Abstract
PURPOSE: To investigate the predictive value of cervical volume, placental volume, and cervical length for placental accreta (PAS) in patients with complete placenta previa and a history of cesarean section. PATIENTS AND METHODS: Three radiologists analyzed MRI images from 157 patients, using 3D Slicer to measure placental volume, cervical volume, and cervical length. The predictive efficacy of combining these three factors for PAS was assessed by plotting Receiver Operating Characteristic (ROC) curves. A scoring system was established, and its predictive accuracy relative to the combined model was evaluated by comparing the Area under Curve (AUC). RESULTS: Among 157 patients, 72 were diagnosed with PAS. Results showed that increased placental volume and decreased cervical volume and length were significantly associated with PAS occurrence (P < 0.001). The combined AUC for predicting PAS was 0.891, with sensitivity of 87.31%, specificity of 83.95%, positive predictive value (PPV) of 82.16%, and negative predictive value (NPV) of 88.65%. The scoring system demonstrated superior predictive performance, with AUC increasing to 0.902, sensitivity at 88.72%, specificity at 84.21%, positive predictive value (PPV) at 82.66%, and negative predictive value (NPV) at 89.82%. CONCLUSION: Cervical volume, placental volume, and cervical length are effective indicators for predicting preterm labor. A scoring system developed based on these factors demonstrates higher accuracy.