Two-Phase Deep Learning Approach for Diagnosing Pediatric Obstructive Sleep Apnea Using Lateral Cephalometric Images

1

A two-phase deep learning framework was developed to diagnose pediatric obstructive sleep apnea-hypopnea syndrome using lateral cephalometric images.
2

The framework achieved a mean Dice Similarity Coefficient of 0.931 for upper airway segmentation and an AUC of 0.945 for OSAHS classification.
3

AI assistance improved diagnostic accuracy by 0.165 for junior dentists and 0.237 for senior dentists in a reader study.
4

The study included 188 children, with 150 lateral cephalograms used for cross-validation and 38 for independent testing.
5

This model offers a promising automated tool for early detection and management of pediatric OSAHS in dental settings.

1