Expert review of respiratory medicine · 2025
Application of artificial intelligence for the detection of obstructive sleep apnea based on clinical and demographic data: a systematic review.
Систематический обзорСтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Casal-Guisande Manuel, Mosteiro-Añón Mar, Torres-Durán María, Comesaña-Campos Alberto, Fernández-Villar Alberto
Аннотация
INTRODUCTION: Artificial intelligence (AI) has shown promise in enhancing the detection and stratification of obstructive sleep apnea (OSA) using clinical and demographic data. This systematic review assessed the effectiveness of AI models, methodological quality, and future research needs.
METHODS: Following PRISMA guidelines, a systematic search of PubMed (2014-2024) identified studies applying AI to detect or stratify OSA using clinical/demographic data, validated against polysomnography or cardiorespiratory polygraphy, and reporting performance metrics such as the area under the curve (AUC). Studies primarily based on wearable devices were excluded. Methodological quality and risk of bias were evaluated using the PROBAST tool.
RESULTS: Of 447 records, 26 met inclusion criteria. Common algorithms included decision trees, support vector machines, and neural networks, frequently using variables such as age, BMI, neck circumference, and comorbidities. AUC values ranged from 0.62 to 0.93, with most exceeding 0.80. Research output increased substantially between 2021 and 2024. Methodological heterogeneity and limited external validation hindered comparability. Exclusion of incomplete cases was a recurrent issue.
CONCLUSIONS: AI models show potential for improving OSA detection, but methodological limitations restrict generalizability. Future studies should prioritize external validation, diverse populations, and adherence to standardized reporting frameworks to enable clinical translation.
PROTOCOL REGISTRATION: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251025868 identifier is CRD420251025868.
METHODS: Following PRISMA guidelines, a systematic search of PubMed (2014-2024) identified studies applying AI to detect or stratify OSA using clinical/demographic data, validated against polysomnography or cardiorespiratory polygraphy, and reporting performance metrics such as the area under the curve (AUC). Studies primarily based on wearable devices were excluded. Methodological quality and risk of bias were evaluated using the PROBAST tool.
RESULTS: Of 447 records, 26 met inclusion criteria. Common algorithms included decision trees, support vector machines, and neural networks, frequently using variables such as age, BMI, neck circumference, and comorbidities. AUC values ranged from 0.62 to 0.93, with most exceeding 0.80. Research output increased substantially between 2021 and 2024. Methodological heterogeneity and limited external validation hindered comparability. Exclusion of incomplete cases was a recurrent issue.
CONCLUSIONS: AI models show potential for improving OSA detection, but methodological limitations restrict generalizability. Future studies should prioritize external validation, diverse populations, and adherence to standardized reporting frameworks to enable clinical translation.
PROTOCOL REGISTRATION: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251025868 identifier is CRD420251025868.