Abstract:Objective To explore the risk factors for early neurological deterioration (END) in patients with acute ischemic stroke (AIS), to develop a prediction model based on IScore, and to evaluate its predictive validity. Methods A total of 452 AIS patients were divided into an END group (n=73) and a non-END group (n=379).Risk factors for END between the two groups were compared. Predictive model was developed by binary logistic regression and the predictive validity of the model was tested using the area under the ROC curve (AUC). Results The incidence of END in this study was 16.15%.The predictive model included IScore score, white blood cell counts, moderate to severe stenosis of the middle cerebral artery M1,and moderate to severe stenosis of the carotid artery. The AUC was 0.790 with the specificity of 0.834,sensitivity of 0.635, accuracy of 0.798, and the Youden index of 0.469. Conclusion The prediction model based on IScore has good predictive power and repeatability for END in AIS patients.