Chest · 2026
Personalized obesity hypoventilation syndrome risk assessment among bariatric surgery candidates through explainable machine learning: A multicenter cross-sectional study.
СтатьяНазвание 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.
Ni Min, Tang Yangyang, Zhang Junzheng, Lu Yujin, Liao Jiankun, Dong Zhiyong, Wang Cunchuan, Feng Jia, Chen Wenhui
Аннотация
BACKGROUND: Obesity hypoventilation syndrome (OHS) is frequently underdiagnosed, which substantially increases risk of perioperative complications. Limited provider awareness and the need for cumbersome diagnostic testing may hinder timely diagnosis. Machine learning (ML) may facilitate individualized OHS risk assessment by integrating multiple clinical and laboratory variables.
RESEARCH QUESTION: Can an explainable ML model using readily available clinical and laboratory variables accurately identify bariatric surgery candidates at risk for OHS.
STUDY DESIGN AND METHODS: This retrospective cross-sectional cohort study included 673 bariatric surgery candidates from the Chinese Obesity and Metabolic Surgery Database, randomly divided into training (70%, n = 471) and internal testing (30%, n = 202) cohorts. An independent cohort of 175 patients from another tertiary hospital was collected for external validation. Feature selection was performed using LASSO regression and Boruta, with bootstrap resampling used to assess selection stability. Ten ML models were constructed and evaluated using discrimination, calibration, and clinical utility measures. SHAP analysis was performed to enhance model interpretability, and a publicly accessible web-based calculator was deployed clinical application.
RESULTS: The final logistic regression model included 7 predictors: serum bicarbonate, type 2 diabetes, mean corpuscular hemoglobin concentration, tiredness, body mass index, observed apnea, and neck circumference. The model demonstrated good discrimination in the training, internal testing, and external validation cohorts, with AUCs of 0.884 (95%CI:0.844-0.917), 0.863 (95%:0.784-0.935) and 0.870 (95%:0.806-0.926). SHAP analysis identified serum bicarbonate as the most important predictor. The model showed favorable calibration and clinical utility as well as was implemented as an online calculator.
INTERPRETATION: Our explainable ML model integrating 7 readily available clinical variables demonstrated good predictive performance for OHS in bariatric surgery candidates. This model may facilitate earlier risk stratification and targeted diagnostic evaluation.
RESEARCH QUESTION: Can an explainable ML model using readily available clinical and laboratory variables accurately identify bariatric surgery candidates at risk for OHS.
STUDY DESIGN AND METHODS: This retrospective cross-sectional cohort study included 673 bariatric surgery candidates from the Chinese Obesity and Metabolic Surgery Database, randomly divided into training (70%, n = 471) and internal testing (30%, n = 202) cohorts. An independent cohort of 175 patients from another tertiary hospital was collected for external validation. Feature selection was performed using LASSO regression and Boruta, with bootstrap resampling used to assess selection stability. Ten ML models were constructed and evaluated using discrimination, calibration, and clinical utility measures. SHAP analysis was performed to enhance model interpretability, and a publicly accessible web-based calculator was deployed clinical application.
RESULTS: The final logistic regression model included 7 predictors: serum bicarbonate, type 2 diabetes, mean corpuscular hemoglobin concentration, tiredness, body mass index, observed apnea, and neck circumference. The model demonstrated good discrimination in the training, internal testing, and external validation cohorts, with AUCs of 0.884 (95%CI:0.844-0.917), 0.863 (95%:0.784-0.935) and 0.870 (95%:0.806-0.926). SHAP analysis identified serum bicarbonate as the most important predictor. The model showed favorable calibration and clinical utility as well as was implemented as an online calculator.
INTERPRETATION: Our explainable ML model integrating 7 readily available clinical variables demonstrated good predictive performance for OHS in bariatric surgery candidates. This model may facilitate earlier risk stratification and targeted diagnostic evaluation.