Using Preoperative and Intraoperative Factors to Predict the Risk of Surgical Site Infections After Lumbar Spinal Surgery: A Machine Learning-Based Study

Abstract

Objective: To develop a model based on machine learning to predict surgical site infection (SSI) risk in patients after lumbar spinal surgery (LSS).

Methods: Patients who developed postoperative SSI after LSS in the First Affiliated Hospital of Nanchang University between December 2010 and December 2019 were retrospectively reviewed. Preoperative and intraoperative variables, including age, diabetes mellitus, hypertension, body mass index, previous spinal surgery history, surgical duration, number of fused segments, blood loss, and surgical procedure were analyzed. Six machine learning algorithms-logistic regression, multilayer perceptron, decision tree, random forest, gradient boosting machine, and extreme gradient boosting-were used to build prediction models. The performance of the models was evaluated using the area under the curve, accuracy, precision, sensitivity, and F1 score. A web predictor was developed based on the best-performing model.

Results: The study included 288 patients who underwent LSS, of whom 144 developed SSI and 144 did not develop SSI. The extreme gradient boosting model offers the best predictive performance among these 6 models (area under the curve = 0.923, accuracy = 0.860, precision = 0.900, sensitivity = 0.834, F1 score = 0.864). An extreme gradient boosting model-based web predictor was developed to predict SSI in patients after LSS.

Conclusions: This study developed a machine learning model and a web predictor for predicting SSI in patients after LSS, which may help clinicians screen high-risk patients, provide personalized treatment, and reduce the incidence of SSI after LSS.

Keywords: Lumbar; Machine learning; Prediction; Surgery; Surgical site infection.