THE ESSENTIAL
- Researchers have developed an AI capable of spotting placental infections by analyzing photos.
- Early identification of a placental infection could speed up the care of affected mothers and babies.
- The team plans to integrate their AI into an application so that it is accessible to as many healthcare professionals as possible.
“Discarding the placenta without examining it is a common, but often overlooked, problem”notes Alison D. Gernand of the Department of Nutritional Sciences at the Penn State College of Health and Human Development. “This is a missed opportunity to identify problems and provide early intervention that can reduce complications and improve outcomes for mother and baby.”
This is why the researcher and her colleagues have developed artificial intelligence capable of detecting signs of infections or neonatal sepsis in the placenta using a photo of the latter.
Their work was published in the journal PatternsDecember 13, 2024.
Placenta infections: photos analyzed by AI to detect anomalies
The researchers used cross-modal contrastive learning – an AI method capable of analyzing and relating different types of data (in this case, images and pathology reports) – to develop a tool that can spot signs of infections or other disorders in the placenta by analyzing photos. They took placental images and analysis results spanning a 12-year period to build a model capable of making predictions based on new images.
This artificial intelligence, called PlacentaCLIP+, presented very satisfactory results when it was put into real conditions. “The machine learning model has been validated internationally to confirm consistent performance across populations”specify the authors in their press release.
“The placenta is one of the most common specimens we observe in the laboratory”recalls Dr. Jeffery Goldstein, co-author of the study. “When the neonatal intensive care unit treats a sick child, even a few minutes can make the difference in medical decision-making. With a diagnosis made from these photographs, we can have an answer several days sooner than in part of our normal process.”
AI incorporated into an app for more accessible and rapid diagnostics
The team plans to improve its artificial intelligence so that the tool maintains its accuracy with photos taken during daily practices and to improve its accessibility.
“Our next steps include developing a user-friendly mobile app that can be used by healthcare professionals – with minimal training – in low-resource clinics or hospitals“, adds Yimu Pan, lead author of the study. “The user-friendly app would allow doctors and nurses to photograph placentas and get immediate feedback and improve care.”
“In low-resource areas, that is, places where hospitals do not have pathology laboratories or specialists, this tool could help doctors quickly spot problems such as infections due to the placenta”adds the researcher. “In well-equipped hospitals, the tool could help doctors determine which placentas need a more thorough and detailed examination, making the process more efficient and ensuring that the most important cases are prioritized.”
