Cervical varicosities may predict placenta accreta in posterior placenta previa: a magnetic resonance imaging study.

PURPOSE: The aim of this study was to prenatally predict placenta accreta in posterior placenta previa using magnetic resonance imaging (MRI). METHODS: This retrospective study was approved by the Institutional Review Board of our hospital. We identified 81 patients with singleton pregnancy who had undergone cesarean section due to posterior placenta previa at our hospital between January 2012 and December 2016. We calculated the sensitivity and specificity of several well-known findings, and of cervical varicosities quantified using magnetic resonance imaging, in predicting placenta accreta in posterior placenta previa. To quantify cervical varicosities, we calculated the A/B ratio, where "A" was the minimum distance from the most dorsal cervical varicosity to the deciduous placenta, and "B" was the minimum distance from the most dorsal cervical varicosity to the amniotic placenta. The appropriate cut-off value of the A/B ratio was determined using a receiver operating characteristic (ROC) curve. RESULTS: Three patients (3.7%) were diagnosed as having placenta accreta. The sensitivity and specificity of the well-known findings were 0 and 97.4%, respectively. Furthermore, the A/B ratio ranged from 0.02 to 0.79. ROC curve analysis revealed that the area under the combined placenta accreta and A/B ratio curve was 0.96. When the cutoff value of the A/B ratio was set 0.18, the sensitivity and specificity were 100 and 91%, respectively. CONCLUSION: It was difficult to diagnose placenta accreta in the posterior placenta previa using the well-known findings. The quantification of cervical varicosities could effectively predict placenta accreta.