Validation of CSR model to predict stroke risk after transient ischemic attack

Area of Science:

• Neurology
• Stroke Research
• Medical Diagnostics

Background:

• Identifying high-risk patients after a transient ischemic attack (TIA) is crucial for preventing recurrent strokes.
• The Comprehensive Stroke Recurrence (CSR) model, a neuroimaging tool, has demonstrated high predictive ability for recurrent stroke in prior studies.

Purpose of the Study:

• To validate the predictive performance of the CSR model in TIA patients.
• To compare the predictive ability of the CSR model against the established ABCD³-I score for stroke recurrence risk.

Main Methods:

• Analysis of a prospective hospital-based database of TIA patients meeting World Health Organization time-based criteria.
• Outcome assessment focused on stroke occurrence within 90 days post-TIA.
• Receiver-operating characteristic (ROC) curves and C statistics were employed to evaluate predictive accuracy.

Main Results:

• Among 1186 TIA patients, the CSR model (C statistic=0.754) showed similar predictive ability to the ABCD³-I score (C statistic=0.717).
• In the subgroup of 474 TIA patients with positive diffusion-weighted imaging (DWI), the CSR model (C statistic=0.725) demonstrated statistically higher predictive value than the ABCD³-I score (C statistic=0.626).
• 9.9% of patients experienced a stroke within 90 days.

Conclusions:

• The CSR model is a valuable tool for predicting stroke risk in TIA patients.
• The CSR model offers superior predictive accuracy compared to the ABCD³-I score specifically for TIA patients with positive DWI findings.