System Maintenance Notice

Please note that maintenance work is scheduled between 4 AM and 5 AM - CET. During this time, you may experience reduced performance. Thanks for your patience.

Publications

Journal Article Observational Study Ppublish 1 citation

PubMed: 39028323 DOI: 10.1007/s00540-024-03377-7

Pediatric cardiac surgery: machine learning models for postoperative complication prediction.

J Anesth 38(6):747-755

Authors: RĂ©mi Florquin, Renaud Florquin, Denis Schmartz, Philippe Dony, Giovanni Briganti

Affiliations

  • Department of Anesthesiology, CHU Charleroi, ChaussĂ©e de Bruxelles 140, 6042, Lodelinsart, Belgium.
  • Chair of Artificial Intelligence and Digital Medicine, Mons University, 7000, Mons, Belgium.
  • Floconsult SPRL, 1480, Tubize, Belgium.
  • Department of Anesthesiology, HĂ´pital Universitaire de Bruxelles (H.U.B), UniversitĂ© Libre de Bruxelles, 1070, Brussels, Belgium.
  • Department of Anesthesiology, CHU Charleroi, ChaussĂ©e de Bruxelles 140, 6042, Lodelinsart, Belgium.
  • Chair of Artificial Intelligence and Digital Medicine, Mons University, 7000, Mons, Belgium.

Abstract

PURPOSE

Managing children undergoing cardiac surgery with cardiopulmonary bypass (CPB) presents a significant challenge for anesthesiologists. Machine Learning (ML)-assisted tools have the potential to enhance the recognition of patients at risk of complications and predict potential issues, ultimately improving outcomes.

METHODS

We evaluated the prediction capacity of six models, ranging from logistic regression to support vector machine, using a dataset comprising 33 variables and 1364 subjects. The Area Under the Curve (AUC) and the F1 score served as the primary evaluation metrics. Our primary objectives were twofold: first, to develop an effective prediction model, and second, to create a user-friendly comprehensive model for identifying high-risk patients.

RESULTS

The logistic regression model demonstrated the highest effectiveness, achieving an AUC of 83.65%, and an F1 score of 0.7296, with balanced sensitivity and specificity of 77.94% and 76.47%, respectively. In comparison, the comprehensive three-layer decision tree model achieved an AUC of 72.84%, with sensitivity (79.41%) comparable to more complex models.

CONCLUSION

Our machine learning-assisted tools provide an additional perspective and enhance the predictive capabilities of traditional scoring methods. These tools can assist anesthesiologists in making well-informed decisions. Furthermore, we have successfully demonstrated the feasibility of creating a practical white-box model. The next steps involve conducting clinical validation and multicenter cross-validation.

TRIAL REGISTRATION

NCT05537168.

MeSH

Anesthesia, Machine Learning, Anesthesiologists, Learning, Postoperative Complications

Cited by