A New Model for Early Prediction of Persistent Low-Lying Placenta

Abstract

Objective Finding a predictive model for persistent low lying placenta (LLP) based on early clinical and laboratory parameters. Methods This retrospective cohort study included patients with LLP detected during early anatomy scan. Additional transvaginal ultrasound exams assessed for resolution at 22–24 weeks and 36–39 weeks. Patients were categorized as: Group 1–LLP resolved by the second trimester scan, Group 2- LLP resolved by the third, and Group 3–LLP persisted to delivery. Clinical and laboratory parameters were compared between groups. A linear support vector machine classification was used to define a prediction model for persistence. Results Among 236 pregnancies with LLP, 80% resolved by 22–24 weeks, 10.5% resolved by 36–39 weeks and 9.5% persisted until delivery. Second trimester hCG levels were higher the longer the LLP persisted (0.8MoM + 0.7 vs. 1.13MoM + 0.4 vs. 1.7MoM + 1.5, P = 0.03, respectively) and Cervical length was shorter (P = 0.008, P = 0.02, respectively). A linear SVM classification model was calculated based on these parameters. The predictive accuracy of this model was 90.3%. Conclusion LLP persistence can be predicted with an accuracy of 90.3%, as early as the beginning of the second trimester. Persistence vs. resolution of LLP may represent different entities and not a spectrum of the same condition over time.