Retinal AI Predicts Neonatal Lung Disease

Retinal scans already performed in the NICU could flag infants at risk for serious lung disease before standard diagnosis

1

Deep learning models using retinal images can predict bronchopulmonary dysplasia and pulmonary hypertension in premature infants.
2

The study utilized images from the i-ROP study, focusing on infants at 34 weeks' postmenstrual age or less.
3

A multimodal model combining retinal features and demographic data outperformed models using either data type alone for bronchopulmonary dysplasia.
4

For pulmonary hypertension, the imaging-only model demonstrated strong diagnostic capability, achieving an AUC of 0.91.
5

The findings suggest retinal imaging could aid in early identification of infants at risk for severe cardiopulmonary complications.

1