BPD-Neo: An MRI Dataset for Lung-Trachea Segmentation with Clinical Data for Neonatal Bronchopulmonary Dysplasia

Bronchopulmonary dysplasia (BPD) is a common complication among preterm neonates,with portable X-ray imaging serving as the standard diagnostic modality in neonatal in-tensive care units (NICUs). However, lung magnetic resonance imaging (MRI) offers anon-invasive alternative that avoids sedation and radiation while providing detailed insightsinto the underlying mechanisms of BPD. Leveraging high-resolution 3D MRI data, advancedimage processing and semantic segmentation algorithms can be developed to assist cliniciansin identifying the etiology of BPD. In this dataset, we present MRI scans paired with corre-sponding semantic segmentations of the lungs and trachea for 40 neonates, the majority ofwhom are diagnosed with BPD. The imaging data consist of free-breathing 3D stack-of-starsradial gradient echo acquisitions, known as the StarVIBE series. Additionally, we providecomprehensive clinical data and baseline segmentation models, validated against clinicalassessments, to support further research and development in neonatal lung imaging.