Journal of clinical anesthesia · 2026
Predicting postoperative atrial fibrillation after cardiac surgery using machine learning.
СтатьяВалидационное исследованиеНазвание 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.
Cho Youn Joung, Ahn Jang Ho, Park Jin-Woo, Lee Hyeonhoon, Nam Karam, Ju Jae-Woo, Chung Jaeyeon, Kim In Jung, Lee Hyung-Chul, Jeon Yunseok
Аннотация
BACKGROUND: Postoperative atrial fibrillation (POAF) is a frequent complication after cardiac surgery that is associated with increased morbidity. However, existing risk scores have limited accuracy, highlighting the need for more effective prediction models. In this study, we developed and validated a machine-learning (ML) model predicting new-onset POAF following cardiac surgery using various perioperative variables.
METHODS: A total of 6859 adult patients who underwent cardiac surgery at Seoul National University Hospital (SNUH) or Seoul National University Bundang Hospital (SNUBH) between October 2004 and October 2021 were included. Top 20 perioperative variables were input to develop a prediction model based on extreme gradient boosting. The final model was internally validated using 516 patients from the SNUH test set and externally validated in 1701 patients from the SNUBH cohort. The areas under the receiver operating characteristic curve (AUROC) and precision recall curve (AUPRC), and 95% confidence intervals (CIs) were evaluated to assess model performance.
RESULTS: The incidence of new-onset POAF after cardiac surgery was 37.2%. The AUROC and AUPRC were 0.891 (95% CI, 0.859-0.920) and 0.861 (95% CI, 0.819-0.898), respectively, for internal validation and 0.726 (95% CI, 0.700-0.751) and 0.646 (95% CI, 0.607-0.684) for external validation. The negative predictive values were 0.867 and 0.773 for internal and external validation, respectively, supporting the model's utility in identifying low-risk patients.
CONCLUSIONS: A ML model developed with perioperative data demonstrated strong performance in predicting new-onset POAF after cardiac surgery. This approach may enable risk-stratified postoperative monitoring while providing the basis for improved perioperative decision-making.
METHODS: A total of 6859 adult patients who underwent cardiac surgery at Seoul National University Hospital (SNUH) or Seoul National University Bundang Hospital (SNUBH) between October 2004 and October 2021 were included. Top 20 perioperative variables were input to develop a prediction model based on extreme gradient boosting. The final model was internally validated using 516 patients from the SNUH test set and externally validated in 1701 patients from the SNUBH cohort. The areas under the receiver operating characteristic curve (AUROC) and precision recall curve (AUPRC), and 95% confidence intervals (CIs) were evaluated to assess model performance.
RESULTS: The incidence of new-onset POAF after cardiac surgery was 37.2%. The AUROC and AUPRC were 0.891 (95% CI, 0.859-0.920) and 0.861 (95% CI, 0.819-0.898), respectively, for internal validation and 0.726 (95% CI, 0.700-0.751) and 0.646 (95% CI, 0.607-0.684) for external validation. The negative predictive values were 0.867 and 0.773 for internal and external validation, respectively, supporting the model's utility in identifying low-risk patients.
CONCLUSIONS: A ML model developed with perioperative data demonstrated strong performance in predicting new-onset POAF after cardiac surgery. This approach may enable risk-stratified postoperative monitoring while providing the basis for improved perioperative decision-making.