基于深度学习的新生儿呼吸系统疾病的多类分类在新生儿重症监护病房的胸部放射图上
Hye Won Cho1,2, Sumin Jung3, Kyu Hee Park1
1Department of Pediatrics, Korea University Ansan Hospital, Ansan-si, Republic of Korea.
Neonatology
|March 6, 2025
概括
一个新的深度学习算法准确地从胸部X射线中分类新生儿呼吸系统疾病,帮助做出关键护理决策. 这种人工智能工具支持新生儿科医生诊断诸如呼吸困扰综合征和支气管肺功能障碍等疾病.
科学领域:
- 人工智能在医学中的应用
- 新生儿成像分析分析
- 计算病理学计算病理学
背景情况:
- 准确解释新生儿胸部X射线图对于管理重症新生儿至关重要.
- 深度学习为复杂的医学图像的自动化分析提供了潜力.
- 新生儿重症监护室 (NICU) 需要有效的呼吸道疾病诊断工具.
研究的目的:
- 开发和验证一种深度学习算法,通过胸部X射线对六种常见的新生儿呼吸系统疾病进行分类.
- 对专家新生儿学家的分类来评估算法的性能.
- 提供一个工具,支持新生儿护理的及时和准确的临床决策.
主要方法:
- 一个大数据集由43,338个新生儿胸部X射线 (34,598培训,4,370验证,4,370测试) 来自10个韩国大学医院.
- 图像由20名新生儿科医生手动对六种情况进行分类:健康的肺部,呼吸困扰综合征 (RDS),新生儿短暂的呼吸暂停 (TTN),空气泄漏综合征 (ALS),乳腺排泄症和支气管肺功能失调症 (BPD).
- 在这个数据集上训练了一种修改后的ResNet50深度学习模型,其中包含了人口统计学变量.
主要成果:
- 自动化算法实现了总体测试准确率为83.96%和F1得分为83.68%,表明与人类专家分类的高度一致.
- 在特定疾病中观察到高F1分数:BPD (92.19%),ALS (90.65%),RDS (90.30%),以及健康的肺 (87.38%).
- 该算法显示了对阿特莱克塔斯 (86.56%) 和TTN (70.84%) 的有希望的性能.
结论:
- 成功开发了一种深度学习算法,用于使用多类标记的胸部X射线对新生儿呼吸系统疾病进行分类.
- 该算法显示有潜力帮助新生儿科医生更快,更准确地为重症新生儿做出诊断决策.
- 非成像数据的整合进一步提高了该算法的实用性,以支持临床管理.
相关概念视频
Assessment of Airway, Skin Color, and Use of Accessory Muscles
950
A thorough assessment of respiratory health is paramount in clinical settings to identify and manage respiratory distress and ensure adequate oxygenation. This article elaborates on the critical aspects of respiratory evaluation, including airway assessment, skin color examination, and the observation of accessory muscle use, which are integral to effectively diagnosing and managing patients with respiratory conditions.
Introduction
The initial evaluation of a patient's respiratory system...
Introduction
The initial evaluation of a patient's respiratory system...
950
Assessment of Ventilation II: Respiratory Depth and Rhythm
1.3K
Respiratory Depth
Respiratory depth measures the volume of air inhaled or exhaled during a breath. It can vary from shallow to deep and typically remains consistent when a person is at rest or asleep. Occasionally, individuals will automatically inhale deeply, known as sighing, which inflates the lungs with more air than normal breathing.
To assess respiratory depth, observe the degree of chest excursion or movement:
Respiratory depth measures the volume of air inhaled or exhaled during a breath. It can vary from shallow to deep and typically remains consistent when a person is at rest or asleep. Occasionally, individuals will automatically inhale deeply, known as sighing, which inflates the lungs with more air than normal breathing.
To assess respiratory depth, observe the degree of chest excursion or movement:
1.3K
Assessment of Respiration
1.0K
The respiratory system's basic structures and primary functions lay the foundation for nurses' comprehensive respiratory assessments. This assessment includes subjective and objective data to gauge the patient's respiratory health.
Subjective Assessment: Nurses interview the patient to gather information directly during the subjective assessment. It includes questions about the individual's medical history, medications, and symptoms, focusing on past respiratory conditions like...
Subjective Assessment: Nurses interview the patient to gather information directly during the subjective assessment. It includes questions about the individual's medical history, medications, and symptoms, focusing on past respiratory conditions like...
1.0K
Respiratory System Abnormal Finding I: Inspection and Percussion
192
Respiratory system abnormalities are a significant concern in healthcare due to their potential to indicate underlying severe conditions like Chronic Obstructive Pulmonary Disease (COPD), asthma, and pneumonia. These abnormalities can often be detected through physical examination methods like inspection and percussion.
Inspection Findings
During an inspection, several findings may suggest the presence of respiratory distress or disease. Pursed-lip breathing, where exhalation is slowed by...
Inspection Findings
During an inspection, several findings may suggest the presence of respiratory distress or disease. Pursed-lip breathing, where exhalation is slowed by...
192


