AI Tool Analyzes Placentas at Birth to Detect Problems in Newborns and Mothers More Quickly

A newly developed tool that harnesses computer vision and artificial intelligence (AI) could help clinicians quickly assess placentas at birth, improving care for newborns and mothers, according to the results of a new study by scientists at Northwestern Medicine and Penn State University. The study, published in the print edition of the journal Patterns, describes a computer program called PlacentaVision that can analyze a simple photograph of the placenta to detect abnormalities associated with infections and neonatal sepsis, a life-threatening condition that affects millions of infants worldwide.

Do Not Discard Placenta Without Examination

“The placenta is one of the most common things we see in the lab,” said study co-author Dr. Jeffery Goldstein, director of perinatal pathology and associate professor of pathology at the Feinberg School of Medicine at Northwestern University. “When a sick infant is being treated in the NICU, even a few minutes can make a difference in medical decision-making. With a diagnosis from these photos, we can get an answer days sooner than we would through our normal process.” Northwestern provided the largest set of images for the study, and Goldstein led the development and troubleshooting of the algorithms. Alison D. Gernand, senior researcher on the project, had the initial idea for this tool through her work in global health, particularly in pregnancies where women deliver at home due to a lack of medical resources. “Discarding the placenta without examining it is a common but often overlooked issue,” said Gernand, an associate professor in the Penn State College of Health and Human Development (HHD) Department of Nutritional Sciences. ”It’s a missed opportunity to detect problems and intervene early to reduce complications and improve outcomes for both mother and baby.”

Why Early Placental Examination is Important

The placenta plays a critical 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 could save lives and improve health outcomes,” said Yimu Pan, a doctoral student in the computer science program in the College of Information Sciences and Technology (IST) and lead author of the study. “It could make placental examination more accessible and benefit the research and care of future pregnancies, especially for mothers and babies at higher risk for complications.”

Early detection of placental infections through tools like PlacentaVision could enable clinicians to take prompt action, such as administering antibiotics to the mother or baby and closely monitoring the newborn for signs of infection, the scientists said. PlacentaVision is intended for use in a range of medical specialties, according to the researchers. “In low-resource settings – in places where hospitals don’t have pathology labs or specialists – this tool could help doctors quickly identify issues like infection from a placenta,” Pan said. “In well-equipped hospitals, the tool can ultimately help doctors determine which placentas require further detailed examination, making the process more efficient and ensuring that the most important cases are prioritized.”

Before such a tool could be used worldwide, we first had to overcome the core technical obstacles. This included making the model flexible enough to handle various placenta-related diagnoses and ensuring that the tool was robust enough to handle different delivery conditions, including variations in lighting, image quality, and clinical settings.

How the Tool Learned to Analyze Placenta Images

The researchers used cross-modal contrastive learning, an AI method for aligning and understanding the relationship between different types of data – in this case, visual (images) and textual (pathology reports) – to teach a computer program how to analyze placenta images. They collected a large, diverse dataset of placenta images and pathology reports from a 12-year period, studied how these images were associated with health outcomes, and built a model that could make predictions based on new images. The team also developed various image-processing strategies to simulate different imaging conditions so that the model’s robustness could be properly evaluated. The result was PlacentaCLIP+, a robust machine learning model that can analyze photos of placentas to detect health risks with high accuracy. It was validated across countries to confirm consistent performance in diverse populations.

According to the researchers, PlacentaVision is designed to be easy to use and could potentially work through a smartphone app or integrated into a medical record software, so that doctors can get quick answers after delivery. The next step is to develop a user-friendly mobile app that can be used by healthcare professionals – with minimal training – in clinics or hospitals with limited resources. The user-friendly app would enable doctors and nurses to photograph placentas and receive immediate feedback, improving care. The researchers plan to make the tool even smarter by incorporating more types of placental features and adding clinical data to improve predictions while contributing to research on long-term health. They will also test the tool in different hospitals to ensure it works in a variety of settings.