Researchers at Washington University in St. Louis, Missouri, have developed an artificial intelligence (AI) model that can predict preterm births by analyzing the electrical activity in a pregnant woman’s uterus. A survey by the Centers for Disease Control and Prevention (CDC) found that 10% of babies born in the US in 2021 were preterm, and around 16% of baby deaths are attributed to preterm deliveries.
The researchers utilized electrohysterograms (EHGs) to recor
d the electrical currents in the uterus. They trained the AI model using EHG recordings from 159 pregnant women who were at least 26 weeks along. Factors such as the woman’s age, weight, fetal weight, and bleeding in the first or second trimester were taken into account to predict premature birth.
The study revealed that the AI model could predict preterm births by analyzing data as early as the 31st week of pregnancy, up until the 37th week. This suggests that preterm birth is not just an early ending to a pregnancy but an abnormal physiological condition. The researchers believe that incorporating EHG measurements into routine pregnancy check-ups could enable early intervention and lifestyle changes to protect babies.
One advantage of this AI model is its cost-effectiveness and potential ease of use in clinical or even home settings. However, there are limitations to consider. EHG readings typically take 30 to 60 minutes, including instrument placement, which may pose challenges in low-resource areas. Furthermore, the algorithm’s predictions lack interpretability, making it difficult to identify the specific causes behind its predictions. Additional medical examinations are needed to determine effective therapies to reduce preterm birth and improve outcomes.
While the researchers acknowledge the need for a larger dataset and further validation, this study represents a promising development in using AI to predict preterm births and potentially enhance prenatal care.