Scholarlink

<p>Pediatric obstructive sleep apnea (OSA) is a serious sleep disorder that can lead to long-term consequences, including learning difficulties, cardiovascular problems, and a delayed physical development. Although polysomnography (PSG) is an established gold standard of diagnosis of OSA, it is often expensive, takes a long time to achieve, and may not be very attainable. Nevertheless, alternative diagnostic measures have been examined, particularly those focusing on craniofacial features. In this review, we outline both manual approaches, such as cephalometric analysis, and AI-based techniques, including machine learning and deep learning applied to medical and facial images, which aim to automate parts of the manual assessment process. We review studies published between 2000 and 2025 that explored associations between facial characteristics and pediatric OSA, including those integrating imaging with clinical and demographic data for diagnostic prediction. In identifying trends in traditional and AI approaches, we identify the current status, hurdles, and future opportunities to develop more accessible and accurate measures for early detection of OSA in children.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A review of artificial intelligence and craniofacial imaging approaches for obstructive sleep apnea detection in children

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月17日

查看全文
  • Soulaimane Bahi,
  • Mohamed Youssfi,
  • Abdelmajid Bousselham,
  • Sara Retal,
  • Ouajih Ramzi

摘要

Pediatric obstructive sleep apnea (OSA) is a serious sleep disorder that can lead to long-term consequences, including learning difficulties, cardiovascular problems, and a delayed physical development. Although polysomnography (PSG) is an established gold standard of diagnosis of OSA, it is often expensive, takes a long time to achieve, and may not be very attainable. Nevertheless, alternative diagnostic measures have been examined, particularly those focusing on craniofacial features. In this review, we outline both manual approaches, such as cephalometric analysis, and AI-based techniques, including machine learning and deep learning applied to medical and facial images, which aim to automate parts of the manual assessment process. We review studies published between 2000 and 2025 that explored associations between facial characteristics and pediatric OSA, including those integrating imaging with clinical and demographic data for diagnostic prediction. In identifying trends in traditional and AI approaches, we identify the current status, hurdles, and future opportunities to develop more accessible and accurate measures for early detection of OSA in children.