Story perspectives

AI Model Detects Lung Risks in Premature Infants with High Accuracy

3/3/2026

Summary

Story summary

Researchers developed an artificial intelligence model that analyzes routine retinal images from retinopathy of prematurity (ROP) screenings to identify risks of bronchopulmonary dysplasia and pulmonary hypertension.
The study, published in JAMA Ophthalmology, reports the AI model achieved 82% accuracy for bronchopulmonary dysplasia and 91% for pulmonary hypertension.
Data came from 493 premature infants across seven United States NICUs.
Further validation studies are necessary before integrating this technology into clinical care.

History

Historical Context

The FDA's approval of IBM Watson Health's AI-based diagnostic tool in 2016 marked a pivotal moment in healthcare, achieving a 90% accuracy rate in cancer detection. This advancement reflects a broader trend in AI's integration into medicine, enhancing early disease detection and patient outcomes, particularly in neonatal care, as seen in recent models analyzing retinal images in premature infants.

1 of 2

Story summary

Researchers developed an artificial intelligence model that analyzes routine retinal images from retinopathy of prematurity (ROP) screenings to identify risks of bronchopulmonary dysplasia and pulmonary hypertension.
The study, published in JAMA Ophthalmology, reports the AI model achieved 82% accuracy for bronchopulmonary dysplasia and 91% for pulmonary hypertension.
Data came from 493 premature infants across seven United States NICUs.
Further validation studies are necessary before integrating this technology into clinical care.

1 / 2