AI model could improve early detection of deadly pregnancy complication, new research suggests
A novel artificial intelligence model could improve the identification of a potentially deadly pregnancy complication that often goes undetected.
Placenta accreta spectrum happens when the placenta improperly attaches, typically too deeply, to the uterine wall. Due to its potential to trigger life-threatening hemorrhage during or after delivery, it is considered a high-risk complication of pregnancy and is a leading cause of maternal mortality and morbidity.
Despite the dangers associated with the condition, its presence often (up to half the time) evades detection during pregnancy. With instances of PAS on the rise, there is a critical need for methods of identifying the condition prior to delivery.
This week, during the Society for Maternal-Fetal Medicine (SMFM) 2026 Pregnancy Meeting, researchers shared new data on a novel AI tool they believe could help improve the diagnosis of PAS.
“Our team is very excited about the potential clinical implications of this model for accurate and timely diagnosis of PAS,” researcher Alexandra L. Hammerquist, MD, a maternal-fetal medicine fellow at Baylor College of Medicine in Houston, said during the presentation. “We are hopeful that its use as a screening tool will help decrease PAS-related maternal morbidity and mortality.”
Researchers trained a convolutional neural network (CNN) model on nearly 40,000 grayscale 2D placental ultrasound images. Mean gestational age at ultrasound was 30.89 ± 3.67 weeks. The model produced PAS scores based on US findings and patient data, including risk factors such as prior cesarean section.
In testing, the model predicted the presence or absence of PAS in 88% of cases, with a sensitivity of 100% and a specificity of 75%. It ably identified all women who were positive for PAS, yielding just two false positives and no false negatives.
Researchers acknowledged that more work is needed before the model can achieve clinical utility, but they maintained optimism regarding its performance.
“The results of the model support its use as a screening tool to reduce missed diagnoses in PAS screening, warranting future prospective trials,” the team concluded.
Read more research from the conference here.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.