Predicting invasive placenta accreta spectrum and adverse clinical outcomes using magnetic resonance imaging combined with intravoxel incoherent motion parameters

To investigate the value of magnetic resonance imaging (MRI) combined with intravoxel incoherent motion (IVIM) parameters for predicting invasive placenta accreta spectrum (PAS) and adverse clinical outcomes. A total of 110 patients with PAS were retrospectively enrolled and divided into an invasive PAS group ( n = 47) and a non-invasive PAS group ( n = 63). Clinical data, conventional MRI findings, and IVIM parameters (f, D*, and D) were compared. Multivariate logistic regression was used to identify independent predictors and construct a prediction model. Model performance was evaluated by discrimination and calibration; internal validation was performed using bootstrap resampling (1,000 iterations). Receiver operating characteristic (ROC) analysis was used to evaluate predictive performance for adverse clinical outcomes. Six independent predictors were identified. The logistic regression model achieved an original area under the ROC curve (AUC) of 0.926 (95%CI: 0.889–0.953); after bootstrap internal validation, the optimism-corrected AUC was 0.887 (95%CI: 0.841–0.933). The combined IVIM parameters predicted adverse clinical outcomes with an AUC of 0.866. Invasive PAS was significantly associated with adverse neonatal outcomes (RR = 5.203, 95%CI: 1.646–16.446). The overall association with fetal congenital anomalies was not statistically significant (RR = 6.787, 95%CI: 0.939–49.039), none of the individual malformation categories reached statistical significance, and these fetal-anomaly estimates were based on only eight events. A model combining conventional MRI features with IVIM-derived parameters identified invasive PAS with good internally validated performance. IVIM parameters also showed discriminatory ability for adverse clinical outcomes. Invasive PAS was significantly associated with adverse neonatal outcomes, whereas fetal congenital anomalies and all individual malformation categories did not reach statistical significance; no established association with congenital anomalies can be inferred from this dataset. External validation in independent cohorts is required before clinical implementation.