Radiomics-Based Intracranial Thrombus Features on CT and CTA Predict Recanalization with Intravenous Alteplase in...

Area of Science:

• Neuroradiology
• Medical Imaging Analysis
• Stroke Intervention

Background:

• Thrombus characteristics on non-contrast CT (NCCT) may predict recanalization with intravenous (IV) alteplase in acute ischemic stroke (AIS).
• Identifying predictive thrombus features is crucial for optimizing IV alteplase treatment in AIS.

Purpose of the Study:

• To identify the best radiomics-based thrombus features on NCCT and CT angiography (CTA) associated with recanalization.
• To evaluate the predictive performance of these radiomics features for IV alteplase treatment response in AIS patients with proximal intracranial thrombi.

Main Methods:

• A nested case-control study included 67 AIS patients with ICA/M1 MCA segment thrombi treated with IV alteplase.
• 326 radiomics features were extracted from NCCT and CTA images.
• Linear discriminative analysis and a linear support vector machine classifier were used to identify and evaluate predictive features.

Main Results:

• Thrombus radiomics features significantly predicted early recanalization with IV alteplase (AUC = 0.85).
• The combination of radiomics features from NCCT and CTA outperformed individual features like thrombus length, volume, and permeability.

Conclusions:

• Radiomics features derived from NCCT and CTA are superior predictors of IV alteplase recanalization in AIS.
• These advanced imaging biomarkers offer greater predictive power than conventional thrombus imaging metrics.