Respiratory medicine · 2025
Integrating clinical phenotypes, impulse oscillometry and fractional exhaled nitric oxide: a robust machine learning model for early asthma prediction in preschool children.
СтатьяПоддержка исследования: вне правительства СШАНазвание и аннотация: Перевод в очереди
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.
Hu Ying, Hu Yuan, Dong Milan, Zheng Shouyan, Chen Zhiqiang, Zhou Ping, Li Xuemei, Liao Wei
Аннотация
BACKGROUND: Preschool wheeze diagnoses prove challenging due to the significant heterogeneity and lack of objective laboratory markers. To design and validate a predictive model powered by machine learning, which integrates impulse oscillometry (IOS), fractional of exhaled nitric oxide (FeNO), and clinical features to facilitate early asthma prediction in preschool children.
METHODS: A total of 264 preschool children with recurrent wheezing were identified from medical records. They were divided into training set and internal validation set. Twenty-four variables were analyzed, including clinical features, IOS, and FeNO measurements. Feature selection was performed using logistic regression and Lasso regression. To develop the asthma predictive model, seven different machine learning algorithms were utilized. Furthermore, data from 55 preschool children with recurrent wheezing at Chongqing Maternal and Child Health Hospital, China (2020-2021), were used for external validation.
RESULTS: Variable selection identified six predictors of asthma in preschool children: wheezing frequency, personal allergy history, FeNO, R5 percent of predicted, X5, and X5 percent of predicted. Based on these six factors, Logistic Regression model performance was best at an AUC of 0.88 (95 % CI: 0.79-0.97) internal validation and 0.89 (95 % CI: 0.80-0.97) for external validation. A nomogram derived from the Logistic Regression model was developed for free access by pediatricians.
CONCLUSION: This study innovatively incorporates IOS, FeNO, and clinical features to construct a logistic model, offering a robust predictive tool for early asthma identification in preschool children. However, as an exploratory investigation, these findings warrant validation through large-scale, multicenter prospective investigations to establish their generalizability across diverse clinical settings.
METHODS: A total of 264 preschool children with recurrent wheezing were identified from medical records. They were divided into training set and internal validation set. Twenty-four variables were analyzed, including clinical features, IOS, and FeNO measurements. Feature selection was performed using logistic regression and Lasso regression. To develop the asthma predictive model, seven different machine learning algorithms were utilized. Furthermore, data from 55 preschool children with recurrent wheezing at Chongqing Maternal and Child Health Hospital, China (2020-2021), were used for external validation.
RESULTS: Variable selection identified six predictors of asthma in preschool children: wheezing frequency, personal allergy history, FeNO, R5 percent of predicted, X5, and X5 percent of predicted. Based on these six factors, Logistic Regression model performance was best at an AUC of 0.88 (95 % CI: 0.79-0.97) internal validation and 0.89 (95 % CI: 0.80-0.97) for external validation. A nomogram derived from the Logistic Regression model was developed for free access by pediatricians.
CONCLUSION: This study innovatively incorporates IOS, FeNO, and clinical features to construct a logistic model, offering a robust predictive tool for early asthma identification in preschool children. However, as an exploratory investigation, these findings warrant validation through large-scale, multicenter prospective investigations to establish their generalizability across diverse clinical settings.