January 29, 2025

By Shriti Pant, UNC Gillings School Communications Fellow

The use of Artificial Intelligence (AI) in medicine has improved pregnancy care worldwide.  

Knowledge of gestational age is crucial in maternal and neonatal care as it can help detect when to screen for gestational diabetes, administer certain vaccines, determine whether preterm delivery is anticipated or to assess if clinician-initiated delivery is needed.  

Research recently published in the Journal of the American Medical Association (JAMA) presents findings on how AI-enabled ultrasonography can be used by novice clinicians, such as midwives, to evaluate gestational age (GA) in low-resource settings. 

This research was funded by the Bill and Melinda Gates Foundation and conducted by researchers from UNC Gillings School of Global Public Health, UNC School of Medicine and University of Zambia School of Medicine. Jeff Stringer, MD, a professor of Obstetrics and Gynecology at the UNC School of Medicine, was the principal investigator. 

The study enrolled 400 pregnant individuals with viable first-trimester pregnancies from Chapel Hill, North Carolina, and Lusaka, Zambia. To establish the most accurate estimate of GA, a trained sonographer recorded transvaginal crown-rump length measurements. During follow up visits, novice users conducted blind ultrasound sweeps of the maternal abdomen using a handheld portable ultrasound device, the Butterfly QI+. The data from this battery-operated device was extracted onto Android tablets using an app to allow image processing and analysis in real time. Blind sweeps are conducted by scanning specific sections of the abdomen without targeting a predetermined region in order to detect any potential concerns during pregnancy. 

While conducting blind sweeps, the novice users were prevented from seeing any ultrasound image. The gestational age estimate from the model was also blinded during the study to minimize user bias. 

Using this low-cost AI device,  users with little to no training in sonography were able to estimate gestational age during the 14-to-27-week gestation period as accurately as trained sonographers who use advanced ultrasound machines.  

Dr. Teeranan (Ben) Pokaprakarn

Dr. Teeranan (Ben) Pokaprakarn

“The cost of large ultrasound machines and the lack of trained sonographers are significant barriers to accessing obstetric care in many low-resource settings,” explained Teeranan (Ben) Pokaprakarn, PhD, the UNC researcher who led the development of the AI model used in the study. 

He emphasized the importance of portable ultrasounds in addressing these challenges, saying, “As long as you have a tablet and the probe has batteries, it can be used anywhere.” 

“It’s rewarding to work on technology that can impact low-resource settings, unlike most tech designed for developed countries,” Pokaprakarn said. This project is unique because it is expanding healthcare access through AI. The simplicity of the ‘blind sweep,’ which takes only a few minutes and can be used by novices, makes it highly scalable. If it’s accurate and reliable, this solution could have a global impact on the well-being of mothers and babies.” 

Read the full journal article online. 


Contact the UNC Gillings School of Global Public Health communications team at sphcomm@unc.edu.

RELATED PAGES
CONTACT INFORMATION
Visit our communications and marketing team page.
Contact sphcomm@unc.edu with any media inquiries or general questions.

Communications and Marketing Office
125 Rosenau Hall
CB #7400
135 Dauer Drive
Chapel Hill, NC 27599-7400