Nomogram model based on magnetic resonance imaging and clinical factors for prenatal diagnosis of placenta accreta spectrum disorders.

This study aimed to develop a magnetic resonance imaging (MRI) and clinical factor-based nomogram for prenatal diagnosis of placenta accreta spectrum disorders (PAS). Clinical and MRI data from pregnant women were retrospectively analyzed and split into training (70%) and validation (30%) sets. Binary logistic regression identified age, placenta previa, T2 dark band, placental/uterine bulge, loss of T2 hypointense interface, myometrial thinning, and abnormal placental bed vessels as significant predictors of PAS. A nomogram incorporating these factors demonstrated strong diagnostic performance via ROC analysis (AUC >0.8), calibration curves, and decision curve analysis (DCA) in both training and validation sets. The model showed high accuracy and clinical utility, supporting its use in prenatal PAS diagnosis.