Risk Prediction for Postoperative Morbidity of Endovascular Aneurysm Repair Using Ensemble Model

摘要

Endovascular aneurysm repair (EVAR) is an advanced minimally invasive surgical technology that is helpful for reducing patients' recovery time and postoperative morbidity. This study proposes an ensemble model to predict postoperative morbidity after EVAR. The ensemble model was developed using a training set of consecutive patients who underwent EVAR between 2000 and 2008. The research outcomes consisted of an ensemble model to predict postoperative morbidity, the occurrence of postoperative complications pro-spectively recorded, and the causal-effect decision rules. The probabilities of complication calculated by the model were compared to the actual occurrence of complications and a receiver operating characteristic (ROC) curve was used to evaluate the accuracy of postoperative morbidity prediction. In this series, the ensemble of BN, NN and SVM models offered satisfactory performance in predicting postoperative morbidity after EVAR. Moreover, the Markov blankets of BN allow a natural form of causal-effect feature selection, which provides a basis for screening decision rules generated by granular computing.