AI Model Improves Detection of High-Risk Pregnancy Condition
A recent study highlights an artificial intelligence model that more accurately diagnoses placenta accreta spectrum (PAS), a life-threatening disorder. The technology analyzes patient risk factors and imaging data, reportedly outperforming current standard screening methods. This development points to the growing role of AI in perinatal risk management.
- The study, presented at the Society for Maternal-Fetal Medicine's 2026 meeting, was conducted by researchers at Baylor College of Medicine who retrospectively analyzed 2D ultrasound images from 113 high-risk patients at Texas Children's Hospital. - In the study, the AI model achieved 100% sensitivity, successfully identifying every confirmed case of placenta accreta spectrum (PAS); it produced two false positives but had no false negatives. - Current prenatal screening relies on ultrasound and identifying risk factors, but these methods fail to diagnose up to 50% of PAS cases before delivery. - The incidence of PAS has risen to as high as 1 in 272 pregnancies, largely attributed to increasing rates of cesarean deliveries. For women with placenta previa and two prior C-sections, the risk of developing accreta is as high as 40%. - Undiagnosed PAS is a leading cause of severe maternal morbidity and mortality, with the maternal death rate estimated to be as high as 7% in some studies. The condition can lead to massive maternal hemorrhage, organ failure, and the need for a hysterectomy. - This AI application is part of a broader trend of using machine learning in perinatal care to predict complications such as preeclampsia, gestational diabetes, and preterm labor by analyzing large datasets. - Other emerging digital technologies in high-risk pregnancy management include wearable devices for remote monitoring of biometrics like blood pressure and digital platforms designed to support patients with conditions like peripartum depression. - The full oral abstract detailing the AI model, titled "AI-based ultrasound screening for early, accurate identification of placenta accreta spectrum," is scheduled for publication in the February 2026 issue of the medical journal *PREGNANCY*.
