The medical information that can be gained from examining the placenta immediately after a birth is very valuable. To make this examination better, researchers have developed an AI tool that can help analyze the placenta more quickly and better diagnose any neonatal and maternal problems.
The AI tool is part of PlacentaVision, a computer program that, by taking a picture of it, allows the placenta to be analyzed and examined for abnormalities such as infections or neonatal sepsis, a condition that affects millions of newborns worldwide. The research, conducted by scientists at Northwestern Medicine and Penn State, was recently published in Patterns.
AI tool development
To develop the AI tool, the researchers used cross-modal contrastive learning. This is an AI method for matching and understanding relationships between different types of data - in this case, visual (images) and textual (pathology reports). With this, the computer program was then taught how to analyze pictures of placentas.
To do so, a large, diverse dataset of placental images and pathology reports was collected over a 12-year period, which was then studied by the researchers to determine how these images were related to health outcomes. With that knowledge, a model was built that could make predictions based on new images. The team also developed different image change strategies to simulate different conditions when taking pictures so that the resilience of the model could be properly evaluated.
Machine learning model
From this emerged PlacentaCLIP+, a robust machine-learning model that can analyze photographs of placentas to detect health risks with high accuracy. It was validated internationally to confirm consistent performance across populations. According to the researchers, PlacentaVision is designed for easy use, possibly through a smartphone app or integrated into medical records software, so that doctors can get answers quickly after delivery.
“Our next step is to develop an easy-to-use mobile app that can be used by medical staff - with minimal training - in clinics or hospitals with limited resources,” they said. The user-friendly app would allow doctors and nurses to photograph placentas and get immediate feedback and improve care,” said Yimu Pan, a doctoral student in the computer science program at the College of Information Sciences and Technology (IST) and lead author of the study.
Placental research
The placenta plays a vital role in the health of both the pregnant woman and baby during pregnancy, but is often not thoroughly examined at birth, especially in areas with limited medical resources. When infections or other conditions that can be detected earlier by examining the placenta with tools such as the now-developed AI-powered PlacentaVision, it allows doctors to take quick action. Consider administering antibiotics to the mother or baby and closely monitoring the newborn for signs of infection.
“When the neonatal intensive care unit is treating a sick infant, even a few minutes can make a difference in medical decision-making. With a diagnosis based on these photographs, we can get an answer days earlier than in our normal process,” says Dr. Jeffery Goldstein, director of perinatal pathology and associate professor of pathology at Northwestern University Feinberg School of Medicine.
Placental research is also being conducted in the Netherlands. In 2023, Dr. Michelle Broekhuizen received her doctorate on this topic. At Erasmus MC's Placenta Lab, which has a device that can keep the placenta alive for 4 hours after birth, Broekhuizen examined hundreds of placentas from both healthy and mothers with conditions such as pre-eclampsia, high blood pressure or gestational diabetes.
