AI Now Detects Health Risks in Premature Infants

The new AI model specifically targets two dangerous conditions in premature infants: bronchopulmonary dysplasia (BPD), a chronic lung disease, and pulmonary hypertension (PH), a type of high blood pressure affecting the lungs and heart. These are leading causes of illness and death in these vulnerable babies and are often difficult to diagnose early. Currently, diagnosing these conditions can be invasive and often happens after the diseases have progressed. For instance, BPD diagnosis relies on prolonged oxygen dependency, while PH often requires an echocardiogram (an ultrasound of the heart) to confirm. The AI taps into a screening that's already standard procedure for premature infants: eye exams for retinopathy of prematurity (ROP), a potential blinding disorder. The AI analyzes subtle patterns in the retinal blood vessels from these routine eye photos, patterns that are not visible to the human eye, to predict the risk of lung and heart issues. In a study involving 493 premature infants across seven neonatal intensive care units, the AI model demonstrated high accuracy. It predicted the risk of bronchopulmonary dysplasia with 82% accuracy and pulmonary hypertension with 91% accuracy. This breakthrough is part of an emerging field called "oculomics," where eye images are used to diagnose a wide range of systemic health problems, including cardiovascular and neurodegenerative diseases. The eye offers a unique, non-invasive window into the body's vascular health. Developed by researchers at the University of Colorado Anschutz, the study was published in the journal *JAMA Ophthalmology*. The lead author, Praveer Singh, PhD, highlighted that this approach could allow a simple photograph to provide deep insights into an infant's overall health.