Magnetic resonance imaging features of uterine sarcoma and mimickers.

Uterine myometrial tumors are predominantly benign conditions that affect one-third of women and represent the main indication for hysterectomy. Preoperative imaging is of utmost importance for characterization and for precise mapping of myometrial tumors to best guide therapeutic strategy. New minimally invasive therapeutic strategies including morcellation, myolysis, uterine artery embolization and image-guided radiofrequency or focused ultrasound ablation have been developed for the treatment of uterine leiomyoma. However, preoperative differentiation between atypical leiomyomas and leiomyosarcomas is critical on imaging as uterine sarcoma requires a specific surgical technique to prevent dissemination. A single, rapidly growing uterine tumor, associated with endometrial thickening and ascites, in post-menopausal women is suspicious of uterine endometrial stromal sarcoma and carcinosarcoma. Suggestive magnetic resonance imaging features have been described, but overlap in imaging appearance between uterine leiomyosarcomas and cellular leiomyomas makes it challenging to ascertain the diagnosis. This review aims to illustrate the imaging features of uterine sarcomas and potential mimickers to make the reader more familiar with this serious condition which needs special consideration.

Suspected invasive placenta: evaluation with magnetic resonance imaging.

PURPOSE: To determine the utility of magnetic resonance imaging (MRI) in diagnosing invasive placenta (IP). MATERIALS AND METHODS: MRI findings in 32 women with suspected IP were evaluated independently by four readers. Interobserver agreement was calculated with kappa (κ) statistics. Associations between MRI findings and IP were assessed by univariate and multivariate analyses. Sensitivity, specificity and accuracy of MRI for the diagnosis of IP were estimated. RESULTS: Sixteen women (16/32; 50%) had confirmed IP. Interobserver correlation for the diagnosis of IP was fair (κ = 0.40). Univariate analysis revealed that thinning or focal defect of the uteroplacental interface (P < 0.0001) was the most discriminating MRI variable in the differentiation between normal and IP. Overall sensitivity and specificity of MRI for the diagnosis of IP were 84% [95% CI: 75-94%] and 80% [95% CI: 66-93%], respectively. Thinning or focal defect of the uteroplacental interface was the most accurate finding (88%) in the diagnosis of IP. Multivariate analysis revealed that thinning or focal defect of the uteroplacental interface was the single independent predictor of IP (P = 0.0006; OR = 64.99). CONCLUSION: MR imaging has 84% sensitivity [95% CI: 75-94%] and 80% specificity [95% CI: 66-93%] for the diagnosis of IP. Thinning or focal defect of the uteroplacental interface is the most discriminating independent MR variable in differentiating between normal placenta and IP. KEY POINTS: MR imaging has acceptable degrees of accuracy to diagnose invasive placenta. Focal uteroplacental interface defect is the best finding to diagnose invasive placenta. Focal uteroplacental interface defect is the single independent predictor of invasive placenta.