Supporting data for "Application of Deep learning in the radiographic image analysis of upper airway and adenoid"

Obstructive sleep apnoea (OSA) is a sleep disorder characterised by recurrent episodes of partial or complete upper airway obstruction during sleep. Recurrent interruptions in breathing during sleep lead to disrupted sleep patterns and inadequate oxygen supply to the body. The effects of chronic sleep disturbance and inadequate oxygen supply during childhood are concerning. Children with OSA often exhibit diverse symptoms, such as daytime sleepiness, hyperactivity and behavioural problems. Persistent OSA in children has been associated with various craniofacial deformities, such as a long face, constricted maxillary arch, narrow nasal base, and backward rotation of the mandible.

Currently, polysomnography (PSG) is the standard method used for diagnosing OSA; however, it is complex, time-consuming, and expensive. As numerous studies have indicated a correlation of airway morphology and adenoid hypertrophy (AH) with OSA severity, analysis of these structures through radiographic images can help healthcare professionals, particularly dentists, to identify certain anatomical features or abnormalities that could be associated with OSA.In recent years, artificial intelligence (AI) techniques have gained significant traction across various fields, leading to the development of numerous specialized branches. The clinical applications of dental image-based machine learning methods in orthodontics during the past decade were reviewed (Chapter 1). However, the application of AI techniques in upper airway and adenoid assessment has not been thoroughly explored. Therefore, we aimed to explore the possibilities of applying AI techniques to automatic upper airway and adenoid assessment to achieve more accurate and efficient evaluations.

In this thesis, the fully automatic AI-driven systems using convolutional neural network algorithms (CNNs) were developed and validated for upper airway segmentation and minimum cross-sectional area (CSAmin) localisation in two-dimensional (2D) radiographic airway images (Chapter 2). And cone-beam computed tomography (CBCT) images were taken as training data to develop AI models for upper airway segmentation and CSAmin localisation in three dimensions (Chapter 3). The AI models achieved segmentation accuracy exceeding 90.0%, and the height difference error between AI processing and human annotation was within 3 millimetres. In efficiency comparison, AI processing was much more efficient than manual processing in both tasks.

In addition to upper airway assessment, the AI model for children adenoid hypertrophy (AH) diagnosis was established. The diagnostic model Category-based Relation Consistency Mean Teacher Network (CRC-MT) was established for AH diagnosis based on 679 latreal cephalograms obtained from 12-year-old Chinese and Caucasian children. In comparison to human detection results, the AI methods exhibited significant higher accuracy across various evaluation metrics. In blocked region tests, our results showed that AI methods may detect features and characteristics from images that may not be discernible to the human eyes. .

In conclusion, our study contributes to the advancement of AI applications in upper airway and adenoid imaging processing, with potential implications for broader applications in orthodontics. Regarding AI applications, upper airway and adenoid imaging assessment are basic and initial steps in the development of AI systems for further pathology detection and disease diagnosis. From the clinical perspective, this study provided an innovative AI-driven method that will help improve clinicians’ efficiency.