EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

An Explainable Machine Learning Model for Predicting Enteral Nutrition Feeding Intolerance in ICU Patients After Cardiopulmonary Bypass-Assisted Cardiac Surgery

Xueqing Chen, Shengqiang Zou, Yiting Wang, Jing Li, Ruijie Zong, Weiwei Chen, Yin Fang

The Heart Surgery Forum · 17.06.2026 · 29 (6)

DOI: 10.31083/hsf51865 ↗

Kurz gefasst

Fachartikel in der Zeitschrift The Heart Surgery Forum, online erschienen am 17.06.2026. Verfasst von Xueqing Chen und 6 weiteren Autor:innen. Die Arbeit nennt 55 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background:Enteral nutrition (EN) feeding intolerance (ENFI) frequently occurs in intensive care unit (ICU) patients after cardiopulmonary bypass (CPB)-assisted cardiac surgery and may adversely affect postoperative recovery. However, predictive tools specifically tailored to this population remain limited. This study aimed to develop and validate an explainable machine learning model to predict ENFI in postoperative ICU patients following CPB-assisted cardiac surgery.Methods:This retrospective study included 519 adult patients admitted to the ICU following CPB-assisted cardiac surgery between June 2019 and December 2023. ENFI was assessed within 5 days of EN initiation according to standardized clinical criteria. Thirty-nine perioperative variables were collected. The dataset was randomly divided into a training cohort (70%) and a testing cohort (30%). Least absolute shrinkage and selection operator (LASSO) regression was applied for feature selection within the training cohort. Five machine learning algorithms, including support vector machine (SVM), random forest (RF), decision tree (DT), naïve Bayes (NB), and extreme gradient boosting (XGBoost), were developed and compared. Model performance was evaluated using area under the receiver operating characteristic curve (AUC), calibration analysis, Brier score, and decision curve analysis (DCA). Shapley Additive Explanations (SHAP) were used to enhance interpretability.Results:Among the 519 patients, 253 (48.7%) developed ENFI. Fourteen predictors were selected by LASSO regression. In the independent testing cohort, the RF model demonstrated the highest discriminative performance, with an AUC of 0.864 (95% confidence interval [CI]: 0.804–0.916), sensitivity of 0.747, specificity of 0.815, and a Brier score of 0.155. Calibration curves indicated good agreement between predicted and observed probabilities, and DCA demonstrated favorable net clinical benefit across a range of threshold probabilities. SHAP analysis identified vasoactive-inotropic score (VIS), serum creatinine, lactate, delayed EN, and estimated glomerular filtration rate (eGFR) as the most influential predictors.Conclusions:An explainable machine learning model, particularly the RF algorithm, showed good discrimination and calibration for predicting ENFI in ICU patients after CPB-assisted cardiac surgery. The identified predictors highlight the importance of hemodynamic instability and organ dysfunction in the development of ENFI and may support early risk stratification and individualized nutritional management in postoperative ICU patients. However, as this model was developed and internally validated in a single-center cohort, further external validation in independent populations is required before broader clinical application.

Zitieren

APA

Chen, X., Zou, S., Wang, Y., Li, J., Zong, R., Chen, W., & Fang, Y. (2026). An Explainable Machine Learning Model for Predicting Enteral Nutrition Feeding Intolerance in ICU Patients After Cardiopulmonary Bypass-Assisted Cardiac Surgery. The Heart Surgery Forum, 29(6). https://doi.org/10.31083/hsf51865

Vancouver

Chen X, Zou S, Wang Y, Li J, Zong R, Chen W, et al. An Explainable Machine Learning Model for Predicting Enteral Nutrition Feeding Intolerance in ICU Patients After Cardiopulmonary Bypass-Assisted Cardiac Surgery. The Heart Surgery Forum. 2026;29(6). doi:10.31083/hsf51865

BibTeX
@article{Chen2026,
  author = {Chen, Xueqing and Zou, Shengqiang and Wang, Yiting and Li, Jing and Zong, Ruijie and Chen, Weiwei and Fang, Yin},
  title = {An Explainable Machine Learning Model for Predicting Enteral Nutrition Feeding Intolerance in ICU Patients After Cardiopulmonary Bypass-Assisted Cardiac Surgery},
  journal = {The Heart Surgery Forum},
  year = {2026},
  volume = {29},
  number = {6},
  doi = {10.31083/hsf51865}
}
Kostenlos starten →