AI tool analyzes placentas at birth for faster detection of neonatal, maternal problems

The study was published in the print edition of the journal on December 13 pattern The magazine’s cover describes a computer program called PlacentaVision that analyzes simple photos of the placenta to detect abnormalities related to infection and neonatal sepsis, a life-threatening disease that affects Millions of babies born around the world.

“The placenta is one of the most common specimens we see in the laboratory,” said study co-author Jeffrey Goldstein, Ph.D., director of perinatal pathology and director of pathology at Northwestern University Feinberg School of Medicine. Associate Professor. “When treating a sick child in the NICU, even a few minutes can have an impact on medical decisions. Through diagnosis from these photos, we can get answers days earlier than normal.”

Northwestern provided the largest set of images for the study, and Goldstein led algorithm development and troubleshooting.

Alison D. Gernand, the project’s principal investigator, conceived the original idea for the tool during her work in global health, specifically in situations where women are giving birth at home due to a lack of health care resources.

“Discarding the placenta without examining it is a common but often overlooked problem,” said Gernand, an associate professor in the Department of Nutritional Sciences in Penn State’s College of Health and Human Development (HHD). “This is a missed opportunity to identify problems and provide early intervention, thereby reducing complications and improving outcomes for mothers and babies.”

Why early placenta testing is important

The placenta plays a vital role in the health of the mother and baby during pregnancy, but it is often not thoroughly examined at birth, especially in areas with limited medical resources.

“This research can save lives and improve health outcomes,” said Yimu Pan, a doctoral student in the Informatics Program at the School of Information Science and Technology (IST) and lead author of the study. “It could make placental examinations more accessible, benefiting research and care in future pregnancies, especially for mothers and babies who are at higher risk for complications.”

Scientists say early identification of placental infection through tools such as PlacentaVision may allow clinicians to take swift action, such as giving antibiotics to the mother or baby and closely monitoring newborns for signs of infection.

Researchers say PlacentaVision is intended for use in a variety of medical populations.

“In resource-poor settings—where hospitals don’t have pathology labs or specialists—this tool could help doctors quickly detect problems like placental infection,” Pan said. “In well-equipped hospitals, this tool may eventually help doctors Determining which placentas require further, detailed examination makes the entire process more efficient and ensures the most important cases are prioritized.”

“Before such a tool can be deployed globally, the core technical hurdles we face are making the model flexible enough to handle the wide range of placenta-related diagnoses and ensuring the tool is robust enough to handle a wide range of delivery conditions, including “Our artificial intelligence tools need to maintain accuracy even if many of the training images come from well-equipped urban hospitals,” said James Z. Wang, Distinguished Professor at Penn State IST and one of the study’s principal investigators. Ensuring that PlacentaVision can handle a variety of real-world conditions is critical. “

How the tool learns to analyze placenta pictures

The researchers used cross-modal contrastive learning, an artificial intelligence method for aligning and understanding relationships between different types of data, in this case visuals (images) and text (pathology reports), to teach computer programs how to Analyze the imaged placenta. They assembled a large, diverse dataset of placental imaging and pathology reports spanning 12 years, studied how these images were related to health outcomes, and built a model that could make predictions based on the new images. The team also developed various image change strategies to simulate different photographing conditions in order to properly assess the model’s resilience.

The result is PlacentaCLIP+, a powerful machine learning model that analyzes placenta photos to detect health risks with high accuracy. It has been validated cross-nationally to confirm consistent performance across different populations.

PlacentaVision is designed to be easy to use and can be run through a smartphone app or integrated into medical records software so doctors can get answers quickly after delivery, the researchers said.

Next step: creating user-friendly apps for medical staff

“Our next steps include developing a user-friendly mobile app that can be used by medical professionals in low-resource clinics or hospitals with minimal training,” Pan said. “This user-friendly app will allow Doctors and nurses photograph the placenta and get immediate feedback and improve care.”

The researchers plan to make the tool smarter by incorporating more types of placental characteristics and adding clinical data to improve predictions while contributing to long-term health research. They will also test the tool in different hospitals to ensure it works in a variety of settings.

“This tool has the potential to change the way postpartum placentas are examined, especially in areas of the world where these tests are rarely performed,” Gernand said. “This innovation has the potential to provide greater access in resource-poor and resource-rich settings. With further refinement, it has the potential to transform neonatal and maternal care by enabling early, personalized intervention to prevent serious health outcomes and improve the lives of mothers and babies around the world”.

This research was supported by the National Institute of Biomedical Imaging and Bioengineering of the National Institutes of Health (grant R01EB030130). The team used supercomputing resources from the National Science Foundation-funded Advanced Network Infrastructure Coordination Ecosystem: Services and Support (ACCESS) program.