PURPOSE: to establish the main signs of placental accretism in magnetic resonance imaging (MRI) in patients with clinical suspicion and to estimate the benefit of this method. METHODS: prospective transversal study with 15 patients suspected of placental accretism, referred between March 2003 and February 2006. Gestational age varied from 20 to 31 weeks. All patients underwent MRI to study the placenta and had previously done an ultrasonography. Material was sent to histological study. MRI was done on Magnetom Impact and Sonata Maestro Class Siemens®, with acquired sequences HASTE, TURBO SPIN in axial, sagittal, coronal planes and echo gradient (GE®), pre- and post-dynamic contrast in the best plan for acquisition. Images were analyzed by a team of two radiologists. RESULTS: mean gestational age was 24.3 weeks. We studied seven placenta previa (47%), six anterior placentas (40%) and two posterior placentas (13%). Ultrasonography was positive in 80% of the palcentas and MRI in 53%. However, echography had a low concordance with anatomic pathological studies by Kappa test (11%), revealing 75% of sensitivity, 14% of specificity, 50% as positive predictive value (PPV) and 33% as negative predictive value (NPV). MRI had an excellent concordance with anatomic pathological studies (0.86), showing 100% of sensitivity, 86% of specificity, 89% as PPV and 100% as NPV. CONCLUSIONS: MRI is useful for placental accretism diagnosis. The principal findings are transmural hyper-signal, the loss of continuity in myometrial wall in fast sequences and the identification of vessels invading myometrial layer in dynamic sequences.
Placenta accreta; Magnetic resonance imaging
