<h3>BACKGROUND CONTEXT</h3> Anterior cervical spinal fusion procedures have grown in prevalence due to excellent outcomes, low risk of perioperative outcomes and reduced length of hospitalization. The average age and comorbidity burden of patients undergoing anterior cervical fusion has increased. Given the significant cost and morbidity associated with developing major perioperative complications and unplanned readmissions, accurate risk stratification of patients undergoing this procedure is of great utility. Advanced machine learning (ML) methods have become increasingly employed in spinal surgery due to their ability to recognize complex, non-linear relationships between variables. ML models for preoperative risk stratification of patients undergoing anterior cervical fusion remain limited. <h3>PURPOSE</h3> We aim to develop a novel ensemble ML algorithm for prediction of major perioperative complication and readmission after anterior cervical fusion. Our secondary aim is to compare its performance against benchmark ML models and logistic regression (LR). <h3>STUDY DESIGN/SETTING</h3> Retrospective, cohort study. <h3>PATIENT SAMPLE</h3> Patients 18 years or older at a non-federal California hospital who underwent anterior cervical spinal fusion. <h3>OUTCOME MEASURES</h3> Readmission within 30 days, major perioperative complications (venous thromboembolism within 30 days, myocardial infarction within 7 days, pneumonia within 7 days, systemic infection within 7 days, surgical site bleeding within 90 days, and wound complications within 90 days). <h3>METHODS</h3> This is a retrospective cohort study of adult patients who underwent anterior cervical fusion at any California hospital between 2015-2017. We developed a ML-based model predicting complication risk using AutoPrognosis, an automated ML framework that configures an optimally performing ensemble of ML-based prognostic models. We compared our model with LR and four standard ML models (XGBoost, Gradient Boosting, AdaBoost, Random Forest). Predictive performances were assessed using area under the receiver operating characteristic curve (AUROC). Calibration was assessed using Brier scores. We ranked the contribution of the included features to model performance. <h3>RESULTS</h3> Of the 23,184 patients who met inclusion criteria, there were 1,886 cases of major complication or readmission (8.13%). The ensemble AutoPrognosis model had superior risk prediction (AUROC 0.724 + 0.015) compared to LR (0.717 + 0.012). Furthermore, our model outperformed the four standard ML algorithms. The AutoPrognosis model was well-calibrated (Brier score 0.069 + 0.01). The variables that are most important for AutoPrognosis include patient sex, workers' compensation insurance, age, malnutrition, and chronic obstructive pulmonary disease. Teaching hospital status, schizophrenia, and quadriplegia are also among the 10 most important features for AutoPrognosis model performance but not for LR. <h3>CONCLUSIONS</h3> We report the use of a novel ensemble ML algorithm for prediction of major perioperative complications after anterior cervical fusion. This algorithm is well-calibrated and demonstrates excellent risk prediction, superior to LR and four standard ML algorithms. By automating the choice of appropriate model as well as tuning of model hyperparameters, AutoPrognosis can be used by clinicians who may not possess an in-depth knowledge of ML methods – facilitating the use of advanced ML methods in mainstream clinical research. Notably, the predictors most important for AutoPrognosis are different from those for LR. This suggests that the superior discriminative capability of AutoPrognosis results from not just its ability to select among different ML models but also by capturing complex non-linear relationships between variables that LR is unable to capture. <h3>FDA DEVICE/DRUG STATUS</h3> This abstract does not discuss or include any applicable devices or drugs.
Read full abstract