Magnetic resonance imaging and sonography in the diagnosis of placental invasion

Abstract

Objectives-To compare older and newer magnetic resonance imaging (MRI) criteria for placental invasion and to compare the sensitivity, specificity, and accuracy of MRI and sonography in determining the depth of placental invasion. Methods-Forty pregnant patients at high risk for morbidly adherent placenta based on prenatal sonography underwent MRI evaluations. Two reviewers, who were blinded to the original MRI and sonographic interpretations, clinical history, and obstetric/ pathologic findings, reviewed the MRI examinations. The MRI and sonographic scans were analyzed for the presence and depth of invasion. The MRI scans were tabulated for the presence of dark intraplacental T2 bands, bulging of the myometrium, increased vascularity, and indistinct myometrium, loss of the dark T2 myometrial/placental interface, and a thin myometrium. The obstetric/pathologic results served as the reference standards. Results-Eighteen of 40 patients had a morbidly invasive placenta. The sensitivity, specificity, and accuracy of MRI and sonography were not significantly different. The accuracy rates for determining the depth of placental invasion by readers 1 and 2 were 0.65 and 0.55, respectively (P> .05). According to the Cohen κ statistic, there was a good inter-reader agreement between the MRI readers in assessing the depth of placental invasion (κ= 0.45). The features most commonly seen were dark T2 bands, bulging of the uterus, and loss of the dark T2 interface, which were all associated with the presence of placental invasion. Conclusions-The diagnosis of placental invasion remains challenging on sonography and MRI, which perform similarly. The presence of 2 or more criteria adds specificity to the diagnosis of placental invasion on MRI.