AI Accurately Detects Placenta Accreta, Offering Hope for Safer Pregnancies

A potentially life-saving artificial intelligence (AI) model is demonstrating remarkable accuracy in detecting placenta accreta spectrum (PAS) – a dangerous pregnancy complication – before childbirth, according to research presented at the Society for Maternal-Fetal Medicine (SMFM) 2026 Pregnancy Meeting™ in Las Vegas. This advancement addresses a critical gap in care, as current screening methods often fail to identify PAS until it’s too late.

Understanding Placenta Accreta Spectrum (PAS)

Placenta accreta occurs when the placenta abnormally attaches to the uterine wall, often following a cesarean delivery or other uterine surgery. This can lead to severe complications, including massive hemorrhage, organ failure and even maternal death. Currently, approximately half of all PAS cases go undiagnosed during pregnancy [1], leaving mothers and medical teams unprepared for the challenges of delivery.

How the AI Model Works

Researchers at Baylor College of Medicine developed the AI model by retrospectively analyzing nearly 40,000 2D obstetric ultrasound images from 113 high-risk pregnant women who delivered at Texas Children’s Hospital between 2018 and 2025 [2]. The model utilizes a convolutional neural network – a type of machine learning particularly adept at image analysis – to differentiate between a normal placenta and one affected by accreta.

The AI doesn’t simply analyze the outline of the placenta; it learns subtle patterns within the pixels of the ultrasound images and integrates clinical data, such as a history of cesarean section, to calculate a probability score indicating the likelihood of PAS [2].

Impressive Accuracy and Potential Impact

Testing of the algorithm revealed a high degree of accuracy, correctly identifying the presence or absence of placenta accreta 88% of the time. Notably, the model achieved 100% sensitivity – meaning it identified all cases of PAS – with a specificity of 75%, resulting in only two false positives and no false negatives [2].

“Our team is remarkably excited about the potential clinical implications of this model for accurate and timely diagnosis of PAS,” said Dr. Alexandra L. Hammerquist, a maternal-fetal medicine fellow at Baylor College of Medicine in Houston [1]. “We are hopeful that its apply as a screening tool will help decrease PAS-related maternal morbidity and mortality.”

Beyond Ultrasound: AI and Electronic Medical Records

Researchers are also exploring the use of machine learning models that utilize Electronic Medical Record (EMR) data to predict PAS risk even before conception [3]. One such model, developed by Mount Sinai, identified anemia as a potential modifiable risk factor, suggesting opportunities for early intervention and personalized care planning [3].

Looking Ahead

While the initial results are promising, further prospective trials are needed to validate the AI model’s performance in a broader clinical setting. The widespread adoption of this technology could significantly improve the identification and management of placenta accreta, ultimately leading to safer pregnancies and reduced maternal mortality.