorrhagic infarction between the medulla and cortex of the torsed ovarian parenchyma. Results: Our study demonstrates that in the diagnosis of adnexal torsion during pregnancy, the sensitivity, specificity, positive predictive value, and negative predictive value are 62.5%, 83.3%, 90.9%, and 45.5% for sonography and 100%, 77.8%, 90.5%, and 100% for MRI. MRI results in surgical-proven adnexal torsion patients revealed unilocular ovarian cysts (36.8% (7/19)), multilocular ovarian cysts (31.6% (6/19)), and near normal-appearing ovaries (31.6% (6/19)). Pathology in adnexal torsion revealed a corpus luteal ovarian cyst (63.2% (12/19)) and underlying adnexal pathology (46.8% (7/19)). Maternal adnexal torsion during pregnancy was more likely to occur in corpus luteal ovarian cysts than in underlying adnexal masses (odds ratio, 2.14; 95% confidence interval (CI), 0.428–10.738). MRI features for adnexal torsion were as follows: tubal wall thickness, 100% (19/19); ovarian stromal (medullary) edema, 100% (19/19); symmetrical or asymmetrical ovarian cystic wall, 100%(19/19); prominent follicles in the ovarian parenchyma periphery, 57.9% (11/19); periadenxal fat stranding, 84.2% (16/19); uterine deviation to the twisted side, 21.1% (4/19); and peritoneal fluid, 42.1% (8/19). The signal intensity of the ADC values of the ovarian medulla and cortex were compared between the cystectomy and detorsion (CD) and salpingo-oophorectomy (SO) groups. The ADC values of the CD and SO groups were 1.81 ± 0.09 × 10(−3) mm(2)/s and 1.91 ± 0.18 × 10(−3) mm(2)/s, respectively (P = 0.209), in the ovarian medulla and 1.37 ± 0.32 × 10(−3) mm(2)/s and 0.96 ± 0.36 × 10(−3) mm(2)/s, respectively (P = 0.022), in the ovarian cortex. The optimal cut-off value of ADC values for predictable total necrosis in the torsed ovarian cortex was ≤ 1.31 × 10(−3) mm(2)/s (area under the curve (AUC) = 0.81; 95% CI 0.611–1.0; P = 0.028). Conclusion: Our data showed that maternal adnexal torsion during pregnancy occurred in most corpus luteal cystic ovary cases and some normal-appearing ovary during the 1st and 2nd trimesters of gestation. Therefore, this study is the first study to elaborate on the existence or usefulness of the diagnostic MRI for acute maternal adnexal torsion during pregnancy and to provide a predictive diagnosis of the severity of hemorrhagic infarction for deciding surgical radicality.