Detection of placenta accreta spectrum and prediction of adverse perinatal outcomes in pregnant women with placenta previa using ultrasonography and magnetic resonance imaging: A retrospective cohort study

BACKGROUND: Placenta accreta spectrum (PAS), an abnormal placental invasion into the myometrium or beyond the uterine serosa, is associated with adverse pregnancy outcomes. Placenta previa is the most significant individual risk factor for PAS, and Ultrasonography (US) and Magnetic resonance imaging (MRI) are widely used to detecting PAS. However, limited data exist on the effectiveness of US and MRI in predicting maternal and neonatal morbidity. This study evaluated the utility of these imaging modalities, not only in detecting PAS but also in predicting adverse perinatal outcomes. METHODS: This retrospective cohort study included 150 pregnant women with placenta previa who underwent US and MRI at a tertiary hospital between December 2019 and December 2023. PAS was diagnosed clinically or histopathologically after delivery. The predictive performance of US, MRI, and their combination was assessed using McNemar's test, receiver operating characteristic analysis, and trend analysis. RESULTS: Forty-one patients were diagnosed with PAS. The PAS group had significantly higher rates of prior caesarean section, maternal hemorrhagic outcomes, and neonatal complications than the non-PAS group. PAS-suspected on US group also showed increased estimated blood loss (EBL) and number of transfusion units administered, as well as a higher incidence of transfusion requirement, use of intrauterine balloon tamponade, hysterectomy, preterm birth before 37 weeks of gestation, neonatal ventilatory support and 1-minutes Apgar scores (AS) <7 compared to those with PAS-unsuspected on US group. PAS-suspected on MRI group had a significantly increased number of transfusion units administered and higher rates of hysterectomy, maternal ICU admission, preterm birth and neonatal complications. US plus MRI yielded progressively higher detection rates, outperforming either modality alone. A linear trend was observed in the increase of adverse perinatal outcomes when PAS was suspected on both US and MRI. CONCLUSIONS: US and MRI showed complementary strengths. US, but not MRI, is informative for predicting higher EBL, transfusion requirement, and use of intrauterine balloon tamponade, whereas MRI, but not US, predicts maternal ICU admission, preterm birth before 34 weeks of gestation, NICU admission, and 5-minute AS <7. Combining both imaging modalities yields better performance than either modality alone in detecting PAS and predicting adverse perinatal outcomes.