A Primary Screening Predictive Model for Abnormally Invasive Placenta Based on Clinical and Basic Ultrasound Factors in Singleton Pregnancies.

OBJECTIVES

This study investigated risk factors for abnormally invasive placenta (AIP) in singleton pregnancies and developed a risk prediction model using clinical and basic ultrasound features for AIP screening.

METHODS

A retrospective analysis was conducted on 16,735 singleton pregnancies (with/without AIP) delivering at the First Affiliated Hospital of Sun Yat-sen University (2017-2021). Eighteen clinical and 3 basic ultrasound features were recorded. AIP was confirmed via placental pathology or intraoperative observations. The cohort was randomly split into development (70%) and validation (30%) sets. Univariate/multivariate logistic analyses and multicollinearity tests identified AIP predictors. Three models (clinical-only, basic ultrasound-only, combined) were constructed; their performances were compared to select the final model, visualized as a nomogram. Model performance was evaluated using receiver operating characteristic curves (ROC), area under the curves (AUC), calibration curves, and decision curve analysis (DCA) (1000 bootstraps, 5-fold cross-validation).

RESULTS

AIP was confirmed in 161 cases (0.962%). Independent predictors included a history of placenta accreta spectrum (PAS), advanced maternal age, assisted reproductive technologies (ART), pre-pregnancy obesity, placenta previa/low-lying placenta, and placental location (p < 0.05). In the validation set, the combined model had an AUC of 0.844 (95% CI 0.775-0.913), versus 0.781 (95% CI 0.701-0.862) for basic ultrasound-only and 0.698 (95% CI 0.626-0.770) for clinical-only. The nomogram of the combined model showed good discrimination; calibration curves and DCA confirmed satisfactory calibration and clinical utility.

CONCLUSION

The combined clinical-basic ultrasound model outperforms single-component models in AIP prediction. Its nomogram has good predictive performance, enabling clinicians to quickly screen high-risk pregnant women for AIP.