XGBoost, a Machine Learning Method, Predicts Neurological Recovery in Patients with Cervical Spinal Cord Injury

Abstract

The accurate prediction of neurological outcomes in patients with cervical spinal cord injury (SCI) is difficult because of heterogeneity in patient characteristics, treatment strategies, and radiographic findings. Although machine learning algorithms may increase the accuracy of outcome predictions in various fields, limited information is available on their efficacy in the management of SCI. We analyzed data from 165 patients with cervical SCI, and extracted important factors for predicting prognoses. Extreme gradient boosting (XGBoost) as a machine learning model was applied to assess the reliability of a machine learning algorithm to predict neurological outcomes compared with that of conventional methodology, such as a logistic regression or decision tree. We used regularly obtainable data as predictors, such as demographics, magnetic resonance variables, and treatment strategies. Predictive tools, including XGBoost, a logistic regression, and a decision tree, were applied to predict neurological improvements in the functional motor status (ASIA [American Spinal Injury Association] Impairment Scale [AIS] D and E) 6 months after injury. We evaluated predictive performance, including accuracy and the area under the receiver operating characteristic curve (AUC). Regarding predictions of neurological improvements in patients with cervical SCI, XGBoost had the highest accuracy (81.1%), followed by the logistic regression (80.6%) and the decision tree (78.8%). Regarding AUC, the logistic regression showed 0.877, followed by XGBoost (0.867) and the decision tree (0.753). XGBoost reliably predicted neurological alterations in patients with cervical SCI. The utilization of predictive machine learning algorithms may enhance personalized management choices through pre-treatment categorization of patients.

Expert-augmented automated machine learning optimizes hemodynamic predictors of spinal cord injury outcome

“…Alongside the relatively limited sample size of available clinical SCI datasets, the variety of SCI characteristics presents significant challenges for reproducible identification of patient outcome predictors despite the volume of data collected throughout patient hospitalization and treatment [ 20 ]. Various prognostic models for SCI outcome have been developed with algorithms ranging from logistic regression to extreme gradient boosted (XGB) trees and convolutional neural networks [ 12 , 21 , 22 ]. While such studies bear potential for informing clinical care, algorithm selection in many SCI ML studies have primarily depended on the researchers’ familiarity with specific ML algorithms, and prediction accuracy remains the primary metric for comparing models [ 23 , 24 ].…”

Efficacy of a machine learning-based approach in predicting neurological prognosis of cervical spinal cord injury patients following urgent surgery

“…Inoue and colleagues proposed models predicting neurological outcomes 6 months after injury based on the binary classi cation of AIS grade A/B/C or D/E, and XGBoost had the highest accuracy (81.1%). 7 For predicting reacquisition of walking ability, the predictive model with a binary classi cation of improvement to more than AIS grade D is useful. However, not only predicting motor ability but also predicting recovery of one AIS grade or two AIS grades (eg, AIS grade A to B or AIS grade A to C), is considered clinically meaningful neurological recovery from the patient's perspective.…”

“…A previous study reported that ML models were useful for predicting neurological improvements in patients with CSCI. 7 However, it included cases that received both surgical and conservative treatment, and the outcome of the prediction was a binary classi cation of the American Spinal Injury Association (ASIA) Impairment Scale (AIS), grades A/B/C or D/E. Several studies reported that early surgery within 24 hours after injury has a signi cant impact on the neurological prognosis of spinal cord injury.…”

Comorbid insomnia among breast cancer survivors and its prediction using machine learning: a nationwide study in Japan

“…The purpose of this study was to clarify the prevalence, severity and characteristics of insomnia in breast cancer survivors and to determine the clinical characteristics associated with the comorbidity. We also developed a classifier that predicts the comorbid insomnia using two machine learning algorithms, namely, the L2 penalized logistic regression model and the XGBoost model ( 12 ). Furthermore, we used the RuleFit algorithm to extract the hidden rules for segments at high risk of comorbid insomnia.…”