About 10% of babies born in the United States in 2021 will be premature, according to the Centers for Disease Control and Prevention (CDC). That means they gave birth earlier than 37 weeks of gestation.
Premature birth also accounts for about 16% of infant deaths.
Researchers at the University of Washington in St. Louis, Missouri, are now trying to increase that possibility in the following ways. Use of artificial intelligence.
They developed a deep learning model that can predict preterm birth by analyzing the electrical activity of the uterus of pregnant women, and tested the model in a study published in the medical journal PLOS One.
AI tools help doctors make sense of chaotic patient data and identify diseases: ‘more meaningful’ interactions
“The key point is that we can get data as early as 31 weeks and predict preterm births as early as 37 weeks,” said Arie, an electrical engineering professor, who surprised the researchers. Dr. Neholai said: Washington University in St. Louistold Fox News Digital.
Researchers at the University of Washington in St. Louis, Missouri, have developed a deep learning model (not shown) that can predict premature birth by analyzing electrical activity in the uterus of pregnant women. (St. Petersburg)
” AI/deep learning It automatically learns the most informative features from the data associated with predicting preterm birth,” he added.
Furthermore, the results of this study indicate that premature birth is not just an anomaly, but an abnormal physiological condition. A pregnancy that happens to end earlyNeborai said.
During the study, researchers performed electrohysterography (EHG), which uses electrodes in the abdomen to record electrical activity in the uterus.
They obtained recordings of these currents from 159 pregnant women who were at least 26 weeks old and “trained” an AI model on that data.
Ruptured brain aneurysm leaves social media influencer in medical coma after emergency caesarean section
The researchers combined this data with medical information such as the woman’s age and weight, fetal weight, and any bleeding experienced during the first or second trimester of pregnancy.
Nearly 19% of the women who participated in the study had premature births. In theory, data from these women could be used as a benchmark for predicting preterm birth.
According to the CDC, about 10% of babies born in the U.S. in 2021 will be preterm, meaning they were born before 37 weeks’ gestation. (St. Petersburg)
“The advantage of our approach is that it’s cheap to build,” Neholai said of the new study. “Our model was effective in predicting with shorter his EHG records, which makes the model easier to use, more cost-effective in clinical settings, and potentially usable in the home setting.” ”
Looking ahead, researchers believe the method should be adopted by hospitals and obstetricians as part of routine pregnancy testing for women. Pregnant women can then seek care when needed and make lifestyle changes to protect the health of their babies.
“Our study contributes to the goal of accurately predicting preterm birth using an EHG measuring device.”
“For this purpose, we have to build a dedicated device that implements our method,” said Nehorai.
The researchers said it’s difficult to say how long it will take for this kind of test to become widely available.
“There are some EHG measurement devices already on the market, but it’s been difficult to predict preterm birth from EHG data,” said Uli Goldstein, a Ph.D. speaks. University.
New AI ‘cancer chatbot’ offers 24/7 support to patients and families: ’empathetic approach’
“Our study contributes to the goal of using EHG measuring devices to accurately predict preterm birth,” he told Fox News Digital.
An EHG measurement typically takes 30 to 60 minutes, Goldsztejn said, but placing the device on the mother’s abdomen takes longer.
“We have shown that predictions can be made based on short EHG measurements of less than five minutes without significantly reducing prediction accuracy,” he told Fox News Digital. “This finding is important because long EHG measurements are a significant limitation for clinical implementation.”
The “promise” of deep learning – but a caveat
Dr. Suzy Lipinski, Board Certified Obstetrician-Gynecologist, Pediatrix Medical Group in Denver, ColoradoAlthough she was not involved in the research, she shared her opinion on whether deep learning technology could help solve the problem of premature births in the United States.
“Being able to predict who is at risk before birth begins is very beneficial,” Lipinski told Fox News Digital. “The use of deep learning models seems promising, but the relatively small number of patients in this study precludes us from determining how applicable this is to larger populations.”
“Being able to predict who is at risk before labor begins is very beneficial,” an obstetrician told Fox News Digital. (St. Petersburg)
“Previous research using AI More studies and larger patient populations will be needed before this method can be used because it has not been shown to be very reliable,” she added.
Another potential limitation, doctors noted, is that very few sites use EHG measurements.
“It’s standard in most hospitals and offices to use a tocodynamometer that measures pressure rather than electronically,” she explained.
Ozempic, Wegoby and Pregnancy Risk: What You Need to Know About the Issue
If EHG becomes a method for assessing preterm birth, hospitals, midwives and offices will all need to purchase new equipment, which could delay adoption in resource-poor areas such as rural and urban areas, Lipinski said. rice field.
“The higher preterm birth rate in this study than the national average also raises questions about its applicability,” she told Fox News Digital. “We don’t have demographics on the number of cases, so we have no way of knowing how that reflects the population of the country as a whole.”
“It would be very beneficial if we could predict who is at risk before birth.”
False positives are also possible, Lipinski said.
“While this method is more predictive than current methods, there are still many patients identified who are not at risk of premature birth,” she said. “This false-positive result will cause a lot of patient stress and increase medical device usage. medical resources. “
If this becomes the new standard of care, we will need to improve how we treat premature births, Lipinski said.
“We have two problems with premature birth: not only is it poorly predicted, but we are also poorly protected after 26 weeks,” she added.
Researchers share major limitations of study
The study has two main limitations, Goldsztejn said.
“First, we developed the study using about 160 samples from two public datasets,” he said. “While this amount of data was sufficient for initial investigations, development and validation of medical products will require even larger data sets.”
The results of the study show that premature birth is an abnormal physiological condition, not just an accidental premature termination of the pregnancy, said the lead investigator. (St. Petersburg)
A second limitation stems from the nature of deep learning, which can produce accurate results that are typically difficult to interpret, Goldsztejn said.
“In other words, it’s hard to understand how algorithms make predictions,” he explained.
Click here to sign up for our health newsletter
In discussing the findings of the study published in a medical journal, the authors said, “Machine learning algorithms can contribute to improving healthcare, and many studies have led to advances in this field, but important challenges remain. there is,” he pointed out.
“Medical product development and validation will require much larger datasets.”
In some of those challenges, it can be difficult to pinpoint the cause behind the algorithm’s predictions, the researchers wrote.
CLICK HERE TO GET THE FOX NEWS APP
“In our case, our predictions may have implications for pregnancy management, but to determine which treatments are more likely to reduce the risk of preterm birth and improve its outcome. , the predictions will need to be complemented by additional physical examinations,” the researchers also said.
