BMC endocrine disorders · 2026
Machine learning-assisted prediction of obstructive sleep apnea in patients with metabolic dysfunction-associated steatotic liver disease: a multicenter study with external validation.
СтатьяМногоцентровое исследованиеВалидационное исследованиеНазвание 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.
Gao Qihang, Wu Yahui, Lei Ke, Zhao Tie, Xing Yixiang, Kang Min, Shen Yun
Аннотация
BACKGROUND: Metabolic dysfunction-associated steatotic liver disease (MASLD) is closely associated with obstructive sleep apnea (OSA) through shared metabolic and inflammatory mechanisms. Early identification of high-risk individuals remains challenging, highlighting the need for reliable predictive tools.
METHODS: This retrospective multicenter study included patients with MASLD from two tertiary hospitals in China. Least absolute shrinkage and selection operator (LASSO) regression combined with multivariate logistic regression was used for feature selection. Seven predictive models were constructed and compared, including logistic regression, decision tree, random forest, support vector machine, artificial neural network, XGBoost, and LightGBM. A nomogram was subsequently developed for individualized risk visualization.
RESULTS: A total of 310 patients from the Affiliated Heping Hospital of Changzhi Medical College were included as the development cohort, while 200 patients from Tongling People's Hospital served as the external validation cohort. LASSO regression initially identified multiple candidate predictors, and eight overlapping variables retained by both LASSO and multivariate analysis were ultimately selected, including sex, BMI, hypertension, T2DM, ALT, GGT, NLR, and TyG Index. Among the seven predictive algorithms, the logistic regression model demonstrated the most balanced performance in terms of discrimination, interpretability, calibration, and external stability. Decision curve analysis further confirmed favorable clinical utility across both cohorts.
CONCLUSION: The proposed logistic regression model provides an interpretable and clinically practical tool for early prediction of OSA risk in patients with MASLD.
METHODS: This retrospective multicenter study included patients with MASLD from two tertiary hospitals in China. Least absolute shrinkage and selection operator (LASSO) regression combined with multivariate logistic regression was used for feature selection. Seven predictive models were constructed and compared, including logistic regression, decision tree, random forest, support vector machine, artificial neural network, XGBoost, and LightGBM. A nomogram was subsequently developed for individualized risk visualization.
RESULTS: A total of 310 patients from the Affiliated Heping Hospital of Changzhi Medical College were included as the development cohort, while 200 patients from Tongling People's Hospital served as the external validation cohort. LASSO regression initially identified multiple candidate predictors, and eight overlapping variables retained by both LASSO and multivariate analysis were ultimately selected, including sex, BMI, hypertension, T2DM, ALT, GGT, NLR, and TyG Index. Among the seven predictive algorithms, the logistic regression model demonstrated the most balanced performance in terms of discrimination, interpretability, calibration, and external stability. Decision curve analysis further confirmed favorable clinical utility across both cohorts.
CONCLUSION: The proposed logistic regression model provides an interpretable and clinically practical tool for early prediction of OSA risk in patients with MASLD.