Abstract
Background
Approximately 54% of patients with endovascular treatment will have futile recanalization, and the cause of this is not yet clear. Therefore, in this study, we aimed to build a risk prediction model to Identify the characteristics of such patients
Methods
Acute ischemic stroke participants from the Daping Hospital, endovascular treatment Project database were included. The predictors of futile recanalization were identified by single- and multi-factor analyses; then, the least absolute contraction and selection operator regression model (lasso) was used to optimize the characteristic variables. Finally, the prediction model was obtained by multi-factor logistic regression fitting, and a nomogram of futile recanalization risk after endovascular treatment of acute cerebral infarction was drawn. Receiver operating characteristic curve and C-index, calibration curve, and decision curve analysis were used to evaluate the discrimination, calibration degree, and clinical utility of the prediction model, respectively. Finally, a bootstrap algorithm was used to internally verify the C-index of the prediction model.
Results
Finally, predictive models showed an area under the curve of was 0.975 (90% CI: 0.953–0.997). Age (OR: 1.09, 95% CI: 1.00–1.19), the TOAST classification (OR: 0.10, 95% CI: 0.01–0.76), infection (OR: 390.18, 95% CI: 6.18–24656.32), the NIHSS score at discharge (OR: 1.40, 95% CI: 1.18–1.67), and tirofiban use (OR: 0.13, 95% CI: 0.02–0.94) were included in the model.
Conclusions
The clinical utility curve of DCA showed good clinical utility. our results support the argument that Endovascular treatment after intravenous tirofiban use was associated with a lower rate of futile recanalization.