AI enables preterm birth prediction at 31 weeks By examining a pregnant woman’s uterine electrical activity, scientists were able to construct a deep learning algorithm that can predict premature deliveries.The CDC found that in 2021, preterm births (those occurring before 37 weeks of gestation) accounted for 10% of all births in the United States.The study also found that preterm births were a contributing factor in the deaths of roughly 16% of infants.Researchers at Washington University in St. Louis, Missouri, are currently seeking to use AI to improve these odds.
Scientists developed a deep learning algorithm that can predict preterm births by analysing a pregnant woman’s uterine electrical activity. Next, they conducted an analysis of the model and published their findings in the peer-reviewed journal PLOS One.Professor of electrical engineering at Washington University in St. Louis Arye Nehorai, PhD, told Fox News Digital, “The key takeaway is that it is possible to take data as early as the 31st week and predict preterm birth up to the 37th week.”In addition, “the AI/deep learning automatically learned the most informative features from the data that are relevant to the prediction of preterm birth,” he said.
Nehorai elaborated by saying the findings demonstrate preterm delivery is an abnormal physiological state and not just a pregnancy that ends prematurely.Researchers used electrohysterograms (EHGs), which involve placing electrodes on the abdomen to monitor uterine electrical activity, as part of their study.Researchers “trained” an AI model with recordings of these electrical currents from 159 women who were at least 26 weeks along in their pregnancies.Women’s age, body mass index, foetal weight, and first- or second-trimester haemorrhage were all factors in determining the likelihood of a premature birth.Nehorai, when asked about the benefits of the new study, cited the low cost of construction as an advantage.
The researcher elaborated, saying, “Our model was effective in prediction with shorter EHG recordings, which could make the model easier to use, more cost-effective in a clinical setting, and possibly usable in a home setting.”The researchers advocated including EHG measurements in prenatal care, so expectant mothers might get help and make changes to their lifestyles if necessary to protect their unborn children.The researchers also remarked on how tricky it is to foretell if and when this type of test would be widely available.An EHG reading can take anywhere from 30 minutes to an hour, including the time it takes to position the equipment on the mother’s abdomen, according to Uri Goldsztejn, a PhD candidate in the department of biomedical engineering at Washington University working under the supervision of Professor Nehorai.
Emissions of greenhouse gases (GHG) measurements may postpone implementation in low-resource areas due to the necessity of new infrastructure, as the experts also pointed out.The study has two major limitations, both of which are not technology-related: the need for a more extensive dataset in order to properly test and create medical products, and the complexity of deep learning’s predictive abilities.In order to discover which medicines are most likely to prevent preterm birth and enhance outcomes, it is necessary to conduct additional medical examinations; however, pinpointing the factors underlying the algorithm’s predictions is challenging.