Screening Obstructive Sleep Apnea Patients Via Deep Learning Of Knowledge Distillation In The Lateral Cephalogram

This article discusses the use of lateral cephalogram in orthodontics as a valuable screening tool for the diagnosis of obstructive sleep apnea (OSA). A deep learning-based classifier was proposed to differentiate OSA from non-OSA patients using the lateral cephalogram. The model was trained using 500 OSA patients and 498 non-OSA patients from two different devices. The model was able to overcome modality differences and exhibited high performance in the classification of OSA and non-OSA patients.