Predicting Recanalization Failure With Conventional Devices During Endovascular Treatment Related to Vessel Occlusion

Area of Science:

• Neurology
• Interventional Neuroradiology
• Medical Informatics

Background:

• Endovascular treatment is crucial for eligible stroke patients.
• Identifying patients with low recanalization chances pre-procedure is key for anticipating rescue strategies.
• Conventional devices include stent-retrievers and direct aspiration.

Purpose of the Study:

• To develop a preprocedural algorithm for predicting recanalization failure with conventional devices (RFCD).
• To identify clinical variables associated with RFCD.
• To improve patient selection for timely rescue interventions.

Main Methods:

• An observational study analyzed data from 984 stroke patients across 10 centers (2019-2022).
• RFCD was defined by specific imaging scores or need for rescue therapy (balloon angioplasty±stent).
• A gradient boosted decision tree model was built using pre-procedure variables.

Main Results:

• Recanalization failure with conventional devices (RFCD) occurred in 14.3% of patients.
• Absence of atrial fibrillation and no-thrombolytic treatment were independent predictors of RFCD.
• The predictive model demonstrated acceptable discrimination (AUC: 0.72) and accuracy (0.75).

Conclusions:

• Early rescue therapy (balloon angioplasty±stent) in RFCD patients was linked to better outcomes.
• A preprocedural predictive model using accessible clinical variables can identify patients at risk for RFCD.
• This model aids in optimizing endovascular treatment strategies for stroke.